you2idea@video:~$ watch AcaBGRCzrjs [19:49]
// transcript — 292 segments
0:02 You can now hand a brand new developer a codebase you've built alone for months
0:06 or even years and get them up to speed fast. An AI agent reads the whole thing,
0:11 the history and the reasoning included, and hands over the context that used to
0:14 live in your head. That onboarding process used to take weeks of manually
0:19 walking someone through a codebase. So, one half of hiring your first developer
0:22 just got a lot easier. What hasn't changed is the other half, actually
0:26 picking the right person. get that wrong and there's no tool that saves you from
0:30 the wasted months and the wasted runway you just burned. So I sat down with
0:34 Derek Rhyr. Derek and I started Drip together, built it up and eventually
0:37 sold it. And along the way, we hired an entire engineering team. He's since gone
0:41 on to found Savvy. So between the two of us, we've hired a lot of developers and
0:45 learned a lot about what to scream for. Let's start with a question from
0:49 Michael. I founded an accounting source for Swiss freelancers four years ago and
0:53 I'm currently at around 24K MRR and I'm thinking about hiring my first developer
0:57 and I have so many questions around that. So I thought I'd ask you first of
1:01 all, what qualities would you look out for in the first developer you hire so
1:05 that that person can actually take things off my plate and free up my time
1:09 and hopefully take ownership of the codebase in the long run. Also, how do
1:13 you effectively onboard someone onto a codebase you've been the sole
1:17 contributor to? Would you try to create some sort of documentations or just dive
1:22 into it and try to teach them on the go? And also, do you have any other tips on
1:26 how I can derisk choosing the wrong person and wasting time and money and
1:31 making the onboarding process as smooth as possible so that the person can get
1:35 effective as soon as possible? >> So Derek, I'm going to kick this to you
1:38 first. He has several questions. So let's maybe tackle first. What qualities
1:43 would you look for in the first developer you hire so that person can
1:46 actually take things off your plate, free up your time, and hopefully take
1:50 ownership of the codebase in the long run? >> As I was thinking about this question, I
1:56 pulled up my most recent job description when I was hiring uh my latest full
2:00 stack dev who replaced my first developer for Savvy. So, kind of running
2:05 running through the list of things that I was listing in this cuz I sort of use
2:10 the job description as my spec for the things I'm going to be evaluating
2:13 against. You know, I guess first piece is like experience obviously like if
2:18 you're hiring somebody to kind of be in this technical founder handing off
2:23 control of the codebase or part of control of the codebase to somebody, you
2:26 want them to have a decent amount of experience. Obviously, budget
2:30 permitting. If you're super early stage, maybe you can't afford that. With Drip,
2:36 we ended up training developers um because we were budget constrained.
2:39 >> That's something I would recommend. Yeah, if you can avoid it, I would avoid
2:42 it. >> So, I was looking for someone with a minimum of 5 years experience as a
2:47 software developer just sort of as it's an arbitrary number, but you know,
2:51 senior level and someone with a full stack skill set. So this was like really
2:57 trying to hone in on someone who is a generalist who's comfortable
3:01 conceptualizing a feature from front end to back end and not somebody who's a
3:06 specialist. I think about like when we were later stage at Drip, we started
3:10 hiring, you know, backend engineers and JavaScript developers and these people
3:14 who were like hyper specific on their skill set and could really own, you
3:18 know, part of the stack, but that's going to work against you if you're
3:22 trying to like offload any and all development tasks to uh to a
3:28 number two. I put thrive working autonomously um because I think that's
3:33 interesting. my previous podcast co-host Ben Orinstein, founder of Tupil pair
3:38 programming app. A lot of uh a lot of developers enjoy pair programming
3:42 regularly. Some people are accustomed to doing it all the time or just, you know,
3:47 constantly collaborating on features. >> And so, you know, this is something like
3:52 you probably as the founder don't have time to be pair programming all the time
3:57 on stuff. And bringing in a a lone engineer, you know, not everyone is cut
4:03 out for that like flow of working. Like some people just really thrive on a team
4:07 and don't work so well independently. So, um, I think that's another important
4:11 one, assuming that you're just hiring one developer to start. Um,
4:14 >> for sure. >> And then confidence to ship. What I'm
4:18 trying to get at with this is sort of like a a a tamed perfectionism. Like,
4:22 you want someone who pays attention to the details and cares a lot about
4:27 quality, but also is willing to accept, you know, better done than perfect and
4:32 and just get stuff shipped. Because I think a lot of times on larger teams um
4:36 maybe where there's less of a sense of urgency or just more hands involved like
4:41 you can just spend weeks on one small task and making it perfect and making
4:46 sure it checks all the boxes and um when you're when you're tiny startup phase
4:50 you don't really have the luxury of being able to do that. So someone who's
4:53 very >> biased towards action. I think the other thing I would add, tell me tell me if
4:59 you think this is accurate still, but there's something to me about an a
5:04 general agreement on the the ethos of how you as the founder code. And what I
5:09 mean by this is back in the day, you may have been doing TDD, whether you're t
5:14 test first or whether you're we wanted tests in the codebase. There were folks
5:17 who were like, nah, don't believe in that. Don't want to do that. I would not
5:20 have hired them if they're not willing to do that. and and test is just one
5:24 example I'm bringing up, but there's a b there's tabs and spaces. No, there are a
5:28 bunch of things though, right? And some for some people, they're just super
5:32 flexible, right? And they are willing to kind of go, h you want tests, great,
5:36 I'll do it. You want tabs, great, I'll do it since that's, you know, the right
5:40 way to do it anyways. Um, you didn't you didn't have a reaction to
5:42 that. Is that not a thing anymore? Because I thought you were spaces.
5:44 [laughter] >> Oh, I am spaces. Yeah. Yeah. >> Yeah. I just said tabs. I just said tabs
5:49 is the right way to do it and you had no reaction. reaction. >> [laughter]
5:50 >> [laughter] >> you you're not even listening to me.
5:54 You're backgrounding me. >> Anyway, >> running along. [laughter]
5:57 >> Yeah, totally. I'm like, why? I'm waiting for the reaction. Um, but
6:00 whatever. And you know, tabs and spaces and tests are just examples of this, but
6:04 there's something I think about being in alignment and potentially having some
6:08 flexibility and willing to change. Like, let's be honest, if you'd hired a senior
6:12 first developer, you're working with them and you know, a cursor, Cloud Code,
6:17 Windsurf, like all these tools. tools. I know wind surf like gold news now, but
6:20 you know these tools came out. There are devs out there who who refuse to use
6:23 them. So what do you do then? Like do you want that person on your team? You
6:26 know there's some there's some type of flexibility and willingness to it's
6:30 having opinions because they should have opinions if you're a senior but being
6:33 willing to >> to go with the flow to let you make the final call to have whatever it there's a
6:38 collaborative nature of that. Do you feel like that's a like that's something
6:41 you look for? >> Yes, for sure. And I I think it's kind
6:45 of hard. It can be hard to pin that down exactly because it's like on the one
6:48 hand, you don't want someone who doesn't have any opinions or any taste about
6:53 things. But on the flip side, I've definitely spoken to developers before
6:58 who are so opinionated and so strong willed in their convictions that they're
7:03 almost hung up on that. So, like I think this is something that maybe would come
7:06 out even more if you were to do like a test project with someone, which I
7:10 highly recommend during a during a hiring process to like, >> okay, let's work on something together
7:15 for an afternoon or a few hours >> and let's see how you react to
7:19 encountering something that is maybe not done the way you would have done it, but
7:23 like this is the norm established in the codebase and are you willing to go along
7:29 with those norms and maybe you know recommend shifting pragmatically in a new direction if
7:35 there's something that is truly better. But like also recognizing like we can't
7:39 stop the world and rewrite the codebase or >> it doesn't make sense to fight against
7:44 the conventions that we've established in such a way where like now it's super
7:48 confusing anyone coming into the codebase is not able to you know to know
7:52 how to how to operate. So >> yeah that's a tricky one. What you don't
7:56 want to find yourself in the place of is like arguing small small things all the
8:02 time because opinions are too strong you know >> you know this you know the person we all
8:05 do we've worked with them not and some are developer a lot are developers some
8:09 are not you know there's people in other [laughter] parts of the company yeah and
8:14 then the other one I thought of that I do whether they're a developer or not
8:17 when hiring at even tiny seed micro comp these days it's almost a hard pass for
8:21 me if they've only worked at big companies >> if they've never worked at a let's say
8:29 sub 50 person company cuz tiny se Microsoft's 10 people right and in this
8:32 case we're talking about a two or three person company or whatever I mean it'd
8:35 be great if they work at like a 10 or 20 person company but man if someone's only
8:39 worked at say a fortune 500 or five you know 1000 or whatever I the the change
8:46 of mindset is so dramatic it's such a big deal that it is something I
8:50 absolutely scream for >> do you ask folks about that >> oh for sure Yeah, I similar criteria
8:56 where I'm just kind of not willing to entertain if someone exclusively has big
9:00 company experience like you said because yeah, I think dev teams just operate so
9:05 differently. I mean, even watching Drip grow to a team of what was it 20 20 plus
9:09 when we left >> and then 20 25 somewhere in there. Yeah.
9:13 Event that was just the engineer product team. >> Yeah. Yep. And so even watching how work
9:18 gets organized, how things, you know, the pace that people work at, it just it
9:22 morphs into a different thing, you know. Um, >> yeah, >> and that's not necessarily a bad
9:26 experience to have, but exclusively having that can be a culture shock. And
9:30 I think some people have a harder time adapting to that. And it it can be a lot
9:35 to reorient someone towards the small smaller company mindset and maybe you
9:38 don't have the luxury to do that if you're if you're hiring your first dev.
9:42 >> I remember hiring someone and and they got in the first day and they're like,
9:45 "So, what's the process for da da da?" And I was like, "There isn't. ask me,
9:50 you know, and what's the what's the uh do you have a form that I fill out cuz I
9:54 I want to get like a new mouse and a keyboard and I'm like, no, just buy it
9:57 on Amazon, tell me the total, I'll PayPal, I'll Venmo you, you know, out of
10:01 the company account or whatever or I can >> come to my computer, put it on the
10:04 thing, I'll do with the company card, you know, it's like that's that's the
10:07 process, you know. So, it's just different mindsets. And so, his next
10:12 question then was around he said, "How do you effectively onboard someone into
10:15 a code base you've been the sole contributor to? Do you try to create
10:19 documentation or do you dive in and try to teach them on the go? How have you
10:23 handled this? And I'm sure this has changed over the years with with AI and
10:26 all that stuff. >> Yeah. Yeah. I think this has changed
10:30 almost completely from my first iteration of hiring pre-AII versus post
10:37 AI. Um so I think before before AI like I I recall drafting some documents where
10:42 I was like giving a lay a high level lay of the land of the codebase like these
10:45 are the major subsystems. I don't think I tried to put too much in writing
10:49 though. It was more like on calls like let's together walk through the codebase
10:54 and let me you know show you kind of in the editor where things are and how they
10:58 fit together at a high level. But I think these days that is squarely in the
11:03 sweet spot of an LLM and I think it's just so much easier in general to get
11:06 onboarded into a codebase because you can literally fire up an agent and just
11:10 start asking questions about the codebase. I was going to ask that
11:12 because I haven't done I mean I've used claude code to do some little you know
11:16 pitly projects and stuff index all the podcast episodes blah blah blah but I
11:20 have not you can just ask it like what are the major subsystems and yeah of
11:24 course of course you can that is incredible wow >> it's [laughter] like we live in the
11:29 future all right and so that's it so at this point you'd be like
11:34 >> it's ask cloud code or ask your ask your coding tool
11:36 >> basically and you give them a you'd obviously give them an overview I mean
11:40 we used to pull someone in for an power on the whiteboard, right? And be like,
11:43 "Here's here's some boxes and here's some stuff. Here's the general
11:46 orientation. Here's the servers, whatever." >> Yeah. You know what? You're going to get
11:50 a kick out of this, too. It will, of course, read git history. So, it'll it
11:55 can look at a file and know like this was introduced by Derek in 2023. And
12:01 it'll read the the git, you know, commit comments and the surrounding, you know,
12:05 it'll look up the pull request that it was attached to and read the surrounding
12:09 tickets from around that time. So that its ability to gather context is just
12:13 insane. It'll say like yeah this was this was put in in 2023 as part of this
12:17 workstream to optimize >> uh this part of the you know slot
12:22 calculation engine cuz that's always the thing is we'd be like why does this code
12:26 operate this way and we're like nobody remembers you know where we'd have to go
12:29 back through and go through notes or wipe holy crap that's amazing.
12:33 >> Yeah. Yeah. [laughter] Well, remember remember when when you were starting to
12:37 code drip? I was like so we had a conversation and I was like um so you're
12:40 going to I don't know how we got here but it was like you're going to add code
12:44 comment you know comments you used to add comments in code people to like say
12:47 what it says and you're like I'm not going to do that because it's rail it's
12:51 Ruby and Rails and it's supposed to be written so simply that you can read it
12:53 and that comments go out of date and we had this whole conversation I was like
12:57 fine I believe you and with tests it's it's less brittle anyways now that would
13:02 be laughable to like write comments in your code there's no there's just no
13:06 reason to you know Yeah. Yeah. But it's funny too like even even these days just
13:12 writing a commit message by hand feels archaic because it's like these days
13:16 you're building with an agent generally the agent has so much context and the
13:20 agent can summarize so much better what we did and why we did it.
13:24 >> So it's it's crazy how much better the commit history has gotten about
13:27 capturing context. >> In a minute, Dererick and I get into how
13:30 to make sure you don't hire the wrong person and waste months of time and
13:34 runway. But first, hiring your first developer is a big milestone. It means
13:38 you've already cleared the riskiest part of the journey. You found an idea worth
13:41 building, and you got people to pay for it. But if you're not there yet, that
13:45 early part is exactly what my new book is about. It's called [music] Idea to
13:49 Traction. It's a step-by-step guide to finding an idea worth pursuing,
13:52 validating that people actually want it, and launching something they'll pay for.
13:56 Basically, how to stop building SAS nobody wants in the age of AI.
14:00 Pre-orders open in a few days. They might even be open now depending on when
14:04 you're watching this. Either way, head to idea toractionbook.com and get on the
14:09 wait list. You'll get my 9to5 audit right away that walks you through how to
14:13 find a 10K a month SAS idea at your day job. You'll also get an invite to a live
14:17 Q&A where I'm going to answer your early stage questions. Plus, you'll get access
14:20 to some other bonuses that won't be available after the book is published.
14:23 All right, let's get back to Michael's last question about making the
14:26 onboarding process as smooth as possible. Would you pair program like an
14:31 hour or two a day for the first week or two with like a new a new hire?
14:36 >> Yeah, I've always done that. I feel like it's um you know, especially early on,
14:41 that's your time to really infuse the the ethos of like what we do and how we
14:47 do it. Um and getting some of those particular patterns that we like to
14:52 follow when we do things infused into that person is really crucial early on,
14:57 I think. And some of that can come from an AI agent reading the codebase and dis
15:01 and trying to discern some of that. But also there's the element of like
15:05 building the rapport, building the here's how we work together on things,
15:09 here's how we disagree on things and how we resolve those disagreements. And so
15:12 there's a lot of that kind of human level collaboration value in spending
15:18 time together pairing. So yeah, I've always spent a good couple weeks, you
15:22 know, kind of each day for an hour or two or whatever. um pair programming
15:28 >> and then in terms of not choosing the wrong person I mean obviously this is
15:31 this is the big risk of hiring anybody this is tough the things there's a
15:34 couple things that come to mind like when I'm looking to hire a developer
15:38 specifically I think we have technical acumen we have interpersonal dynamics
15:44 and that includes do we get along the thing we said earlier are they flexible
15:47 in their thinking you know will they are they good team player there's a ton of
15:51 stuff you can wrap up into that one and then the third is Like I guess it's just
15:56 reliability. This is for any employee. Like are they do they ship stuff on
16:01 time? Do they do what they say they're going to do? And do they show up to work
16:06 and do they you know not call in sick all the time or whatever, right? There's
16:08 like this kind of consistency, reliability cuz I've had folks who are
16:12 great technically in different roles whether the dev or copyriter, marketer,
16:16 great technically, great interpersonally and they're super flaky. Yeah. Right. So
16:20 you need all three of these. So, the thing you can't test is the third one,
16:24 the flakiness. You just have to work with someone and whether that's for a
16:28 couple weeks on a probationary period or whatever, aside from asking someone when
16:31 they're going to tell you the right answer, you know, but the other two you
16:36 can test in conversations and in pair programming. So, when you and I started
16:40 hiring engineers for drip, remember it was go it was a take-home project. Now,
16:45 that is laughable because they will do it with AI, which they should, but like
16:50 it's just you're not going to do that. And so, we but we switched to pair
16:54 programming once we really started hiring engineers because we found it was
16:58 so much more productive, not just to see the results, but to see how they
17:01 thought, to see what it's like when they were challenged. You would challenge
17:03 them. Usually with one thing, you're like, "Oh, I'm a really good throw for a
17:06 loop here. I'm going to tell them to use suddenly use spaces instead of tabs or
17:10 whatever and see if they like completely go you know go postal or whatever. So
17:15 those are my thoughts like you know we we would have one or two conversations
17:18 usually two >> that was after an initial HR screening but then you and I think we have one or
17:25 two then we bring them in for we would bring them in in person because we were
17:27 hiring in Minneapolis but now you do it remote right with Tupel or something uh
17:32 for what was it a twoh hour was it usually two hours a pairing I don't
17:35 remember. Yeah, I think something like that. I ended up going maybe three hours
17:40 on my last one. It went really fast in my recollection. Like I thought we were
17:43 going to make so much more progress and it we didn't make a ton of progress on
17:45 what we were doing, but it was still a high value exercise. You know,
17:48 >> here's the thing with this. This is someone I'm basically like, yes, they've
17:52 said all the right things. They get I like this person. I think they're going
17:55 to be a good fit for the team. As long as they're a decoder, they're in. You
17:59 know what I mean? This is the last step for me before we hire someone. So, cuz
18:03 you might think two to three hours of pairing, wow, that's a long time. But
18:08 most of the time, not all actually, I remember there was one or two that you
18:10 paired with and you were just like, "No." And it was great. That saved us a
18:14 ton of time. Um, but most of the time by the time they got there, you know, we we
18:17 did wind up hiring them. >> Yeah. I will say, uh, using a recruiter
18:22 was super helpful on this last round of hiring. Um just cuz you know you you
18:27 ideally you're you're promoting this widely that you that you're hiring you
18:31 know to to get it in front of >> hopefully a good amount of people so the
18:35 right fit can emerge but then you just have so much uh so many applications to
18:40 wade through so much sifting and and initial screening >> with AI now AI generating applications
18:44 and all that. >> Yeah. And I often recommend remote first
18:50 recruiting. It's Dan and Ian's um recruiting service for I think it's flat
18:54 fee and or I know it's flat fee and a lot of tiny seat companies use that.
18:57 There are other options out there who I'm curious who you use.
18:59 >> I use them. Yep. >> Use them. Okay, great. And they were and
19:03 they hire so folks know they hire around the world. They hire developers. They
19:07 hire like any role at a SAS for sure. Customer success and sales and stuff and
19:10 they've helped uh and they hire marketers and freelancers and all kinds
19:14 of stuff. So remote first recruiting if you're interested. Thanks again to Derek
19:18 for joining me to help answer Michael's question. If you've got your own
19:22 question, head to robwalling.com/quests and we'll consider it for an upcoming
19:26 episode. Hiring your first developer means you've made it further than most
19:29 founders. You got real revenue and now you're building a team. But there's a
19:33 good chance your first hire shouldn't be a developer. In this video, I give a
19:36 talk that walks through how to think about who your next hire should be and
19:40 the different directions you can go depending on the skill set of the
19:44 founding team. Go watch that one next. If you found this video helpful, please
19:47 give it a like and a subscribe. Thanks
$

The Developer Red Flag Most Founders Miss

@RobWalling 19:49 7 chapters
[developer tools and coding][AI agents and automation]
// chapters
// description

Hiring your first developer just got easier in one way: AI can now read your whole codebase and get someone up to speed in days. But picking the right person hasn’t gotten any easier, and it’s still the part that can burn your runway. I sat down with Derrick Reimer to answer a listener’s question about hiring his first developer: what to screen for, how AI changes onboarding, why pair programming beats take-home projects, and how to avoid hiring the wrong person. 🔗 Links in the video: 📖 Preord

now: 0:00
// tags
[developer tools and coding][AI agents and automation]