🎧🍌 Solving the Hardest Problems in Dev Tools | Jake Cooper, Founder of Railway

Why there are only two moats, how AI is changing backend infrastructure, why managers should be operators, and why you shouldn't work on the weekends

This episode with Jake Cooper gets into Railway’s explosive growth in the first six months of 2025, how AI accelerates the need for strong backend infrastructure, when to build vs buy in AI software, and why there are only two moats: solving hard problems and doing hard things.

We also unpack Railway’s bold product bets, like enabling creators to earn revenue with backend templates, building their own data centers and cloud service, and not building their own AI models.

Jake also talks about their four week new hire onboarding, how they build a problem roadmap, why operators should be managers, and why you should almost never work on the weekends.

Thanks to Angelo Saraceno at Railway and Erica Brescia at Redpoint for their help brainstorming topics for Jake! 🙏

---

Presented by Ramp

If you're running a finance team, you know how much time gets wasted on expense management. Chasing receipts, categorizing transactions, waiting for expense reports. It adds up quickly.

Ramp handles all of this automatically. Its a corporate card and expense management platform that over 40,000 companies, like Shopify, CBRE and Stripe use to streamline their financial operations.

But here's what makes their corporate card different: every transaction gets automatically categorized and matched to receipts. No more wondering what that $47 charge was for three weeks later. You can set spending controls, get real-time alerts, and even block certain merchant categories. It's like having a finance team member embedded in every purchase.

The platform integrates with your accounting system and ERP, so everything flows through without manual data entry. Whether you're issuing cards to a few employees or managing spend across departments, Ramp gives you visibility and control without the paperwork.

Stop chasing receipts: check out ramp.com/ThePeel to get $250 and see what a corporate card can actually do for you.

Time is money. Save both with Ramp.

To inquire about sponsoring future episodes, click here.

---

👉 Stream on Spotify and Apple

---

Timestamps to jump in:

• 3:33 Solving the hardest problems in dev tools

• 8:16 Starting with the hardest thing

• 11:18 How AI accelerated the need for Railway

• 12:50 Importance of backend in AI-native software

• 16:52 Jake’s angel fundraise strategy

• 20:51 Resisting AI for so long

• 25:32 Using AI to get leverage

• 29:57 Build vs buy in AI software

• 33:22 When Jake knew Railway was working

• 34:27 Creating infrastructure templates

• 38:04 Building data centers and a cloud service

• 40:27 Two moats: Hard problems and hard things

• 46:25 Hitting 8-figures in revenue

• 48:47 Railway’s four week onboarding

• 54:25 Building a problem roadmap

• 56:16 You can’t set your own culture

• 1:01:58 Railway’s viral “How We Work” post

• 1:08:39 Using Discord instead of Slack

• 1:11:25 How hypergrowth companies mess up org design

• 1:14:03 Why you shouldn’t work weekends

• 1:19:45 Not betting big on AI models

• 1:21:53 Lessons from Zuck, Martin Scorsese

Referenced:

• Railway

• Careers at Railway

• The Inward Draw of Capitalism

• How We Work (volumes 2, 3, and 4)

Find Jake on X / Twitter and LinkedIn

---

👉 Stream on YouTube, Spotify, and Apple

---

Transcript

Find transcripts of all prior episodes here.

Turner Novak:

Jake, welcome to the show.

Jake Cooper:

Thanks for having me.

Turner Novak:

Yeah, thanks for coming. I think this will be fun. Really quick, for people who don't know, I've heard Railway described as solving the hardest problem in DevTools. What is that actually, and then what is Railway?

Jake Cooper:

I think it's the intersection of a lot of hard problems, and so that's what makes it quite difficult. So we are really easy-to-use cloud platform. You kind of just go to dev.new or railway.com, you start almost kind of spewing out the things that you want. You say, "Give me Postgres, deploy my GitHub repository. All right, give me a starter Python app, start writing some code, et cetera." And we, as the underlying infrastructure cloud provider, will automatically go in and provision those resources for you.

So we solve a lot of really, really hard problems in terms of delivering user experience to users. When people think about cloud infrastructure, they think of servers and... Like Kubernetes, and YAML, and all these other complicated things. And we try and abstract a lot of those things into really, really easy-to-use infrastructure by starting on a lot of the user experience stuff, and ending all the way at building our own data centers to power a lot of that experience. So there's a lot of hard problems in between, whether they're user experience problems, infrastructure problems, configuration management problems, versioning, deployment, rollouts, stuff like that. Live infrastructure.

Turner Novak:

So pre-Railway, what was I doing to do all this stuff? What did the world kind of look like?

Jake Cooper:

So, a lot of it or runs on things like Terraform. So you'll go and basically write your infrastructure as code, and then you'll be able to roll out and apply it. Which requires you to know every knob, bell, whistle, config, et cetera. So you'll pull in a definition, you'll go and set the properties you want to go in and do. Usually, you end up doing this actually after you explored some of the really, really complicated cloud dashboards that Google or Amazon or whatever have. You have to figure out how to version your config when you work on teams, and go and roll out these services. There's a little bit of knowledge or stuff you have to know as you're going to roll them out. So like, how do you do things like blue-green deployments? Which is I will spin up a new version of the service, I'll let it warm up, and then I will seamlessly cut the traffic. So that whenever you deploy new versions of your code, the thing doesn't go down.

So, those are things that you have to know the underlying bell stops, whistles, et cetera. And for us, the alternative is you just give us your GitHub repository, we will go and roll it out for you. And then we'll go and automatically manage things like that, and slowly expose all of those knobs, bells, and whistles instead of you having to swallow the whole thing at once.

Turner Novak:

So for someone who's maybe confused with a lot of the technical terminology that you just said, the idea is that you make it super simple to do a lot of these more complicated, disparate, traditionally separate, and harder to orchestrate all together infrastructure things. And you kind of tie it into one easy-to-use platform.

Jake Cooper:

Yep. Exactly right. We make all of that stuff that you would need to potentially retain somebody to go and manage those things. You can just do it yourself by saying, "Hey, I've got this thing that I had working locally on my machine. I wrote a little web server. It returns my gym sessions or my to-do list or anything else like that. I'm going to go and deploy it." And you can deploy databases, you can deploy functions, you can deploy pretty much anything. And we make it trivial to not just deploy simple stuff, but everything under the hood so that you can scale to infinity.

Turner Novak:

Why'd you call it Railway?

Jake Cooper:

There's a couple different reasons for that. I really like the kind of infrastructure, almost like monopoly utilities thing, because Railway is basically a utilities company for the modern web. We want to make it so that you can spin up everything super trivially. You just only pay for what you use. And it's just a thing that slides into the background, right? It doesn't need to be crazy or insane or anything else like that. It just works. And so trains are really, really solid from an infrastructure perspective, really easy to get around, et cetera. And so yeah, that was a good one.

But there's also kind of a nod to Ruby on Rails or anything else like that on there. So, it was a couple things.

Turner Novak:

I didn't even think of that one. That makes sense. And you said something about cheaper to you. So, what's the value prop? Is it that you're saving me time? Is it I'm cutting costs by using Railway? Why would I use it?

Jake Cooper:

So we want to make it as cheap as possible in both time and money. So saving a ton of time to go and spin up these things, and then only paying for what you actually use when you're going and doing that. We want you to spend as little time as possible on the configuration, drudge work, et cetera, so that you can actually go and build the thing you actually wanted to go out and do. We won't start a business to go and learn about Ansible or Kubernetes or any of this other thing. But current generation of software, if you want to start a business that'll actually scale for you, you do have to go and learn these things, unless you use the tools that we were building so that you can actually spend all of the time working on the actual problem that you started your business to solve.

Turner Novak:

So when you were starting this, what was order of operations like? What do you start with? Because I think, when I think... All the things you do are separate companies and separate products, so you have to build all of those different things. How did you get this initially off the ground? Was it certain things you started with? What was the process of doing that?

Jake Cooper:

So I started with the thing that I thought would be the hardest, if that makes sense. Because I was like, "Cool, if we can actually solve this, then we have a really, really good shot at solving all of the other stuff." Turns out, I was really dumb and a lot of hard things. So, what I thought was hard was actually... It's still hard.

Turner Novak:

It was the easiest thing.

Jake Cooper:

Yeah. Well, it wasn't the easiest, but it was definitely somewhere in the middle. Maybe, I'm in the 75 percentile, but...

Turner Novak:

So if you go back, would you do a different thing first?

Jake Cooper:

I don't think so. I'm a big believer in the butterfly effect. I think things are going well, so I wouldn't mess with anything in general. The thing that I started with was databases. And so, databases are actually pretty hard. Stable storage ends up being actually quite difficult on the almost iceberg of complicated infrastructure things. Because people have data in their databases, and if you lose the data in your database, then you have a terrible time or whatever. So, I started with that because I was trying to spin up a whole bunch of different side projects. And every single time, I was like, "Oh, I want to use side projects or something like that." I'd have to wait essentially, and go and spin up something, the AWS dashboard. And then, configure any firewall rules or anything else like that to get this working. So I was like, "Oh man, there's got to be a better way."

So I started with a really crappy script that just ran out of a box. I could call the endpoint and it would spin up a database, and then I'll just kept using that. And then as I was iterating through different ideas for stuff I wanted to work on, I just kept expanding this script and people were like, "Oh, this is super cool. Can I spin up not just Postgres?"

And I was like, "Yeah, sure. Why not?" And so then, they started using it. And then I was like, "Okay, I'll just launch it on Twitter or whatever, and people can use it." So pushed it out, and then people thought it was super cool. And then I basically asked. I was like, "Okay, if I'm... I want to work on this thing. This is where I think that I'm going to go in general. Who's working on this?" So, I went around to a lot of different people. I went around to Jason Moore who was the CTO of GitHub at the time, and I was looking on poolside. I was like, "Hey, Jason, were you all working on this?"

And he was like, "Yeah. Something's got moved around as we absorbed into from GitHub to Microsoft, but this is super cool. We're not actually actively working on this."

"I think it's awesome. Do you want to take an angel check?" And so, I went around to a bunch of different people. And at the end of it, I had no job offers, but a bunch just commits for an angel check. So I was like, "Okay, let's see where this goes."

Turner Novak:

You were actually trying to get a job sort of in this process, too?

Jake Cooper:

Yeah, because I wanted to work on this thing. And I was like, "Okay, I know I'll start a company at some point in general, but just tell me where I can work on the cool problem and we'll just work on the cool problem."

Turner Novak:

So, you wanted to join a company and build this?

Jake Cooper:

Well, I just wanted to build this. I didn't really care in general. I was like, "Okay, this seems good. Seems like a thing that is pretty evident people are going to want it." Et cetera. "Somebody must be building this." And then went around, couldn't find anybody who was actually working on it from the angle that I wanted to. So I was just like, "Okay, let's just run with it."

Turner Novak:

Because you think if it's all these different features that already exist in all these different places, shouldn't someone have started to expand what they did and start to do the real... Like, your thesis of someone else should do this. Did you know why it didn't already exist?

Jake Cooper:

I think it was kind of a good enough thing, if that makes sense. Probably like, "This has all the things you need." In aggregate, right? It is the everything store. And I don't think people were building software at the rate that they are right now. And I think especially with the AI stuff, it's accelerated a lot of those things. And so we're seeing massive, massive uptake in terms of people who are using and starting to use our Railway. We've five extra top-of-funnel in the last six months. So tons and tons of people are signing up because the rate of iteration of software is growing so quickly, and now those old primitives really aren't keeping up. If somebody wants to vibe code some sort of thing, and then they slam it into Ansible, Kubernetes, etc, they're just going to be like, "What do I do?"

I don't know if you've seen it, but there's people on some subreddit being like, "Hey, how do I not lose my progress for the last six months? Cursor just erase it." And it's like, "Oh my god, these people." They don't get versioning or any of these underlying things. So there's this whole kind of oasis or ocean of people that are coming to go and build software, whether they are people who are already software engineers moving a lot quicker or new software engineers trying to build something. They're always looking for something way quicker because the iteration loop has gotten so much, so much tighter with a lot of the AI tooling.

Turner Novak:

So if you're a backend pro, you're like the Gilfoyle in Silicon Valley, you might see where I'm like, "I don't need that. I can yield all this." But that's probably a pretty small percentage of people. The average person starting a company, I feel like, most people are not backend experts.

Jake Cooper:

Yeah. So the long tail of people who would not be suited for Railway is diminishing by the day basically, because we can deploy anything now. Actually, anything Asterisk. We can't deploy GPU infrastructure just yet, but that is dwindling. And so too, is the market growing of people who are just saying, "I just want to move a lot quicker." And so as our capability gets significantly better and better and better, and those people wanting to play faster and faster and faster, we're reaching this inflection point where most people are actually almost capitulating out of that older mechanism of building software and moving into the realm of the thing that we have built. So, that's super cool because I think for a long while, we're building a thing and we go and pitch it to VCs, and they'd be like, "What problem is it solving?"

"Well, it's offering infrastructure sucks."

And they're like, "Yeah, but tell me more about this thing." And now, it's people want to move significantly quicker. And in the age of AI, there's almost a very, very material risk of being out-executed by your competitors because they're moving so quickly. And if you don't have the tooling or infrastructure to move at that pace, you're going to get left behind. So now, we're in a more interesting spot where it's like people are moving because it is simultaneously cheaper and faster to use the tools that we've built.

Turner Novak:

And it seems like probably at first, you would getting a big customer. Like, a bigger enterprise or mid-market business switchover may have been hard. It was just like you got to get somebody starting a company to start building on it. But then is it now, it's a little, you're seeing people actually switch and bring their whole... Like, re-architect their entire system on top of Railway?

Jake Cooper:

So the nice thing is that we support all existing open source primitives, and so you can actually just go and import things from existing infrastructure that you have. You have docker files, that's fine. You have your code base. We can go and automatically go and detect these things, et cetera. We can import just various different configuration files to automatically import you. So, you don't have to do too much of that in general. So bigger companies, then the benefit actually in this is that nobody wholesale cuts their infrastructure. They don't do it.

Turner Novak:

Which, what does that mean?

Jake Cooper:

Nobody's just going to basically show up and say, "Well, I spend $2 million a month on AWS and I want to move all of it over here next week." They just don't do it.

Turner Novak:

So it's just like, "Let's just try. Let's just dig in, turn, and set this up." One team use for this was...

Jake Cooper:

The way that it usually happens is somebody either hears about us or already starts bringing it inside of the organization for some sort of like, either Tiger Team project, or thing that they were wanting to set up internally. And so they ran into, essentially, this platform team edge of like, this team potentially internally that manages all of this tooling. They've exposed them in various different ways that ends up being cumbersome. And that's through no fault of the platform tooling. That's actually the underlying infrastructure. The infrastructure that they're trying to abstract is very, very difficult. And in my opinion, is almost unabstractable. It is at its most baseline form, the least complicated you could make that infrastructure. So, any attempt to move around actually just ends up pushing the complexity around.

So these teams who would end up otherwise spinning up things on that infrastructure can't, right? So, they're looking for a new way to move things around. So they'll spin up something on Railway just to start it, and then go from there. And so, that's what easily happens. Somebody spins it up, it grows inside a little bit in the organization. And then at some point, the OPSEC, antibodies of the organization basically say, "Oh, what is this tool? It's not on an approved list or anything else like that." But that gets us into a sales conversation with the people at the company because there's hard things running on it, because it was so easy to go and spin up anything on it.

Turner Novak:

So then going back to when you were first going through this process of it's sort of a side project, became a company, you... I mean, you got a lot of really impressive people though to invest early on. Did they invest before you had a fully baked product, and vision? Customer signed up? Or was it kind of like parallel path, that? How do that come play out?

Jake Cooper:

So yeah, they definitely invested. The demo I gave them was one Postgres instance with a shared user. So if anybody actually just got into it and ran anything on it, they could just nuke the entire database. So, it was a really, really poorly fleshed out in general. So it was just the inklings, just a bit of that almost aha moment of, "Oh, I can do this thing, and here's how I do it in the future." And that was the demo that we went forward with. And as for how we found them, I think it was X or stuff like that. I was just doing the build in public thing, sharing, et cetera. And then I said, "Oh, if I were to actually go in and raise money for this thing, who should I talk to in general?"

Turner Novak:

So was your approach like, "Hey, Turner, I'm raising money. Please invest." What was your approach?

Jake Cooper:

It was more like, "I am building this interesting thing. Here's this interesting thing. Who would be interested in giving money to me to build this interesting thing?" And then I had a bunch of people like Tom Preston-Werner, who's really, really lovely. He raised his hand in the comment and was like, "Oh, we should just chat."

I'm like, "Oh my God, I love Twitter or X or whatever." It's just this bridge that brings a ton of people together. So it was a lot of that honestly, of people just putting up their hand or me seeking them out from that job offer perspective of those two things. And then at the end, we had a half a million dollars worth of angel commits. And then it kind of just all came together in that, and then started doing it.

Turner Novak:

So, would you recommend approaching him more like that is a... When you were getting that first round, you need a million bucks, whatever the number is to get off the ground, less of a, "Hey, I need money." You're talking to people in the context of, "I'm fundraising. Please invest." Treat it more of like a looking for advice, looking for how to unlock, maybe meeting new people. You approach it more that way. That's generally the approach that you took, it seems like?

Jake Cooper:

Yeah. I think this is philosophy that I have just generally. It's almost like you should delay, delay, delay, and then just go all in. It's almost you want to wait until you're really, really confident, and then you just be really, really confident. And almost like turn on a dime. So as an example, I wasn't super into any of the AI stuff as early as basically this... I think January basically is when I started just really digging into it. And then I saw it, and I was like, "Oh, this is really, really cool. And it will work." And I see the way that it'll work. I see the way this is going to break in general for a lot of stuff, but at least it's solid enough ground to build basis on.

So in the same vein, I think that you should continue to work on your project, and make sure that it's going to be something that you're actually going to want to spend. Worst case scenario, 3 to 5 years on. Best case scenario, like 10 to 15 years on or whatever. Because that's how long it is. It's a long journey to be able to go in and build something, et cetera. And once you raise money, you're sense of committing to run that gambit, right? You are committing to deploy the capital, run after the strategy, go and build the team, et cetera.

Turner Novak:

Raise more capital, increase the team.

Jake Cooper:

Increase team.

Turner Novak:

More products.

Jake Cooper:

Once you step on the treadmill, you are on the treadmill, and you should know that you're going to continue to go and do that. And that's when your marathon starts in general. And I think a lot of... And they just angels out there who are like, they're... I know there's a whole meme of VCs want to be helpful, right? But angels aren't genuinely... Some of them feel like they got lucky with a lot of their hard work. And obviously, they've thrown a lot of hard work to it, but they now have money in spades that they want to actually go and almost pay it forward.

So they're potentially looking to say, "Okay, these people who are really, really hungry just as I was that are ready to go on the journey, how can I potentially go in and help them?" We've had the fortune of working with those people who just want to pay it for it. So, that's awesome.

Turner Novak:

And you mentioned you were sort of resisting AI. You're like... I don't know. But then, there're, suddenly, like a foot switch. So, what's been going...

Since ChatGPT launch, what's been your evolution of thought process on everything AI?

Jake Cooper:

So my evolution of thought process was like... That's just funny because my roommates been at open air for the last seven years. So, I'm sitting here being like, "It's cool, man." Or whatever, right?

And then he's like-

Turner Novak:

Like, a nonprofit. Come on.

Jake Cooper:

Yeah. It's fine, right? We go to the sauna every month that there's like, how's working? Ah, okay. Cool. I wonder when that thing will calm down or whatever. And then at some point, it was basically just like, "Oh, I started using it for a bunch of different things." I think I honestly was probably holding it wrong, if that makes sense.

Turner Novak:

Right. Like, you were not using it appropriately or you didn't know how to take advantage?

Jake Cooper:

Yes. Didn't know-

Turner Novak:

If you were one of those people... When we see those threads, like 99% of people are using ChatGPT wrong.

Jake Cooper:

Yes.

Turner Novak:

You remember those? This was a long time ago.

Jake Cooper:

Yes.

Turner Novak:

So you didn't read those threads, obviously.

Jake Cooper:

No. Not a fan, dude. I'd be so powerful if I could read anything. It's just I miss things like that. So yeah, it was one of those things where, I think simultaneously, it crossed two bridges in my mind. It got sufficiently good where I could outsource small pieces of work or anything else like that, and I got a good enough flow on it where... And this is still my flow to this day, which is I ask it to do something. I'm very, very concrete about it. I already know how to solve the problem. I don't ask AI to do that I don't know how to go in and do. I only ask it to do things that I already know how to go in and do, so that I can gut-check it, etc. Right? So for me, it was a lot of... I tried to ask it to solve things in various different levels of success, and I would also just end up with a ton of stuff. So, a lot of the interesting stuff that's happening right now in my mind is actually...

People talk a lot about the greenfield stuff. I can go from nothing to to-do list app or whatever. I actually find that not that interesting because I actually want to set up the initial scaffolding, because I want to know how it works in general, and then I want to ask AI to fill out something way quicker than I would. And so as a very reductive example, I would be like, "Parallelize this for loop for me. This is blocking. I need you to actually just parallelize this." I already know how that's going to work in general, but I can save a ton of time of just like, all right, write the keystrokes or whatever. And so I've been using it to various degrees on that where I'm just like I'm going to write down an RFC on how to do this database migration. We're going to move sessions from Postgres to Redis. This is how it's going to work. And then I just hand it to Claude or Cursor or whatever, and I say, "Do the first step. That's our first pull request, come back to me." Et cetera. Very, very deliberate.

So, it's almost that outsourcing to a junior engineer that is a bit more concrete, kind of type of person.

Turner Novak:

So, you're just a junior engineer. I feel like before, it was outsourcing to an intern or a shitty intern.

Jake Cooper:

Yes.

Turner Novak:

But I feel like that's how I initially... When all these people explaining what they're using, ChatGPT 4, and I would try it. And I would be like, it's kind of like the intern that you hope doesn't get a return offer and leaves, and they aren't very good. It's like, that's kind of was my relationship where it was like it wasn't a good product, but you could see that it was getting better. And I feel like probably to your point, within the past 6 months, 9 months, for the things I use, I'm like, "It was actually pretty good. It was actually usable. This is interesting in a work product." I think it's always been better Google. That's always been there and it worked for that or whatever. I feel like in the last, probably since very early this year or late last year, I've noticed the same thing where it's like you can actually rely on it.

Jake Cooper:

Yep, it's good. It's like a bionic arm, that once in a while just malfunctions. But it malfunctioned in ways where you can be like, "Oh. Okay, stop. Just don't do that." So the way that I've taken to it is basically saying, "Can you one or two-shot this task? And if you fall off the rails too quickly, I'm just going to bail and just do this myself."

Turner Novak:

I do that too, too.

Jake Cooper:

Right? And so that's the utility that I actually found really, really good for it because it's almost a separation of concerns. When I try and work with it to solve a hard Brahma, it's like me and Claude are sharing one brain cell, and it's really not a great time. And we're both like... You know, right?

Turner Novak:

Like, you're trying to figure it out together.

Jake Cooper:

Yeah, we're trying to figure it out together.

Turner Novak:

You see blind people wandering through the dark.

Jake Cooper:

Moving around where you're just like, "Okay." Right? And so that, I haven't found great success on it. But separation of concerns, like the splitting of the... I think it's a bicameral mind of, again, where you basically say, "Do this for me."

"Okay, I'll do that for you." That has actually worked quite well, right?

Turner Novak:

I know you're really big on automating things, and just getting leverage and automate. Do you use AI for that? Are there other ways to automate things? How should somebody think about trying to do that? Getting more leverage from automation?

Jake Cooper:

One thing that I say often is grit process repeat. So, don't try and automate things right off the bat, right? There's that very famous xkcd, which is, say, you automate a very small task that happens very infrequently, but it takes you an entire week to go in and do. It's like, that's a waste of time. You're not going to get the return on investment there.

So, we talk about leverage. I'm looking for opportunities that I can put in a bit and get a lot out of it. And so, running that calculus on every single thing that you can actually go in and do, that helps us quantify leverage. So I think there's a lot of different spots, and I think the reason people really, really like AI is because you put in a little and you get out a lot. And so, this is your equation regardless of whether you're using AI to go and do this thing. Put in a little, get a lot.

Turner Novak:

Have there been any examples where, personally or internally, that you guys have gotten the most leverage?

Jake Cooper:

One that I reach for often and people are like, "Whoa, you did this? This is nuts," is we built our own support tooling.

Turner Novak:

There's a couple of startups that do that.

Jake Cooper:

There are a couple of startups that do that.

Turner Novak:

There's also, when you look at the whole meme of AI is going to replace all software, that's a common example of we built our own support system.

Jake Cooper:

Yes, exactly. So we built our own support system, and it's awesome. I was using it the other day. And it's a bit of an mess internally this last quarter, because we've been pushing a ton on growth and also turning off a bunch of our Google Cloud machines to move around bare metal infrastructure. So it's been a total zoo from just people have questions, stuff like that. And we built this system that's like... I feel like I'm basically playing StarCraft because it's like high APM gaming where I'm just in there, just-

Turner Novak:

In the support system?

Jake Cooper:

In the support system. And genuinely, it's a deeply wonderful experience where you're just tearing through this thing and you're just like, it's this. I've got macros for common questions. I've got a little AI prompt on the side that gives me context on the thread. And all of this is stuff that we've built internally. And so, this is a thing where we have a million users, we have three support people. And we all rotate throughout the cube, but we build a system so that we can actually go and operate, and figure out how quickly can we get back to these users with the level of quality that we really, really need. Because for a time, we were actually really, really bad at support. People would give us tickets, and it would take 5 or 6 days to respond on some of these. It was awful. And we looked at this thing, we're like, "Damn, this is horrible." Totally understand where customers are coming from, et cetera.

I was like, "How do we get to 24 hour SLA on everything, and then lower the 24 hours on enterprise and anything else like that?" So, we build all of the tooling around that. Some of that is AI, some of that is building these things ourselves. Some of it is just solving hard problems of design and engineering, so that you can actually gain all the contacts you need. Railway is interesting because since you can deploy anything, people ask a lot of questions about a lot of disparate things. And so, trying to figure out ways that you can pull together knowledge or start putting people together in the community. We can't answer every single Python question. It's just untenable. So, how do we give these people topics? Or we have these things called bounties in the community an hour. It's like, "Oh, cool. You're..." If you're having throughput issues on Python or anything else like that, we can actually go and kick it to a queue of people in the community who are going around and solving these bounty questions. They get paid every single time they go and solve any of these problems.

And we also go and push those people together. So it's like, "Oh, that person is an expert on Python." So the Python questions are actually getting routed specifically to that person, and they can go and help and solve those things.

Turner Novak:

So is there like a internal forum or something somewhere, or... Okay.

Jake Cooper:

Yep. Maybe a contrarian tick that we have is forums unfortunately have come the way to dinosaur and everything's been pulled into Discord communities, so there's not a ton of stuff that's indexed. So, we'd actually went out and built a ton of tooling around indexing these threads, writing users together, pulling any telemetry and information from, actually, any of the Railway projects to say, "Oh, this person is deploying Python and they've recently done these things, et cetera." These are all hints of how we could actually potentially solve the problem for them.

Turner Novak:

You build a lot of custom stuff, it sounds like, but it's not core to the actual Railway product. How have you rationalized, "Let's build this cool, interesting, unique thing," versus, "Let's just buy one of the AI customer support, or it's a forum software out there." How did you decide on what to do there?

Jake Cooper:

So I guess for us, it was what are other people bad at almost. That's an arm for us, basically. And so-

Turner Novak:

So, what are your competitors bad at?

Jake Cooper:

I wouldn't say competitors necessarily, but almost like what is the industry, as a whole, bad at. Software infrastructure as an industry.

Turner Novak:

Like, at customer support?

Jake Cooper:

I don't think they're very good at customer support. I don't think they're very good at developer experience, et cetera. So ease of use, all of those other things. And I don't think they're good at speed. So we've chosen to invest really, really deeply in all three of those, and have them all essentially feed into each other. So our customer support experience is a product experience that we design. We work in it in-house, our designers work on it, and engineers work on it. Everybody works on a piece as part of the product. So, that's kind of tied in there. And it's fast. You should be impressed with the level of support that you get, and you should be impressed with the product. And now, we routinely have people being, "What is this piece of software? How did you get... Is this open source?"

And we're like, "No, we built this. Sorry."

Turner Novak:

Would you ever open source it or sell it, or-

Jake Cooper:

Well, what we'd like to do at some point is actually bring more... We're having conversations with companies to bring them onto the platform as almost a discussion space. Because you see a lot of these things where people are like, "How do I get the voice of my community? How do I go and do all of this other stuff?" What we can do, because we have all of the telemetry, we can figure out... Let's say, you publish a new version of your software. You go and send it to the community. Now you have issues, et cetera. We can take all of the stack traces, anonymize them, and go and send them back to you and say, "Hey, listen. On this version of Redis or on this version of whatever, you introduced a bug on this line. You should go and fix it."

And it's now affecting these 10,000 people. "Here's a funnel to these 10,000 people if you want to talk with them and figure out actually how to solve reproduces, et cetera. These things." So those are things that I don't know if we would actually sell the software outright, but we want to basically make sure that it's a, not just a cost center, but actually a revenue generation center. And it is, because we do an open-source kickback program. So if you host anything on Railway and you go and build a template for it, you get 25% of any of the revenue that you go in. That people get on... Or that we get when people go and deploy that piece of software.

Turner Novak:

Oh, interesting. So it's almost like the Notion-template type of strategy for infrastructure.

Jake Cooper:

Yeah, right. Or the figment plugins or anything like that.

Turner Novak:

Yeah, interesting. Okay. And then, we talked about complex, complicated product. You started with just databases. How did you think about expanding over time? Any key product decisions that you feel like you've made?

Jake Cooper:

It's a lot of the same thing that I mentioned previously, which is like, no, no, no. Fuck, yes, let's do it. So listening to the users, not necessarily saying we're going to ship everything because we don't want to make software nachos. We want to make actually a really, really solid product.

Turner Novak:

What's a software nacho?

Jake Cooper:

It's like everything under the sun. You're just adding everything to it. It's from The Bear. When there's a chef, that's basically-

Turner Novak:

The restaurant show?

Jake Cooper:

Yeah. Where it's like you're working on fine dining thing, and the chef is basically like, he made nachos. This is not a fine dining experience. So, we want to make something that's a very, very concrete and deliberate, and it is this thing. So, that's what we want to go and do from that perspective.

Turner Novak:

Was there a point when you felt things were really starting to work? Like customer... Like, there's almost like a tipping point in business modeling. It's in Columbia College. It's like finding PMF. It was a certain thing where it's like, "Holy shit. I think Railway is a thing. This is going to be a pretty big..." Like, it's really working.

Jake Cooper:

Yeah. I would say honestly, it's in the last six months where it's like we had a conversation with the team, where I was basically like, "The oil rig is set up. We can run this thing straight upward." We just make it faster for people to build and deploy software. We make the experience better. We go and push towards this vision that we have of a unified new way to build software. This can definitely work. And we're seeing it now. We're seeing it with the inflection point of so many new people coming onto the platform, and trying to build and deploy something. And then running to edges of the platform, they were like, "Oh man, I guess we got to fix that." But it's just so many new people just doing this where it's starting to really take off in my mind.

So, it's the closest to product market fit I've ever felt. But it only keeps getting crazier. So let's say, it just keeps going.

Turner Novak:

One thing you mentioned, you built your own data centers.

Jake Cooper:

Yes. This is a Railway metal T-shirt. It's like metal, and it spells Railway.

Turner Novak:

Okay. So maybe that's the answer to this question, and maybe it's one of the couple answers to this question. But what do you feel have been some of the more bold product decisions that you've made over the past couple of years?

Jake Cooper:

I would say metal is really one of the making one-click infrastructure. And templates is another one that is paying off. We have people in the community making six figures off of these templates. So, they're actually getting...

Turner Novak:

So what is this? Is it like a backend template? If I want to do this... So I'm listening, like, holy shit, I can make six figures a year making Railway templates. What am I making?

Jake Cooper:

So it's software that people want to go in and deploy. So whether-

Turner Novak:

So it's like a system design?

Jake Cooper:

Kind of, yeah. So if you're saying, "Hey, I want to start a software project and I write Python, and I want Sprite." It's like these boilerplate things that people have gone and built. Or it could be packaged versions of software that requires a little bit of almost DevOps handholding to go in and spin and scale. So, we provide the tooling for it. You consume this package in a way that you are, none the wiser, of how it should scale. And then the person maintaining it actually goes and figures out how this thing should scale over time.

So, self-host ClickHouse, or n8n, or Plausible, or any of those other things. So it can start with some community version of the thing being hosted, and then it can move towards a first-party version of Redis or ClickHouse or whatever hosting their self-hosted distribution on Railway.

Turner Novak:

So if I'm using someone's template and they make a change to the template, does it change my infrastructure setup?

Jake Cooper:

No. So you get a notification that there's a new version of it, and so you can go and consume that. You can go and read about the changes. You can go and see all of those things in there. So-

Turner Novak:

Do I want to accept their change of, like, they figured out a way to actually speed something up by changing the connector-

Jake Cooper:

So it's like 25% faster on workloads for asynchronous rights or something like that. And then you're saying, "Oh. Okay, cool. I've had those issues in general." So instead of this being a thing where GitHub is, through no fault of their own, a series of broken pointers, you're basically saying, "Okay, how do I go and get the new version of that software? How do I go and update? Is it a home chart? Is it something else?" Like, as in docker version or something? Figure out where they host a change log or anything else like that. It's like we put all those things together so you can riffle through this version of software and then get the latest version of it, and then also pin it so that you go and downstream-push a version of this. And then now, brace everybody to the apps or anything else like that.

Turner Novak:

Like, versioning control for our infrastructure, sort of?

Jake Cooper:

Yeah. And so this is the point of the earlier problem. We're solving a lot of problems with how do you not just deploy your software, but how do you evolve it over time? How do you iterate on it in a safe way? Actually, at large companies, most of the software outages is not part of a software error. Like, you shipped a bug or anything. Those big outages that happen on Facebook, CloudFlare, whatever, that's a config-based outage. It's actually a property that they were trying to go roll out when they were upgrading system.

Turner Novak:

And just something got fucked up during the process?

Jake Cooper:

Something got fucked up because the peak under the hood and how the sausage is made is, essentially, there's a version of a piece of software, and then there's an engineer whose job is to version and update that piece of software. And how that happens can range from one click, really easy to consume to, I'm doing basically open heart surgery on this piece of software to get it from version N to version N + 1. And we automate the entire life cycle of that.

Turner Novak:

Interesting. And then you also build your own data centers.

Jake Cooper:

Yeah.

Turner Novak:

What was the thinking on that? Do you have a cloud service that you're offering to people? Can you just explain this whole concept?

Jake Cooper:

So, the thinking behind it is that we can actually do a lot of really, really cool stuff and build a better experience because we have our own hardware. We are moving from the realm of milliseconds down to microseconds. That's some experience that we can actually make better. It's a portion that we can make faster. We can push the edge of that experience because we're not actually beholden to operating through the primitives that the cloud providers offer. Which they're solid, but at a certain point, those things end up falling flat from both ability to drive the experience, and also, outcomes that you'll have at scale.

Turner Novak:

So, people deploying on the Railway cloud now?

Jake Cooper:

Yeah. It's essentially a cloud that you can go in and deploy out. So yeah, the reason to go and do that is to drag the product experience forward, and do things that other people possibly can't.

Turner Novak:

What's an example of something that they wouldn't be able to do?

Jake Cooper:

So on AWS or GCP or anything else like that, we can spin up workloads, it'll take a few seconds or anything else like that. We can get into millisecond deploy times now, because we have a lot of our own infrastructure. And it may be possible to go in and push those things out. And we’ve done a lot of really, really cool work on just spinning stuff up, but it ends up being on the realm of a second-ish. And for something to be imperceptible from a user's experience perspective, you're looking to try and be in the 100 millisecond range.

So we want to be as quick as possible, and racking our own services allows us to move even closer to that medal of moving really, really quickly. It also allows us to offer a significantly reduced price to our users because we are running our own infrastructure, and because we've built our own orchestrator, and because we've done essentially all of the hard work of going and doing these peering relationships. So, we pay for network at a metered rate instead of a per gigabyte rate. So that allows us to drop that price significantly, which is super cool, and a lot of the AI workloads because they're pretty data-intensive.

Turner Novak:

Yeah, that's right. And you wrote a post a couple of years ago. It was called... I think it was called the Inward Draw of Capitalism. Did I say that right?

Jake Cooper:

Yeah. Yeah, that's a deep cut.

Turner Novak:

Is this related to this concept?

Jake Cooper:

Sort of in the sense that I think there's only two moats. Hard problems, doing hard things.

Turner Novak:

Are they not related?

Jake Cooper:

No. They're related at the edge, basically. So the point of that post was that it's easier to make, basically, money with money than it is to make progress with money. And we've moved because we moved out the gold standard to actually a system where we're moving money around. And that results, essentially, in people doing jobs like being a day trader or a banker or a Twitch team or moving those things. And there's nothing wrong with that in general. It's just that we've made a lot of the exterior a lot more difficult to do. And so, if I want to go and get a loan and build a factory or anything else... This is before AC Steensy and the whole...

Turner Novak:

Yeah, it's easy now. Just raise 50 million from AZZZ and build a couple of factories.

Jake Cooper:

Yeah, right. So maybe, I was ahead of something or whatever. But if you want to go and do that, you have to go through regulatory burden, you have to find construction workers, you have to go... You have to go and do all of these other things. There's a lot easier ways to make money, but that whole point was that we're losing a lot of the actual fringes of society that allow us to push towards progress and actually building things. And so another reason, and we're not here using investor money to fulfill this philosophical burden or whatever, but we believe that there is a strong arb in terms of doing complicated or difficult things that are almost becoming a lost art. If you think of the people who are racking and building data centers, you've like three people. AWS, GCP, Azure, right? Those are the big people that come to mind.

And so over time, you're getting less and less and less of these people who have esoteric skills, and you can extend that to kernel knowledge. You have people who are working on the Linux kernel maintainer list. They're having difficulties getting maintainers to have this knowledge and work on this thing. There's a core part of, literally, everything that we run on every piece of stock. It's every single thing. And that thing is such esoteric knowledge, and they're having issues pulling in maintainers and people who are going to go in and push this thing forward. It's crazy that that knowledge is getting more sparse over time.

Turner Novak:

So, you've made some of a lot of pretty bold batch. Are there any other big ones that you've made? And maybe any that didn't go well or didn't work, or anywhere you got a lot of pushback from investors or people on the team. They're like, "Jake, this isn't going to work. Don't do this."

Jake Cooper:

What are some bad ideas that we had that we did? I mean, we spend a million dollars on a domain, which I think was actually good where it works well.

Turner Novak:

Just the railway.com?

Jake Cooper:

The railway.com domain.

Turner Novak:

Who owned that before?

Jake Cooper:

Some guy in a flyover state. So, I think it was funny. I looked at his LinkedIn, and he was working at some company for a while. And then I think he recently retired, so I don't know if they're correlated or whatever.

Turner Novak:

You buy the domain, you paid for this guy's retirement.

Jake Cooper:

Yeah. Who knows? Or it's college fund or something. So, that was good. The meta data centers was solid. We've done a lot of experimental stuff in terms of how do you build out an organization internally. We talk about a leverage a lot. And so, I've tried to build out a pretty deliberate management stack internally of how do we stay lean but also have some level of hierarchy, so the decisions can get made. There can be deer eyes. There can be stakeholders. All of those other things that you need at a company while also simultaneously avoiding a lot of the bloat that comes with growing really, really quickly or anything else like that. So, we've done a little bit of organizational operating. The org is still mostly flat and we have 28 people. I essentially have 25 people reporting.

And so that's probably a bad idea over time, and so we're lowering the managerial burden on that as we flush it out. But it also makes... It's really, really important to have operators be also managers. The worst thing that can happen is you hire a career manager who can't execute on the tasks at hand. Shit hits the fan, they can't actually step in or anything else like that to aid or help. And it makes the organization feel very, very fragile or helpless over time. And also, from talking with a lot of people, I was at Uber post-Travis, posts-anything. A lot of the experiences that I have from chatting with people, they were saying like, "Oh yeah, we grew really, really quickly."

There was a year where, I think, they Forex headcount or something like that. It was insane. And so, the cultural sheer that actually happened over that period of time was nuts. And you're like, "Yeah, it wasn't the same company."

I was like, "Yeah, kind of no shit. Because 80% of the people after one year are new." And so when you think about that from a cultural perspective, it's like how could they possibly understand the most operandum of that company? So yeah, I think that we've tried to innovate a little bit on the organizational structure to try and grow... Our gold standard is we want to grow 5X year over year, 1 1/2X headcount.

Turner Novak:

Oh, grow the business revenue 5X.

Jake Cooper:

The business, yeah. Metrics, in general.

Turner Novak:

Metrics.? Okay.

Jake Cooper:

So acquisition, top-of-funnel margins, stuff like that.

Turner Novak:

And you said you grew 5X in the last six months or something, right? I think you slipped that out earlier. So, you guys are ahead of plan?

Jake Cooper:

Yeah. So, ahead on that one, and then revenues in a really, really solid spot. We'll have doubled since the start of the year, which is great. So things are going, I would say, quite well.

Turner Novak:

Nice. What's current state of the business you're allowed or open to sharing publicly with people?

Jake Cooper:

We're in the tens of mil from a revenue perspective.

Turner Novak:

So, somewhere between 10 and 99?

Jake Cooper:

Yeah. Basically, yep.

Turner Novak:

Okay.

Jake Cooper:

Like, wide range. So, that. And then we're like 28 people, staying super lean. We've raised some money. The last announced round that we did was a Series A in 22 or something, but we raised an uncapped safe about a year ago.

Turner Novak:

Oh, interesting. What was the thinking on doing that?

Jake Cooper:

We didn't want to take any more money, but people wanted us to plug capital. So it was like, the only way we're going to do it is on an uncapped safe. And they're like, "All right."

Turner Novak:

Sure. Wow. How do you think about that as a founder then? Do you think I should probably try to convert this for them at some point, or is it like...

Jake Cooper:

It's an investor we really, really like working with, so we'll definitely be converting it at not agree. Just... We really like working together in general. So, our philosophy is they wanted to put in more capital, and we really, really like working with them. So this allows us to continue to push the business into a really, really awesome spot as we continue to look for potential excellent, additional board partner for that.

Turner Novak:

I think there's an interesting thing too, about not getting your valuation to extended from a standpoint of also recruiting. Let's say the business is valued on paper at 50 million and 500 million, right? I have no idea how close any of this is to the actual numbers. But if you're trying to recruit someone and you say, "Hey, you're getting your options or your RSU's at 500 million versus 50." You've taken out 10X upside from them and the interest in joining. The opportunity that they get from joining.

So, you want to balance that. You don't want to delude yourself too much. You want to maybe make existing investors feel like they're making money. And so, it's like that balance of not getting too far ahead.

Jake Cooper:

Yeah, 100%. You also want to be able to reprice equity at some point, to show it's like, "Oh, it's not just mystery money or whatever."

Turner Novak:

It's not like, is there valuation or, probably money, to your point.

Jake Cooper:

Yeah, exactly. Right. And so there's actually a whole kind of mechanics surrounding this, in what is the optimal. Right? Even from a Employee Stock Option pool. When you grow it? When you not grow it? When you raise money? When you not raise money? All those other things. So, that's where we're at right now.

Turner Novak:

And talking to people, just before we did this, a couple of people said you have a philosophy around just hiring really opinionated, talented people, and maybe not specifically telling them what to do. It's just kind of like, unleashing them and just seeing what happens. And trying to guide it, I guess. I don't know if I'm explaining this right, but-

Jake Cooper:

Yeah.

Turner Novak:

So, what's the philosophy there?

Jake Cooper:

The philosophy is that nobody joins a startup to be told what to do. They join a startup to work on hard problems, and to go and push themselves to execute on hard problems. So from that kind of standpoint, we're really looking for people who are hyper autonomous. So we try and say, "Hey, we'll give you kind of a structured onboarding. It's like six weeks of here's how we're ramp you from having zero context about how the business works or anything else like that." I mean, maybe not zero, because you have some of it from external or whatever. But our operating procedure, we've written a little bit about it. But it's very, very different when you come and join the organization. So from nothing to, I want you to do things that are very, very important for the business. Like, go in and start, and figure out how are we going to do networking. Go in and figure out, maybe, data center procurement. Go in and... How can we rapidly, rapidly, rapidly get you to a point where you are building real things that you are really truly interested in?

Because I believe that we're going to get the best out of people when they're really, really interested in building something. So I try to ask them, "What thing are you interested in? How can we set you up for that immediately? And we're going to go and ladder your onboarding experience to get you to that point where you were going and making decisions." And so, this can be actually very uncomfortable for some people who have operated a company where they have a manager, and the manager tells them what to go in and do. And so, we've structured the process where you've two weeks of tasks, really straightforward stuff of like, "Go and make this change in the code base. This thing is evidently broken. Here's where it is." We do two weeks of problems, which is we have this problem, how are you going to go in and solve it? You'll have to put together a little document, it's called an RFC.

You just basically say, "Here's the problem, here's requirements for the problem, here's my suggested solution, here's some other stuff that I've got." And then the team will rally around and say, "Okay, maybe this, this, these things." And then, they'll move through that. And then the last two weeks is what we call opportunities, which is what do you think the company needs? So within one month, we've basically stopped telling you what to go in and do.

Turner Novak:

So this is your six-month onboarding process?

Jake Cooper:

Week. Six weeks.

Turner Novak:

Sure, yeah. Six-week onboarding process. It's specific tasks for the first two. It's problem, like, "Figure out how to solve this problem."

Jake Cooper:

Here's a problem. The problem is evident and we've written it down, and we need you to solve it. There's varieties of different ways you could solve it as a solution, et cetera.

Turner Novak:

It would just be like, we need to grow or we need to set up a data center, or we need just like-

Jake Cooper:

Yeah. Things like our observability, product sucks. How do we fix it? Like, these things. And so, they're nebulous because that's what you want in early stages. You want people who are going to basically take these nebulous things and convert them to tractable outcomes. And that's what we're aiming to, not test for, but ladder you up and basically say, "Hey, here's how you should be thinking a lot of these things." We are trying to move the business forward in either solving immediate problems or identifying opportunities that are going to allow us to grow really, really rapidly over time.

So getting you to that point of making those decisions, and identifying them, and making mistakes and all those other things, it's something that we try and teach people in general so that they can operate completely autonomously. Because otherwise, it's just the company is a series of executives ranging from managers to board people, telling people what problems are still like, and things to do. And again, I think the most awesome people, that's not really how they want to live their life in general. They don't want to just be told exactly what to go in and do. They actually want to go in and unleash themselves on, where could I have the most impact? People are looking for impact in general. So we can almost be a conduit for that in saying, "Okay, here's how you can actually ladder up to having really, really awesome impact on the organization. Shipping all the stuff that you've ever wanted to go, building out the thing." Whether it's your sales process or your go-to market, or product or design, or anything else like that. It is a blank canvas for you to go in and push forward.

Turner Novak:

So, it's more help making an environment where people solve problems versus complete tasks?

Jake Cooper:

Yes. Yep, exactly. Right. And so, that's part of the whole operating, I honestly think. And that's the thing that we really, really try and hammer home because we're, again, trying to innovate on some of the org structures of basically saying, how do you keep a company lean? And if you have people that you're consistently being like, "Hey, how's that TPS report or whatever?" You just need managers and people to close these loops, and then the people are pulling away because not that interesting or anything else like that.

Part of the reason I start Railway is it is a ocean of interesting problems to other businesses, but there's some businesses that are actually deeply boring. And I would far rather work on those interesting problems as an engineer, or in my spare time, just trying to design things or anything else like that. Those are the things that will keep my brain occupied. And so that's the space that we want to cultivate for people to come along and say, "Oh, that's really, really cool. I'd love to solve this thing or build this or do this thing." Whether it's networking, product, go-to market, anything support, stuff like that.

Turner Novak:

So, how would you gauge some things that are worthwhile problem to solve? When you're going through this, like, "We're using this to solve this problem, but not that problem." How do you figure out what's the intensity or severity, or mission criticalness, and crucial to solve of a problem versus not?

Jake Cooper:

So, we have a planning process and we're switching, essentially. We went from projects previously, or we just put up a bunch of projects and we're like, "Okay. Roughly, these will solve these problems and we can go execute on them." Shoot problems. And as part of that planning process, what we do is we have essentially co-signers. So you'll see a problem or you'll see a project previously, and you'll go ahead and you'll put your name down beside it. And then we'll essentially stack-rank and prioritize all of these things, and we will go together.

And for that one week that we're actually going to go and plan, we're going to get consensus on which problems are really, really important. Let's stack-rank of these problems. Who roughly might be a good fit to go and solve these things? And, how many do we think we can actually accomplish? And that just provides a list. From tip to tail, let's start on these things. Let's go down the list and just essentially see how far we can get on those problems.

Turner Novak:

So you have, it's like a product roadmap. It's like a problem roadmap.

Jake Cooper:

Kind of, right. But it's also fueled by, we want to go in and do these things from a product perspective. So if this is the vision, I issued this thing what I internally called guidance, which is it's like how the companies issue guidance. It's mostly on financials and whatever. But I issue it from a product perspective of like, this is what we want go in and build. And so if we have this expected state that we want to go in and get to, and we have our current state, the problem is is that this shit doesn't deploy fast enough, or all of those other problems that are along with it.

And so you can almost quantify those either problems or opportunities as like, we want to be here and we're here. And so for us to be here, we have to solve these problems.

Turner Novak:

Do you feel like there's anything that people miss there? Or just generally in culture setting, a lot of founders or startups get wrong?

Jake Cooper:

My hot take is I don't think you set your own culture. I think at best, you as a founder get to steer it and you get to clarify it. So you essentially say, you are going to hire some people. And at some point, you're going to look up and you're going to realize that the people that you've hired, mostly all actually have the same rough-ish kind of core. And that's your culture. And if you can figure out how to write down what those things are that make you you, and that make people... like, people who work at Railway, really, really successful or enjoy the work that they go and do, et cetera. That thing is your culture. And you as the founder are the steward for your own culture. And so, it's your job to identify it. It's your job to be really, really clear about it. It's your job to whenever you go and chat with candidates, to say, "Hey, this is what we are. And the entire subset of everything else, there's a thing that we are not right."

It's really, really important to quantify what you are and what you aren't, so that people have a really, really good shot. And I think this is the thing that I messed up really, really early was that I wasn't as clear or deliberate about these things, about what made us us, to candidates. And as a result, we hired people and we had to let them go at some point. And it was through no fault of their own. It was, we're not clear about, "This is our culture and this is how we operate," and all these other things. And so as a result, I've continued to issue blog posts after blog posts. Like, "Here's how we work. Here's how we think about problems. Here's how we operate."

Jumping on a podcast, talking about how we're trying to innovate and build on your structure to say, if you believe in these things and the same things that we believe, that infrastructure is stuck in the Stone Ages, and you can build beautiful design for this. This lost art of kernel, and data center, and all these other things. And organizations should be filled with people who solve problems and identify opportunities instead of receive tasks, then you may be a fit for Railway. Come on down.

Turner Novak:

Small plug.

Jake Cooper:

You've got to do that kind of work to quantify what that culture is. And so for us, it's like we have four things. We have honorable, accelerating, reliable, and direct.

Turner Novak:

Say those again? Honorable?

Jake Cooper:

Accelerating, reliable, and direct. It spells HARD, and it helps you solve hard problems. And so whenever you're trying to decide, do I shift this thing today or do I shift this thing tomorrow? You should be accelerating. Do I tell this person this feedback? I should be direct, right? Do I show up for this thing? Am I present? Am I putting myself in a situation where I'm on-call, but actually I've just left my laptop at home? I'd rather be reliable. And then honor does a lot of nebulous capturing of like, what is honorable? Try and figure it out generally.

Turner Novak:

That's a cool...

Cool. Did you come up with it? Or was it a team effort of like, "Let's figure out some kind of fortune cookie-type acronym thing"?

Jake Cooper:

So, I came up with it because I think most values are bullshit. And people go in and they flip them up there, and it's like, "What do these things mean?" At most companies, it's like, ask any employee the values and...

Turner Novak:

Integrity.

Jake Cooper:

Some shit like that, right? It's like transparency or anything ensure... Maybe those are good things, but I really, really wanted the values to have deep value for people to help them navigate those things. I don't know if it's a value or if it's a concept or whatever, but Uber has this thing called toe stepping. It was basically-

Turner Novak:

Like, stepping on someone's toes?

Jake Cooper:

Yeah, right. And look at the value that they had. They were like-

Turner Novak:

If you or don't do it.

Jake Cooper:

Do it.

Turner Novak:

Do it. Oh, wow.

Jake Cooper:

Do the toe stepping thing, right? I was like, what the fuck?

Turner Novak:

That's kind of rude.

Jake Cooper:

That's a scam to add value, et cetera. But then I realized that values, for lots of people, are actually almost like a lens through which you can navigate the culture a bit. And so you can think about it. And so if you think about Facebook's most famous value of, "Move fast and break things," it admonished anybody of any sort of, almost like path that they would incur because they had broken something. Because they could point to the culture and they could say, "But we move fast and we break things. That's what makes us Facebook."

And so if you have this kind of shared tenancy that you can all rally around, you can say, "Well, so-and-so wasn't a fit because they weren't moving fast and they weren't breaking things." And then you layer that with multiple values tenets there over time, and then you figure out what is you, right? What makes you you, as part of the company? The people that you bring on and the people that you attract, and all those other things. And so, you're almost cooking this cultural concentrate.

Turner Novak:

So how would you intentionally set that up? Because it sounds like you messed it up a little bit, or you didn't quite stick it and land it in the beginning. And it sounds like maybe, most people see it as just like a, you know what are our values like. Transparency, integrity. Maybe, it's...

How do you get the thing that actually is tangible and moves the needle, and then not make the same mistake you did?

Jake Cooper:

Man, I wish I could tell you other than you just have to trial and error it almost, right? And be really, really aware of what do you want to, I don't know, give people as a value? Like-

Turner Novak:

It seems like values is the things that's acceptable, that the company leans into as being its thing.

Jake Cooper:

Yeah.

Turner Novak:

That's your-

Okay, I see it's your point.

Jake Cooper:

I think it's a compass for how are you going to operate as a company. That's the way that I've looked at values in general. But again, I think it's one of those things where you don't really set it. You hire some amount of people, and at some point, you realize that that's your shared common denominator.

Turner Novak:

Angelo on your team mentioned that you wrote your way or tweeted your way into being a founder. What does that mean?

Jake Cooper:

So I think it's a lot of that kind of build-in public type of thing where I'm writing blog posts about the inward draw of capitalism or whatever. Just putting out things and feelers and all of those other things, you're almost like, "Was it..." I think it's Neval or whatever that's like, "You got to create the surface area for luck or whatever it looks like."

Turner Novak:

It sounds like a Neval quote.

Jake Cooper:

Yeah, right. Or something like that. So it was a lot of, basically, just how do I put myself in the arena? To go in and try and catch some of this luck in general? And for me, it was more just like a... I think that anything has to stand to public scrutiny or doesn't stand at all, right? And so if you put yourself in that arena, and you start going and putting those things out there, you're going to figure out really, really quickly whether you are correct or not on a lot of these things. Because in a public forum, especially on the internet or whatever, people are not shy to call you out on a lot of this stuff.

Turner Novak:

If you got an anonymous account that's nothing to lose, it's like, "You're an idiot."

Jake Cooper:

Oh, yeah. Some person with an anime avatar and six followers can ruin your entire week. It's just...

Turner Novak:

And so you wrote a post, and it was in May of 2021. I think you actually referred to it earlier. It was called How We Work. I think it was How We Work at Railway, or maybe it was just How We Work. What was the post... What was the general framing in how do you work at Railway?

Jake Cooper:

So this was the first version of that blog post of like, "Oh. Oh, no." We've messed up hiring and we had to let go of somebody, and we need to figure out a way to not do that because these are real people's lives, of course. So, you don't want to do that.

So I wrote a post called How We Work. We're a remote company, so we work with people like Dubai, Japan, Spain, Thailand, Montreal. Every time zone. Every time zone has hit. Australia. So we have to work in a different way, and this is part of the... We're innovating a little bit on the work structure because I think that that's really, really interesting. One of our investors pointed it out. He's like, "Not really a lot of big remote companies."

Turner Novak:

Maybe, GitLab is the biggest?

Jake Cooper:

Yeah, GitLab is the one exception in general to that rule. But generally, there isn't. Right? And we had a conversation about why that is in general. And there's a few nuggets of, basically, the standard org structure doesn't work as well from that perspective.

Turner Novak:

You're saying the management's like MBA-type org structure thinking?

Jake Cooper:

Yes, right. And the short reason for this is actually if somebody tells you something that you don't like in an office or whatever, you still have to show up to the office generally. You have to work, et cetera. You continue to have to show up to this job. If you're doing things that people don't like in general as a remote work employee, what do you do? You close the fucking laptop, you go outside. Basically, you can quiet quit or you can whatever. And so essentially, from an org design perspective, you have to design an org which is interesting enough for people to say, "Oh yeah. Actually, the reason I'm signing along, the reason I'm doing all of these things is actually because I'm really, really, really interested in the work."

And so we had to change a few of the structurings in that. But the How We Work blog post is the first pass, the very elementary kind of stuff of like, this is how we work differently than most companies to combat that problem, to make sure that people can continue to ship a lot of the things that they want to go. To attract talent that can work on really, really hard problems. To make sure that we don't have to get into that situation of like, "Hey, you have that TPS report in your..."

And the person says, "Yeah."

Like, "Great." And then goes to the laptop and goes... So-

Turner Novak:

... anything? We talked a little bit about how you've changed, say cadence, team design or structure related to how you work. Are there anything else you probably want to say related to how you work you haven't hit on yet? That kind of makes the structure of the team and the company a little bit unique?

Jake Cooper:

There's a bunch of different things, but I would recommend just in the interest of time, have a look at those four blog posts that we have in terms of how we work.

Turner Novak:

Okay. We'll throw a link in the description of the show. We'll get all the links and throw them down there for people. It feels like my general takeaway was minimize meta work and maximize real work. I think that seemed to be, succinctly, that's what you're trying to do particularly to productivity, like get stuff done.

Jake Cooper:

Yeah.

Turner Novak:

And specifically in a remote setting, you could create a lot of meta. You can always create a lot of meta work, like work to get work done. And you're trying to remove as much as this... The work that isn't actually real work and it's like adjacent work, you try to minimize how often that happens.

Jake Cooper:

Yeah, we try and do that. So, we do meetings on Mondays and Fridays because meetings are the number one contributor to a lot of that meta work. It's like, "Oh, how do we solve this problem?"

It's like, "Well, let's have a meeting. We met this problems." You don't solve the problem at the meeting. Even if you have a meeting to solve the problem, you don't actually solve the problem in the meeting because you didn't ratify that decision by pulling it out somewhere where everybody could have a look at it. And it would stand to scrutiny, et cetera. I don't know why, I assume Jeff Bezos has memos because it concretizes those things. He's that whole thing that clean memo, like messy meeting or whatever. But you really, really need to be concrete in the writing so that, again, it stands to scrutiny. That's the whole point of PowerPoints allows you to smear logic through the thing.

Turner Novak:

If you write it, yeah.

Jake Cooper:

If you write it, it has to stand to scrutiny. And so we try and minimize the amount of meta work. And it's like, "Put it together, identify the problem, identify the requirements, be very, very clear about it. Okay, cool. That's awesome. Let's move on that thing." And it sounds like a lot of work, but once you get actually reasonably good at the muscle, you can crank out these RFCs in five minutes to do a really, really complicated thing.

I wrote one, I don't know, this week, about some sort of database hardware offloading that we're going to go in and do, to make provisioning databases way quicker. And actually, database latency for right speed. Put it out in five minutes, had some comments on it. People were like, "What about this thing? Et cetera." I got alignment. I covered it. Somebody else pointed out a thing that I completely missed. I was like, "Oh crap, let's go and fix that." And then I was like, "Cool. Way to the races." All that happened. Guy in London just saw it randomly and I was like, "Boom. Okay, cool." And then I've worked on it and I'm going to go home after this podcast. I'm going to go and finish it up.

Turner Novak:

And is this all happening on Discord? Do you guys run the company on Discord? Which is not what you normally hear.

Jake Cooper:

No. Most people like running on Slack. And the reason we run it on Discord is drop-in communication. So remote companies don't have that office culture like serendipity, et cetera, which is actually deeply valuable. You want to actually be close to your coworkers in general. And so yes, you can spin up a huddle in Slack, but Discord has this really, really cool feature where you can see everybody in the meeting rooms right at the top.

Turner Novak:

Yeah, that's true. There's different audio jacks or, I guess, video. You can see who's in them.

Jake Cooper:

You can see. It's like the Michael Bloomberg Glassdoor meeting room policy for the internet of things or whatever. Not the internet of things, but the internet or whatever. So, that's why we do that in general. And it's worked quite well for a lot of things. And then, there's some things that we have to build Slack bridge extensions for.

Turner Novak:

Is there any... I mean, Discord is made for gaming, not running in a business organization.

Jake Cooper:

It's funny. We're trying to make a really concrete decision at the company and then you get a Fortnite ad at the bottom of it. You're like, "Okay, awesome."

Turner Novak:

Do you wish that more people, more founders, and more organizations talk publicly about how they do the team and structure things?

Jake Cooper:

Yeah, 100%. I think how you construct a company's work structure is... One, it hasn't been changed in a while. And two actually is one of the number one complaints that I hear from employees who have gone through hyper growth periods. I have friends at various different companies where they're like, "Man, we grew really fast and our management stack really sucks. We are not setting up our managers for success. We're not setting up people to scale into these roles. There's no quant." Almost like JD. It's like, what does a manager do, right? It's like, how do you give these people really, really awesome tools to be able to innovate on this thing that is very, very standard, MBA kind of, et cetera, business stuff.

So, I would love if people wrote on a lot of these things. I don't think a lot of people are doing that in general, not even just the writing thing. I don't even think they're just innovating on it. It's like a standard manager's path, which there's nothing wrong with that. Having a day... That's probably the closest rubric on concrete manager operationalizing. But I think that there is arb in changing a few things there, especially when you run a remote team and I can run one-on-ones at weird different hours. I can go out in the middle of the day, have a run, go have lunch or something like that, brunch with a friend or something, and then come back and have a meeting at, I don't know, 8:00 or 11 or something.

Turner Novak:

When you talk about, do you see the issues with these hyper growth startups that don't do org design, team design correctly? What's the mistake that seems to be made?

Jake Cooper:

It just seems like a, almost just like a lack of concrete operationalizing of what do managers do at those companies. What is a manager supposed to be responsible for? How can they encourage the success of their employees? Just a very, very concrete example of do they or do they not write code? That's the thing that is like, some companies have management stacks where they simply don't even write that, and it's a completely different experience across different culture or different teams on that culture. And so you have this almost shearing or different ways of operating, which there's not like, really, a standard on that other than industry standard, et cetera.

Turner Novak:

What's say the most interesting operating procedure policy that you've seen? Even if you don't use it at Railway, but it's the thing that pops the most is like, that's a very opinionated way of doing something.

Jake Cooper:

The one that I like is the clean memo-messy meeting or one from Amazon. That one just makes sense. We try and write memos generally for meetings, but it doesn't happen every single time, obviously. What other things are really, really solid operating procedure? I like a lot of the Netflix stuff. If you've read the No Rules Rules book, there's a lot of really, really solid stuff in there.

Turner Novak:

I don't think I have. What's an example from that?

Jake Cooper:

A lot of it is on just giving people as much slack or whatever as possible, a ton of different freedom. There's an example from that book where it's like, "Who approves my marketing expenses?" And it's like, you don't. You have a credit card and you go run the marketing expenses. And if you need to write a PO for any of these things, go and do it.

And they're like, "What's the budget? What's all these other things?" And it's like, make good decisions that are on the interest of the company, and go in and execute on those decisions. And if it's great, that's awesome. And if it's bad, tell us why it was bad. It's a lot of freedom, stuff like that.

Turner Novak:

You feel it's maybe too much like a daycare versus a science experiment?

Jake Cooper:

You words, not mine.

Turner Novak:

Okay. Yeah, that's fair. I'm happy to take the brunt on that one. I mean, when it comes to having a remote team, lots of time zones, lots of people doing different things, different procedures or different work modes, different cultures around the world.

One thing I know you have a really strong opinion on is not working on the weekends. So, why do you think that?

Jake Cooper:

I got lambasted on Twitter the other day for this one where I said... I forget what the actual quote was. I think it was my hot take is that you should not be working weekends. And if you end up working weekends, you suck at planning. And so, I personally try to both champion the idea that you should be on when you're on, and you should be off when you're on.

Turner Novak:

Off when you're off.

Jake Cooper:

When you're off. Yeah, sorry. And the reason for this is horsepower gets you a lot of the way, but at a certain point, you need to refine your technique. And if you figure out that you are working all the time, you don't really have a space to decompress, figure out where you essentially lost time. If you think about it in terms of almost like a Slalom skiing thing, I had this deadline and I was ahead of it. Why was I ahead of it? What decisions did I make that had actually allowed me to move really, really quickly there? Or I had this deadline and I was behind it. Why? For what reason was I behind? Where could I have bought back those seconds essentially?

And if you are always absolutely rip it and grip it and gunning it, all of these things, and your default modus operandi is to, since double-down and continue to work weekends, you're not going to be able to identify those areas in which you can actually get outsized value returns on simply changing one small aspect of how you are operating so that you can get faster and faster and faster. It's kind of like performance athletes or anything else like that. They'll push themselves a ton, and then they'll give themselves a little bit of space to cycle out.

Turner Novak:

So it's not like you don't work and you close your laptop on Friday at 5:00 PM, and you open it back up at 9:00 AM on Monday.

Jake Cooper:

No, no, no, no. If I do my job perfectly correctly 100% of the time, which I never ever do because I'm human or whatever, I will not work any weekends and I'll get paged when I'm on vacation. That doesn't happen. I worked the last six weekends getting a bunch of these things. But as a result of working those last six weekends, I've put things into a spot where I can say, "This is the reason I've been working the last six weekends." This class of work, we probably actually didn't even need to let go it down and do. We were just doing it because either org planning or reactionary or anything else like that. So I can go back through there and audit my time and say, "Where could we buy back 20% of our cycles? Where could we go in and do that?" And then we go in and write it down.

And creating that space to be able to go and do that, almost like time audit, decompression, et cetera, is really, really key for figuring out ways that you can actually go really, really quickly. Again, it's like you build the race car, you run the race car. You figure out where it rattles. You bring it into the pit lane. And then you go and you tune it again, and then you run the race car.

Turner Novak:

So, it's almost like you almost try to be a weekend warrior. You live for the weekend because you make your week so efficient and so good that you can take the weekend off. Even if you don't actually necessarily get to do that a lot, but that's what you strive for.

Jake Cooper:

Yeah, that's exactly the goal. And because my brain will naturally just get pulled back towards work anyways, but it's different kinds of work. It's work that's not like doing. It's work that's like thinking. So, I pull myself away from the keyboard. Maybe I'll take my journal, I'll go to the park or the sauna or something like that. It's still thinking about work, but it's almost like the technique of how could I have improved on this thing? How could I have whatever?

So that's why I think that you should delineate these concrete lines to say, "Okay, I'm going to go hard on all..." Monday to Friday, I'm going to go hard. I'm probably working until 11:00, 12:00. Sometimes 1:00 AM, and then I'm up at 8:00 AM. I'm going to go hard during the week. And then buzzer stops, weekend, et cetera. If I have stuff that I have to do because I'm getting pulled in, maybe I'm on call or anything else like that, obviously I have to go in and do it. But I should strive to basically try and disconnect as much as possible, and figure out why I'm getting pulled back into these things, where I lost time or anything else like that so that I can prove myself over time.

Turner Novak:

Do you feel like the culture of Silicon Valley has seeped into that being a necessary thing? Like, you must work weekends? Maybe not Silicon Valley, but just business in general.

Jake Cooper:

I think marketing is truly underrated. It's a skill that if you're really, really good at marketing, you can fuck up almost everything else. I think that marketing has only even gotten more and more important, and so there's a lot of almost gratuitous grind culture of, "I'm just going to grind a ton on this stuff." And because you're putting out that content, those are things that people are generally seeing, they're amping it, et cetera.

So I think it's changed a little bit in terms of people are doing things for almost like visual optics reasons in some ways, versus what can we do that's actually highly impactful or stuff like that. So, I think that that's more where it's gravitated towards. And as a result, you see a lot of, really, just absolute grind culture. Or really flashy things that you're like, "What is this? What product are you actually selling? What thing are you actually solving?" Stuff like that.

Turner Novak:

So one actual question, and we should probably should talk to this a long time ago, but Erica at Redpoint was like, "Oh, make sure you ask this." So, you probably could have taken a really big concentrated bet on AI models at some point. And I don't think you have. What was the thinking on that?

Jake Cooper:

I'm a big fan of Jim Collins. I don't know if you've ever read it.

Turner Novak:

Yeah. Good to Great, Great by Choice. What was the other one?

Jake Cooper:

Yep, I don't know. The one I think of mostly is Good to Great. And it has this principle in it called the Hedgehog Principle. And so it's essentially like, what is the surface area of things that you can be basically top three in the world, I think it is. And if you can't do that and you can't operate on those things, you probably shouldn't move into that adjacency zone just yet. And so, there's a shitload of competition in terms of doing the model hosting, et cetera. And so from this point, I don't think that we can be... We could if we invested all our eggs in that basket, but I don't think we can be top 3 to 5 in the world while also doing all the things that we want to go in and do. So it's almost like a concerted effort to say, "That's really, really cool, but we have to say no to this at this time."

We have to exercise a lot of focus to basically say we're going to execute on the things that we are really, really set up for and really, really do excellent. Well, let's go in and push on those. And then once we get to the point where we are dominating those things in that field, what can we expand to? And then figure out, "Okay, cool. When we can do compute, and data centers, and orchestration, and everything else like that so well that we are literally top in the world?" And we're pushing towards that as of right now, I think in a really good spot. Maybe we can do that for GPUs, right? And maybe we can slice and dice GPUs, and maybe we can give you timeshare on a basis of, "Oh, cool. I really want just this slice of a GPU for this period of time." And that may or may not be a thing that will be in the future.

Turner Novak:

Just like maybe a roadmap sneak peek, but maybe not?

Jake Cooper:

Maybe not. At some point, that will probably be a thing that is worth doing.

Turner Novak:

Okay, that's fair.

Jake Cooper:

This is not like a one drop. Like, next week we're doing GPUs or something else like that.

Turner Novak:

So you mentioned Good to Great by Jim Collins. Do you have any, just like favorite historical businesses or CEOs or founders that you've gotten a lot of inspiration from? And it can be recent, ancient times. Anything that you've... Anyone you've really drawn a lot of inspiration from?

Jake Cooper:

I don't really have founders or whatever. It's not like... Steve Jobs is great. He's a good person to look up for, for a variety of reasons. For some other reasons, not so great. It's kind of rude or whatever. I don't necessarily have a specific person where I'm like, "Oh my God, this is my idol or whatever." I take various different things from people and try to personify them.

One example is not even really a founder. There's this YouTube video of Jonah Hill talking about why Marcus Scorsese is such an excellent director. I swear I'm going somewhere with this. He's talking about it, and he's basically saying, "Imagine yourself doing..." As an experienced chess player with an infinite amount of time. And you make a move, and you keep going or whatever. They are really, really good because given a super, super small subset of time, they're going to make a better move quicker. And so, those are kind of things that I go in and say. I'm like, why is that person excellent? Why are they truly, truly good at their craft, and what makes them so? And is there a way that I can go and take that kind of skill and get really, really good at?

So I try and make decisions really, really quickly and intelligently now, because I believe that that is one part of being excellent. So that's just one example, but I take pieces from various different things I see them.

Turner Novak:

Did you try to find who or what is really good at specific things, and seeing what you can learn from?

Jake Cooper:

Yep.

Turner Novak:

Anything recent you've learned from that?

Jake Cooper:

The one that comes to mind is it's like Mark Zuckerberg. And it's not really like a thing, but I think Mark has really, really, really good patience, and really, really good identification of things that are important. And so, they are very happy to go in and work on the hard problem like VR, because they know they're going to shell a ton of money into that thing. Because they know at some point, VR is going to be really, really, really important. So, they have the patience to go and wait out for a ton, a ton of time so that they can potentially be first in this market and just wait around.

So that's one example of a thing that I've been like, I don't know if I learned it recently but I developed a deep appreciation for how good, I believe, that they are at this thing. It's waiting for the right time and then acting, and then being in a really, really good spot. You know?

Turner Novak:

Mm-hmm. It makes sense with thinking about maybe the Railway product roadmap where you rolled things out, what you've thought about, what you offer, where you compete.

Jake Cooper:

Just be really, really deliberate about that stuff. Wait for users to really tell you exactly what they want out of something. Because if somebody says, "Oh, I want this specific integration," and you're like, "All right, tell me why." Or like, "What problem?" Or anything else like that, it falls a bit flat. But then when a user begs you for it, like, "Oh my God, I need better observability. I can't tell any of those. I don't want to use this specific thing anymore. These are all the problems I need to solve. I need to be able to be notified when my apps crash, or I need to be able to know when latency spikes," or anything else like that. You're starting to get really, really concrete information because you've said no to this person a lot, in these people.

And you're all, for a long while, you're accumulating a lot of this information so that when you actually go and the time is right to go and act on that, you have a litany of feedback that you can actually go in and be like, "Okay, cool. These are all the things that we should go and execute on. These are the problems. These are the things that we're going to actually try and really push the envelope on." Your plan is really, really concrete. Instead of this almost like frenetic reactionary thing of like, "Oh, they want this, so I'm going to go and build this thing."

Turner Novak:

I feel like that's a good way to end it, with the way... It's almost like the whole Railway... The experiment of building this company has been a very intentional way for the right moments to do things. Even thinking about how you've incorporated certain products and features, just make sure you take a feedback and know when you're acting. Like, you act from position of strength almost.

Jake Cooper:

Well, this has been a lot of fun. Thanks for having me. This is excellent. This was a ton of fun.

Turner Novak:

Where if somebody's like, "Man, it's Jake, guys. Seems like a sharp guy." We mentioned your tweets. So, can we find you on Twitter or-

Jake Cooper:

Yeah, you can find me on Twitter or X, whatever you want to call it nowadays. It's Just Jake, so J-U-S-T and then Jake. You can follow me to varying levels of, I don't know, interest or whatever. I'm sorry or you're welcome, depending on how you read the tweets or whatever.

Turner Novak:

And then do you write on Substack, still? I didn't actually look at the dates on your post.

Jake Cooper:

I haven't written in a while. It's mostly kind of like... Actually, I might've written something a little while ago. It's very inconsistent. It's whenever the need strikes to get something out there in general.

Turner Novak:

You mentioned your million-dollar domain, railway.com. Is it where we can find out more?

Jake Cooper:

You can find us a railway.com.

Turner Novak:

Cool. Yeah, we'll throw links to a bunch of stuff we talked about in the description, and people can dig in a little bit more if they want to. Well, this has been a lot of fun. Thanks for doing it.

Jake Cooper:

Cheers. Thanks for having me.

---

Stream the full episode on YouTube, Spotify, or Apple.

Find transcripts of all other episodes here.