1
00:00:03,120 --> 00:00:14,740
Robby (clip): These are problems way beyond Ruby on Rails or Laravel or any of this stuff. This is geopolitical issues. Humanity is having this moment.

2
00:00:18,080 --> 00:00:27,360
Jeremy & Jess (intro): Hey everybody, I'm Jeremy. And I'm Jess. And we're two developers, business owners, and friends exploring the world of independent business and Rails. Welcome to IndieRails.

3
00:00:30,160 --> 00:00:40,480
Jess: All right, everyone, welcome to another episode of IndieRails. Today we are recording live from day two at Rails World in Austin in 2026, right?

4
00:00:40,600 --> 00:00:41,100
Jeremy: Yeah.

5
00:00:41,240 --> 00:00:50,480
Jess: So we've been on a blur here. We've been going from talk to talk and just hitting so many people, but we're lucky enough to bring in a couple of guests.

6
00:00:50,740 --> 00:00:55,140
Jess: We have Matt Swanson of Arrows and we have Robby Russell of Planet Argon.

7
00:00:55,140 --> 00:00:55,740
Jeremy: Hey guys.

8
00:00:55,820 --> 00:00:56,580
Matt: Yeehaw.

9
00:00:56,760 --> 00:00:58,180
Robby: Hello. Yeehaw.

10
00:00:58,720 --> 00:01:01,940
Jeremy: Howdy, partners. I guess that's how we talk down here now. Yeah, I guess so. Is that what it is?

11
00:01:02,000 --> 00:01:09,600
Jess: Yeah. And thank you guys for joining us because I think, like, Jeremy and I both just talked about how our brains are just full of stuff.

12
00:01:10,000 --> 00:01:16,920
Jess: And I think we would, it would be a slow podcast if we were just sitting here ourselves trying to think about what to reflect on.

13
00:01:16,920 --> 00:01:23,220
Jess: So hopefully you guys can spur the juices and get us going with your insights and feedback as well.

14
00:01:23,440 --> 00:01:27,220
Jeremy: Helps to bring on people that are also podcasters, I think. Yes.

15
00:01:27,220 --> 00:01:27,700
Robby: Yeah.

16
00:01:27,700 --> 00:01:28,820
Jeremy: Other talkers.

17
00:01:29,020 --> 00:01:33,060
Robby: I'm used to asking people questions. Yeah. So give me a spare—

18
00:01:33,120 --> 00:01:36,080
Matt: All right. So thumbs up, thumbs middle, thumbs down.

19
00:01:36,160 --> 00:01:36,700
Jeremy: Uh-huh.

20
00:01:36,740 --> 00:01:37,840
Matt: David's keynote.

21
00:01:38,280 --> 00:01:40,660
Jeremy: What if we didn't go to David's keynote?

22
00:01:40,760 --> 00:01:43,100
Matt: Well, then you're going to have to take a guess.

23
00:01:43,100 --> 00:01:47,720
Robby: Okay. Thumbs middle. Did anyone not go to the keynote? I did not. You were not there?

24
00:01:48,800 --> 00:01:51,200
Jeremy: I was there. I was there. That was front row.

25
00:01:51,640 --> 00:01:52,440
Jess: Yeah, I was there.

26
00:01:53,160 --> 00:01:55,240
Jeremy: I'm going to go thumbs up. Yeah. Okay.

27
00:01:55,500 --> 00:02:03,480
Matt: I think a keynote should be energizing. It should be inspirational. I don't think it should be something that you don't have an opinion on.

28
00:02:03,780 --> 00:02:07,840
Matt: So whether you took good or bad from it, I think that did the job of a keynote.

29
00:02:08,180 --> 00:02:10,100
Jess: It definitely did the job of getting people talking.

30
00:02:10,260 --> 00:02:10,760
Robby: Yeah.

31
00:02:10,880 --> 00:02:12,180
Jess: Not just at Rails World.

32
00:02:12,300 --> 00:02:12,560
Robby: No.

33
00:02:12,560 --> 00:02:13,060
Jess: Yeah, yeah.

34
00:02:13,640 --> 00:02:25,460
Robby: I also give it a thumbs up. And I don't think the energy going into the conference, say even the day before, there was already the anxiety of AIs. It's already there.

35
00:02:25,660 --> 00:02:36,080
Robby: So I think some people were hoping for a relief valve. And that's not what he delivered. And so I think it's okay for him to come out and share his perspective.

36
00:02:36,080 --> 00:02:47,340
Robby: But he also comes at running a business in a very different way than most of us are living in. Most of us are not in an environment where we're rebuilding our apps every few years and starting from scratch over and over and over.

37
00:02:47,380 --> 00:02:58,460
Robby: Most of us are taking care of software that's been around for a while. So as a counterbalance, that's what most of us are doing. So that's what he's doing. And we can learn from that and watch him do things like that.

38
00:02:58,460 --> 00:03:00,820
Robby: But it doesn't mean we're all in the same boat.

39
00:03:00,820 --> 00:03:01,940
Jeremy: Yeah, totally.

40
00:03:02,040 --> 00:03:10,900
Matt: Well, I think the strength of Rails has always been that the framework itself is extracted from businesses and real projects.

41
00:03:11,100 --> 00:03:18,940
Jeremy: And what comes with that maybe is not always said is that those businesses were operating in a certain context.

42
00:03:19,100 --> 00:03:29,440
Matt: And so while it's great to get contributions from GitHub, if you don't run a business that's shaped or scaled like GitHub, you maybe don't have the same problems.

43
00:03:30,040 --> 00:03:36,240
Matt: And, you know, certainly things do scale down to smaller companies in smaller situations.

44
00:03:36,260 --> 00:03:39,300
Jeremy: But that's not always the case. And that's fine.

45
00:03:39,900 --> 00:03:45,520
Matt: You know, I think that's something to take away, especially this week is there's good to take away.

46
00:03:45,600 --> 00:03:55,580
Matt: But you have to apply it to your own business, your own team, your own individual sense of what my day-to-day work is.

47
00:03:55,580 --> 00:04:04,760
Matt: And I don't know, there's like an interesting parallel probably because the thing that I think people get excited about with agents is that they have the context of your code base, your business a little bit.

48
00:04:04,920 --> 00:04:16,420
Matt: So in a way, I think like more people should be probably thinking about thinking about that and maybe that is more of the future of some software is like, you know, a great example I think is Mike's CVE skill.

49
00:04:16,420 --> 00:04:25,740
Matt: It was you take that and put it into your context of your project and you get a great result versus him trying to say, well, this is what we did at Basecamp to analyze this incident.

50
00:04:25,820 --> 00:04:36,720
Matt: It's like, well, I don't have my servers on AWS, so before he kind of like couldn't do that. So that to me is kind of like the vibe I'm taking away from it.

51
00:04:36,980 --> 00:04:41,620
Jeremy: Anything that you think you'll apply back at Arrows from so far that you've seen?

52
00:04:41,740 --> 00:04:52,580
Matt: I think I was not terribly surprised by the direction that things were going because I think I was sort of feeling that myself. Yeah. You know, it started the summer and so I don't know.

53
00:04:52,600 --> 00:05:00,800
Matt: I mean, I think we've been more willing to reevaluate decisions that we made about the tools we're using. We're still, you know, mainly building on Rails.

54
00:05:00,800 --> 00:05:12,640
Matt: I don't know that I'm going home and immediately rewriting the back end or anything, but I think for us it's been more like we're more open to being more ambitious than we were before. We have a team of two people.

55
00:05:12,640 --> 00:05:23,000
Matt: And so before that meant tradeoffs like we're going to use Hotwire because we don't want a two language tool chain. But now it's like, well, we're more willing to do that.

56
00:05:23,240 --> 00:05:26,240
Matt: So you kind of have to reevaluate some of the stuff.

57
00:05:26,240 --> 00:05:27,800
Robby: That sounds healthy. Yeah.

58
00:05:28,000 --> 00:05:39,420
Jeremy: Yeah. Yeah. I wasn't even thinking about that. Like I even when they were doing the CFP and I submitted a Hotwire, you know, talk and then it got accepted in the intervening months.

59
00:05:39,860 --> 00:05:49,380
Jeremy: It feels less and less relevant in some ways. I still like Hotwire in some ways, but it does feel like, OK, was that worth doing?

60
00:05:49,500 --> 00:05:57,040
Matt: It's hard. Like, what is the shape of conferences? It's like, I'm sure there's still value in getting together with people talking about practices and things that have worked for you.

61
00:05:57,040 --> 00:06:05,840
Matt: So maybe it just maybe it just shifts and, you know, it's not so much that everything is wrapped up with a language or a tool as the glue that's holding people together.

62
00:06:06,080 --> 00:06:11,580
Matt: It is more of practices or some different identity or something like that.

63
00:06:11,580 --> 00:06:20,940
Jeremy: So for me, it's the friends I made along the way. Exactly. Right. How about you, Robby? Well, actually, before we get into that, I really liked your keynote.

64
00:06:20,940 --> 00:06:22,140
Robby: It was great. Thank you.

65
00:06:22,140 --> 00:06:25,520
Jeremy: I appreciate that. Yeah. It's a good thing to hear.

66
00:06:26,440 --> 00:06:31,360
Jess: I'm curious how long. For the listener, do you want to try to get a TLDR on it?

67
00:06:31,400 --> 00:06:32,360
Robby: Yeah. OK, TLDR.

68
00:06:32,360 --> 00:06:35,600
Jeremy: Tell me more about my talk. Let me try to do that.

69
00:06:38,040 --> 00:06:49,640
Jeremy: Robby, yeah, I feel like I resonate with where you're coming from a lot just because I'm thinking in terms of consulting a lot. And I know that's what you do with Planet Argon. You've been in that space a long, long time.

70
00:06:49,720 --> 00:07:02,340
Jeremy: And you kind of set the stage of thinking about people's workshop and like the environment that we're in, which we're building and how important that is to what comes out of out of the workshop.

71
00:07:02,840 --> 00:07:13,220
Jeremy: And now it's not just the workshop designed for humans, but also for these agents. And I guess I heard a lot of things that were about what was good for often for people.

72
00:07:14,180 --> 00:07:27,940
Jeremy: Sometimes now like there's even more incentive or like more like alignment with the business to say, oh, you know, what's good for, you know, token usage is good for, you know, for us.

73
00:07:27,940 --> 00:07:40,220
Jeremy: And so some in some ways, I don't know, I guess the whole concept of maintainable software, like I felt that coming through from from you, but also applied and it's going to be a new era that that still matters.

74
00:07:40,560 --> 00:07:45,020
Jeremy: So I think it does. I still care about that and it was good to hear.

75
00:07:45,020 --> 00:07:54,580
Robby: Thank you. Yeah, I think the, you know, my spending my good portion of my summer thinking about this and what does it mean to show up and do work?

76
00:07:54,580 --> 00:08:05,120
Robby: And I think we often find ourselves looking at metrics and engineering teams that are related to like number of issues closed, bug fix, you know, how often your PR is getting merged, deployments.

77
00:08:06,040 --> 00:08:16,940
Robby: Yeah, I think these kind of, and there's, those are all metrics that are really kind of easy to, to measure, which is why they get measured. And that's how we compare ourselves and, or see how much more productive our team is.

78
00:08:17,380 --> 00:08:25,160
Robby: But when you talk with a lot of teams that are struggling right now because their PRs are getting bigger and messier and everybody's afraid like to ship things.

79
00:08:25,160 --> 00:08:35,620
Robby: And I'm like, well, is it the code that's the thing that you're afraid to merge? Or is it that your production environment won't protect you from mistakes?

80
00:08:35,680 --> 00:08:39,400
Robby: And how do we, where do we need to be thinking more?

81
00:08:39,560 --> 00:08:50,460
Robby: Like does having, say, an agent help out with shipping stuff or regardless of agents or other human beings shipping more stuff, if you add a bunch more engineers all of a sudden, you lose a little bit of context.

82
00:08:50,600 --> 00:09:02,320
Robby: Yes, but then how do you just make sure that you're focusing on making sure that your overall platform, your workshop is, uh, as safe to make mistakes in. Right. And cause you're going to make mistakes. Yeah. Agents makes mistakes. We make mistakes.

83
00:09:02,480 --> 00:09:14,240
Robby: We just, we have a sense of feeling accountable to the work that we're shipping. And so that's why we're, people are nervous about like, well, if I approve this PR and I didn't really understand it, that's a problem.

84
00:09:14,780 --> 00:09:23,100
Robby: Cause I'm on the hook for it still, that might be arguable because maybe agents are allowed to apparently just hack Ruby gems or whatever. And nobody seems to be getting in trouble for that.

85
00:09:23,100 --> 00:09:31,660
Jeremy: I don't understand. It does feel like there's a re, uh, like a renegotiation around the, um, accountability for developing software.

86
00:09:31,660 --> 00:09:32,900
Robby: And I don't know what that means.

87
00:09:33,260 --> 00:09:39,500
Jeremy: Yeah. Like, uh, if Claude built this and I can go to the business owner and say, shrug my shoulders and be like, I'm sorry. Like Claude did that.

88
00:09:39,640 --> 00:09:40,320
Robby: Like, that's not me.

89
00:09:40,320 --> 00:09:49,920
Jeremy: I didn't write that. Yeah. Uh, so sorry, but it's not my fault. Yeah. In some ways it feels a little bit like that. And almost like the business owners are saying, yeah, that's okay.

90
00:09:50,440 --> 00:09:59,820
Robby: As long as we go faster and it's cheaper. Yeah. I think that's, I think that's accurate. I, well, they, until something goes wrong. Right. And they're like, let's make sure this never happens again. Right.

91
00:09:59,860 --> 00:10:05,260
Robby: And so, and then everybody gets paranoid and then you start adding all these checkpoints and slows down your pipeline again.

92
00:10:05,340 --> 00:10:14,120
Robby: So I, uh, think about, so anyways, my, a big part of my talk was talking about, yeah, like the workshop and we want to encourage people to like spend some time.

93
00:10:14,280 --> 00:10:20,160
Robby: I'm improving their workshops and building tools that, that your agents can use, that your coworkers can use.

94
00:10:20,340 --> 00:10:35,820
Robby: Just, I think there's, we, I think if, if we're able to hand off a lot of these other types of things, whether or not you enjoy handwriting code or not, write features or work on bug fixes, do you have a little bit more capacity to spend more of that time focusing on things that you're curious about improving for yourself or for your coworkers?

95
00:10:35,960 --> 00:10:45,100
Robby: And I think if you can channel some of that anxiety into something productive, I think that's actually maybe a healthy thing. And I think that will make you more valuable for that organization. Yeah.

96
00:10:45,100 --> 00:10:56,240
Robby: Um, if you're the one doing that stuff and a lot of organizations don't have a dedicated platform team to focus on those types of things. So who does that when you're a team of three people or 20 people and, you know, maybe Matt can speak to that.

97
00:10:56,300 --> 00:11:05,820
Robby: Like how do you think about that? Do you, you're like, as you said, you were able to spend more time maybe even using different tech stacks that we wouldn't normally ever touch because like, I don't actually need to know. Yeah. I know how it's, I know it's running.

98
00:11:05,820 --> 00:11:14,040
Robby: We've got some analytics here or something, but like some people are still like nervous or they don't feel like they have the authority to just start doing stuff.

99
00:11:14,620 --> 00:11:18,840
Robby: And I don't know, I, maybe cause I am also a business owner. I have that perspective.

100
00:11:18,840 --> 00:11:20,060
Jeremy: I'm like, just fucking do it.

101
00:11:20,140 --> 00:11:22,740
Robby: Yeah. Just do it. Just, just make your job easier.

102
00:11:22,760 --> 00:11:31,520
Matt: Make, make everybody else's job easier. I like to think, I think I like the motto that's like, if something is painful, you should do it more. And like, that's where, that's where growth is. Right.

103
00:11:31,700 --> 00:11:43,280
Matt: And so it's, you know, in that is like, yeah, something is painful. Right. It, you know, but you have to push through or like experience the pain. I know Jeremy was talking a lot in his Hotwire talk about like, you have to sample.

104
00:11:44,040 --> 00:11:49,420
Matt: The hot sauce. Take a bite and feel the burn. Yeah. Feel the burn. Right. Yeah. And it's the same. I think it's the same for a lot of this stuff.

105
00:11:49,420 --> 00:12:00,100
Matt: Like on our team, uh, we have a designer and I've been pushing and maybe pushing too aggressively, but to say like, you've designed this, like we, we, we were all on board with it.

106
00:12:00,100 --> 00:12:03,500
Matt: Like you should go just implement it and even not just implement it.

107
00:12:03,500 --> 00:12:05,000
Robby: Like you should just push it.

108
00:12:05,080 --> 00:12:14,600
Matt: Yeah. Nice. Because I think it's natural for, for a designer who even is willing and eager to implement, to just say like, Oh, I implemented it.

109
00:12:14,600 --> 00:12:26,080
Matt: But like, I still want the engineers to sign off on it. And I think it's our job as the engineers to, uh, say, no, you push it. And if something bad happens, we need to fix whatever bad, right?

110
00:12:26,080 --> 00:12:35,180
Matt: Is it, is it, Oh, it got deployed and it, it blew up, had weird performance characteristics because you didn't understand the database querying stuff.

111
00:12:35,200 --> 00:12:46,120
Matt: And it's like, okay, well that's something that we can solve and move the knowledge out of the engineering team into like the product or the, you know, the workshop, the factory, whatever metaphor you want, you want to use.

112
00:12:46,380 --> 00:12:58,260
Matt: And to do that is like hard, right? Because you have to admit that something will go bad. Yeah. I've struggled myself with this of like, I, for a long time in my career, like I would feel bad if I shipped a bug, right?

113
00:12:58,260 --> 00:13:05,840
Matt: Because it's like, I'm tying my identity to this, yeah, to this job and this thing. And it's like, Oh, I did bad. Like I am bad. Right.

114
00:13:05,840 --> 00:13:12,200
Matt: And yeah, like that's obviously, you know, it can sound silly if you zoom out or see someone else say that like, Oh, I shipped this bug.

115
00:13:12,200 --> 00:13:13,100
Jeremy: Like I'm such a bad person.

116
00:13:13,100 --> 00:13:25,080
Matt: Like, what are you talking about? It's just, it's just the code. But I think getting over that and realizing that like, instead of, you know, feeling bad about that or feeling anxious or like, this is going to be painful to not review code or not look at code.

117
00:13:25,080 --> 00:13:33,820
Matt: It's like, you'll feel the pain and then you can, you know, use your engineering skills to fix that. And then you're going to end up in a, in a hopefully better, better spot.

118
00:13:34,420 --> 00:13:41,100
Jess: Yeah. Like the talking about the pain or the annoyance, right? Um, you know, cause agents are not going to remember things.

119
00:13:41,520 --> 00:13:59,620
Jess: So if, if you have to tell it over and over again, the same thing, like how to fix something that blew up or just how to remember to do something correctly, then hopefully then I think this is part of your topic was to document those things so that it doesn't happen again and that the agents can see it and take the right path.

120
00:13:59,760 --> 00:14:07,640
Jess: Um, another thing I was going to ask is tell me the title of your talk again. The most valuable engineer is not shipping features or something like that.

121
00:14:08,060 --> 00:14:18,560
Jess: So when I first heard that title, I assumed that, um, it was going to be about, uh, engineers being like more managers, like more product engineers.

122
00:14:18,780 --> 00:14:23,680
Jess: And cause that's kind of partly of where my personal role has gone over the past year or so.

123
00:14:23,680 --> 00:14:31,900
Jess: So, but I was surprised to see, and it made a lot of sense that it was not that you weren't working on the app.

124
00:14:31,900 --> 00:14:41,040
Jess: It, it was that instead of pushing the glamorous feature, you were, you were fixing the thing that enabled people to continue to build and, and make features.

125
00:14:41,420 --> 00:14:53,120
Robby: And that whole perspective is just thinking about like, Hey, don't wait to get promoted to be responsible for these things. Like I think a lot of companies, like there's not a role specifically for that.

126
00:14:53,120 --> 00:15:02,000
Robby: So just become that person that's making, that's helping your coworkers or your agents ship features more smoothly. And, and I don't think we have the answer.

127
00:15:02,220 --> 00:15:10,140
Robby: I knew that when I was preparing this talk, I could not give like a very prescriptive answer. Like, here's how you organize your things for Claude Code to do this the right way. Cause the stuff is moving under our feet.

128
00:15:10,400 --> 00:15:22,820
Robby: So I think those are the problems we need to like spend some time thinking about and look a little bit, I can't say, is that more downstream as you think about your environment or what can we learn from these bigger companies?

129
00:15:22,820 --> 00:15:33,380
Robby: And like, how are they, it's, it's great that Fin is able to have 40% of their PRs get automatically approved and just shipped without a human reviewing them. Most of us aren't there yet. So what could we be?

130
00:15:33,380 --> 00:15:44,720
Robby: And like, I think that's a good question to be asking ourselves. And I'm like, what, what we're having the conversation with our own team. I'm like, why am I reviewing this PR that your Claude wrote with my Claude at some point?

131
00:15:44,720 --> 00:15:47,420
Jess: I'm like, you're not even using different LLMs.

132
00:15:47,420 --> 00:15:58,500
Robby: No. I'm like, what we're like, why do we do PRs the way we do? I heard one person talk, uh, how they were thinking about PRs with including details. To assign to the bots.

133
00:15:58,500 --> 00:16:07,300
Robby: So like after your agent deploys something that it then also has the instructions on how to validate or verify what said, like what is supposed to happen in production.

134
00:16:07,300 --> 00:16:16,980
Robby: So it actually had followup tasks to verify. And if not, then like try to roll back or something. And I'm like, we need, we need to have some interesting creative ideas.

135
00:16:16,980 --> 00:16:23,560
Robby: Cause I don't think the tooling is there to support that quite yet. And there's no frameworks or platforms that are like, use that one specific thing.

136
00:16:23,920 --> 00:16:30,280
Robby: And I think there's a lot of people that are really good at optimizing for one person knowing it like, this is my factory. Yeah.

137
00:16:30,280 --> 00:16:36,020
Robby: But when most of us are playing team sports and we're like, we're figuring this stuff out.

138
00:16:36,040 --> 00:16:44,220
Matt: Yeah. I think some of the anxiety is around like, oh, there's no practices. Everything's out the window. I don't know what to do. And I don't, I don't know that that's actually true.

139
00:16:44,380 --> 00:16:54,540
Matt: I think a lot of times you can look at what we had in the past as a practice and you can almost even just one-to-one instead of a coworker was supposed to do this.

140
00:16:54,540 --> 00:17:05,860
Matt: It's like an agent was supposed to do this. And you know, if that gives you some stability or some sense of the thing, like, I think that is an interesting path for people that are feeling anxious or like, I don't know what the process is.

141
00:17:05,860 --> 00:17:08,300
Matt: It's almost like, well, you can just still do the process.

142
00:17:08,500 --> 00:17:26,940
Matt: I think the difference is that an agent is like actually the most obedient, the most reliable of a coworker in, in some sense, like if you can figure out what, if whatever tooling happens to be of the month or whatever, something that says when you submit code, like you need to test it in all browsers, you need to test it on mobile.

143
00:17:26,940 --> 00:17:39,420
Matt: You need to record a video of you showing me that you clicked through, like that is going to be better than a regular teammate, maybe not on their best day as a teammate, but certainly on their, on their median day.

144
00:17:39,460 --> 00:17:50,920
Matt: Right. And like, I'm guilty of this too. Like there's times when, you know, we've all written a feature. Code looks good. And then you realize you've shipped it and you're like, oh, I didn't actually go in and continually test this. Right.

145
00:17:51,100 --> 00:17:53,540
Matt: Even if I wrote unit tests, even if I wrote system tests and all that.

146
00:17:53,540 --> 00:18:06,440
Matt: And so I think that's like an interesting foothold for people that are, that are struggling with some of this is like, you don't have to necessarily throw out all your practices, but like maybe there's changes you need to make to allow an agent to just follow your practice.

147
00:18:06,740 --> 00:18:12,880
Jeremy: So yeah. Yeah. I think it's hard to, I was thinking about kind of what you were saying about taking responsibility.

148
00:18:14,220 --> 00:18:39,960
Jeremy: I think early in my career, I realized that the key for my growth was to take more and more responsibility of more than my bosses were telling me to, to go ahead and like, you know, document in every incidents that we had in production, but even when they didn't ask, even when they didn't know to, to do that or realize that the site had been down or something like that, but I wouldn't, I'm going to do this practice.

149
00:18:40,280 --> 00:18:49,800
Jeremy: And then that made me better. And so everything that I've been pushing for and been rewarded by for decades has been take more responsibility, do more, take on more.

150
00:18:49,800 --> 00:18:59,960
Jeremy: And now this feels, it feels like a little bit like, uh, now it'd be less responsible. And I know you're saying, don't, that's not what it means, but it, it's what it feels like.

151
00:19:00,000 --> 00:19:07,580
Jeremy: It's like, you know what, throw it all out. Who cares anymore? You know, like I can just blame the bots, you know, I can just like tell the clankers to do this stuff.

152
00:19:07,620 --> 00:19:16,120
Jeremy: I don't have to look at it and it means like all the things that I've been rewarded around for years and everything that made me feel good about what I was doing goes away.

153
00:19:16,260 --> 00:19:18,100
Matt: And what, what's the point?

154
00:19:18,260 --> 00:19:21,660
Jeremy: And change on a dime, right? Yeah, right. Yeah. Yeah.

155
00:19:21,660 --> 00:19:33,380
Jess: Overnight, one year, let's change it all around. Well, you can do that if you're building a personal calculator tool, but maybe not if you have a 10 year old production app that customers that pay you hundreds of thousands a year.

156
00:19:33,580 --> 00:19:35,060
Robby: That's true. You know? Yeah.

157
00:19:35,340 --> 00:19:40,180
Jeremy: They don't even know what Rust is. That's like a thing, right?

158
00:19:40,180 --> 00:19:51,420
Matt: They know what Rust is. They know what Rust is, but yeah. So I think people are imagining that now, you know, we have to change so quickly or like there's so much turmoil. And is it, is it just about the rate of change?

159
00:19:51,680 --> 00:20:03,540
Matt: Because we've always had the issue of tech that we've always had the legacy app that is not on the latest Rails. And, you know, but it felt like people were not coming to Rails World, seeing a new Rails feature, going back.

160
00:20:03,580 --> 00:20:14,460
Matt: My Rails app is three versions behind and having like an existential crisis. It was more just like, this is the reality. And, you know, maybe that's the spark you need to, you know, start pushing that boulder up the hill.

161
00:20:14,620 --> 00:20:26,300
Matt: So I'm just curious if anyone, is that, is that really the, the difference is just the speed or are we, is there something other, other than that, that is causing the more existential questions?

162
00:20:26,400 --> 00:20:33,480
Jeremy: Well, there is a speed at which like humans can't survive. Like there are speeds you pass out, you know, uh, physically.

163
00:20:34,580 --> 00:20:41,220
Jeremy: And I feel like, uh, that's like, yes, it can be speed, but after a certain point, it's not linear, right? It's like, yeah.

164
00:20:41,220 --> 00:20:55,940
Jeremy: And the effects on a, on the, are just the cognitive load and the difficulty around managing that becomes, so it might just be, but I think there is a threshold to handle, uh, you know, a rate of change to handle it after which, uh, becomes just untenable.

165
00:20:56,500 --> 00:21:07,640
Matt: So I don't know if, if there is something else other than that, but I get, I hear what you're saying. I do hear what you're saying. I'm curious, Robby, because a lot of your work is legacy systems and upgrades and things like that of what, what your take would be.

166
00:21:08,200 --> 00:21:13,000
Robby: The, I'm trying to, I don't know that I fully understand your question, Matt.

167
00:21:13,100 --> 00:21:24,300
Robby: Uh, and so I, I don't think I can answer it to the fidelity that I would like to, I would say just with like the existential dread or that the feeling that people have, I'm like, this isn't a programming problem.

168
00:21:24,440 --> 00:21:33,600
Robby: Like this is way beyond, I think if we're in some ways, I think we're like navel gazing a little too much and be like, this isn't just us. Like this is a lot of people in a lot of careers and a lot of industries.

169
00:21:33,600 --> 00:21:42,880
Robby: And like, they're like, am I going to have a job in customer support? You know? And it's like, do I have to start driving a Uber? Maybe that's how you, and then we're in Austin and they're like, that's not a thing.

170
00:21:42,880 --> 00:21:50,740
Robby: And so just like all these like kind of fail safe things, like, well, in a, in a pinch, I'll start delivering food. And they're like, well, there's going to be a bot for that.

171
00:21:50,760 --> 00:22:02,440
Robby: And so it's like, yeah, I think this is the existential thing is more on a macro level, way beyond just like whatever our, you know, benevolent, uh, Rails person, you know, has to say on his top.

172
00:22:02,440 --> 00:22:11,180
Robby: I mean, like, I think people are just looking for something to, we, we feel like a interesting connection to the people that we work with and have a little bit of a, like why I met these people at this conference.

173
00:22:11,380 --> 00:22:25,180
Robby: I've seen them speak so I can feel a certain way about how they are talking about change in our industry and I can reject that because it feels, but like, these are problems way beyond Ruby on Rails or Laravel or any of this stuff.

174
00:22:25,200 --> 00:22:34,580
Robby: This is, this is geopolitical issues. Humanity is having this moment. And so I don't know how to, I don't know how to have that conversation. So I can be like, where, where am I?

175
00:22:34,660 --> 00:22:42,460
Robby: I'm working on these apps that I know clients are paying for me right now to, to have my team work on them and they're just like, Hey, this is what we're working on over the next three to six months.

176
00:22:42,460 --> 00:22:53,520
Robby: Like they're, those companies are having their own existential crisis of like, where's, what's their company going to look like? How do they stay in business? So we're those kind of like, we have a mutual shared, let's not all fail together. Yeah.

177
00:22:53,520 --> 00:23:05,320
Robby: And then of course we can get into the economics of AI companies and our retirement programs and plan— pensions all getting attached to them. And so we're also have to see these things be successful. Otherwise we're screwed. Yeah.

178
00:23:05,900 --> 00:23:17,180
Robby: We didn't sign up for that necessarily, but we also didn't, I don't know. It's a weird, I didn't, I don't know. Yeah. So I can spin out if I spend that. I'm like, all right, well, I guess I'll try to lean in on the optimism a little bit.

179
00:23:17,180 --> 00:23:21,620
Robby: I'm like, well, we saw, we still got our workshops to take care of, workshops to take care of.

180
00:23:21,940 --> 00:23:29,240
Robby: Let's just work on some small improvements here and there and just keep learning and having conversations and try to remember that we're a bunch of humans trying to figure this stuff out.

181
00:23:29,280 --> 00:23:29,780
Jeremy: Yeah.

182
00:23:29,880 --> 00:23:35,480
Matt: I think you're right. And maybe, maybe if I can somewhat summarize, it's like, it's not just the speed, it's the scope. Right.

183
00:23:35,640 --> 00:23:46,780
Matt: So maybe in the past, it's like, you've come out of a RailsConf and you have tech debt and it's in maybe one area, but it's like, in the scope of your project now, it's like everything. Everyone's like, reevaluate, everything. Yeah. Right. Yeah. Everything.

184
00:23:46,780 --> 00:24:00,440
Matt: And, and Robby, you're correct that the scope expands even out of our programming, out of my particular day job into the whole company to, you know, your friends and family and their company and the economy as a whole.

185
00:24:00,540 --> 00:24:09,040
Matt: So, and so maybe, you know, I think that's a good perspective that maybe that is the thing that is causing the dread is maybe not, not just the speed, it's the scope of it.

186
00:24:09,040 --> 00:24:15,660
Robby: Yeah. We're in interesting times. Uh, maybe we'll look back and be like, it was great to, we got to live through that era. And I don't, I don't know.

187
00:24:15,740 --> 00:24:24,240
Robby: Uh, I didn't, it's funny cause I joined, I got into this industry, not thinking that it was as, as a hobby thing. And then I, people would start paying me for it.

188
00:24:24,320 --> 00:24:24,420
Jess: Yeah.

189
00:24:24,420 --> 00:24:35,520
Robby: And now 25 years later, almost it's like, might that not be a thing anymore at some point? And I'm like, am I owed this? I'm like, look at all the work I did to get to this point. I'm like, I didn't go to college for this. I never had any college debt.

190
00:24:35,820 --> 00:24:42,480
Robby: I I'm just been like kind of fumbling my way through this career for a long time. So I'm just like, I'm happy to still kind of be around, I guess. And that's cool.

191
00:24:43,000 --> 00:24:52,060
Jeremy: Um, but it's not guaranteed to be a thing forever. Yeah. I'm definitely grateful. Similar thing. Like, didn't think this was going to be my career. Just like it was my hobby.

192
00:24:52,240 --> 00:25:03,760
Jeremy: And then I found out I could make money at it and I loved it and I couldn't stop. Uh, and then turned out, you know, like I could make a career out of this 25 years later. I'm very grateful. And if that's what that is, then it's worth it. Absolutely.

193
00:25:03,760 --> 00:25:14,600
Jeremy: I would be sad to lose it, but it's still good enough. And a lot of people hate their jobs. I know so many people that just hate what they do. And I've loved what I've done for a long, long time.

194
00:25:14,600 --> 00:25:19,680
Jess: So I know I consider myself. So I'll switch topics a little bit.

195
00:25:19,780 --> 00:25:33,720
Jess: Um, cause I don't know if I've heard anybody talking about this, but I've experienced it within our companies is that, you know, we talk about the speed, the AI with agents and how we can push more code, but then that creates a bottleneck in the PR.

196
00:25:34,060 --> 00:25:44,160
Jess: But now everybody's getting comfortable with like just pushing PRs and not reviewing those. But one thing I'm running into is that my customers can't keep up with the features.

197
00:25:44,720 --> 00:25:56,060
Jess: So like there's, that's, that's another problem to solve is, uh, what, what are you building? And, and, uh, are you making the right choices? Like you have all this power now to build these features. Yeah.

198
00:25:56,220 --> 00:26:05,420
Jess: Are you, are you building the right things? Are you building the things that your customers need? Can they even absorb those things? Because they have a hundred other things to do in their business and their workflows.

199
00:26:05,420 --> 00:26:11,140
Jess: And they're like, Oh, I'll, I'll even like try to do some customer research. I'll call them up and say, Hey, have you tried this new feature?

200
00:26:11,220 --> 00:26:13,200
Jeremy: No, I hadn't, I hadn't had time to try it yet.

201
00:26:13,860 --> 00:26:32,560
Jeremy: Uh, let me attach onto that, not just customers, but I have a friend here who was telling me last night that, um, he's, he runs a small, uh, team for a product and he told the CEO, uh, like recently we're, uh, we're running our, our backlog, our bug backlog down to zero.

202
00:26:32,660 --> 00:26:38,960
Jeremy: Uh, you've got like weeks left until we're coming, knocking on your door saying like, what are the next features?

203
00:26:39,340 --> 00:26:45,680
Jeremy: Cause we're like getting down to zero bugs and like that we like needed to deal with.

204
00:26:45,960 --> 00:26:56,780
Jeremy: And so like all, like that's completely changed, you know, like the, the owner of the business CEO or whatever you like, he's a thought, like, you know, I've got ears or whatever to like deal with this or like, I can always stay out in front.

205
00:26:57,040 --> 00:27:03,580
Jeremy: And now the team has progressed to a point where they're going to be knocking on the door very soon saying, what are we building now?

206
00:27:03,740 --> 00:27:12,440
Jeremy: So it's like both business owners and customers, how do they stay out in front of what can happen there? I I've not experienced that so far. Like I'm definitely seeing higher velocity.

207
00:27:12,700 --> 00:27:17,360
Jess: I have personally felt that like with client work, but, um, it seems like that might be happening more.

208
00:27:17,500 --> 00:27:21,520
Matt: Yeah. Are y'all seeing any of that at arrows? We're in kind of a different spot.

209
00:27:21,600 --> 00:27:36,920
Matt: Cause we, at the start of the year, basically started a new product and in a way we, we felt that like we needed to like rethink what we were offering, given that we can see maybe some trajectory in other industries from what has happened in the software industry.

210
00:27:36,920 --> 00:27:48,020
Matt: So I don't know yet. I think we're still, we're still finding our customers and we're still, you know, obviously we are shipping things out much more quickly, but I think that's also the nature of being early. Right.

211
00:27:48,100 --> 00:27:57,040
Matt: And, and we have pilots that are going on and we have customers in Slack that are telling us what's not working. And, you know, we're turning around and shipping things, you know, in a few hours.

212
00:27:57,040 --> 00:28:06,240
Matt: But, uh, to be honest, that was not a new thing necessarily when you're early. Right. And with Rails, you could always kind of turn things around more quickly than people expected.

213
00:28:06,920 --> 00:28:18,860
Matt: If you were making small changes, if you sort of, uh, didn't have a bunch of existing data or existing customers or existing workflows or, you know, that, that sort of thing.

214
00:28:18,860 --> 00:28:25,800
Matt: So I don't, I don't know that we've, we've seen that yet, but, um, I think it definitely is going to be a thing for people.

215
00:28:26,140 --> 00:28:34,200
Matt: Like it's always hard to get people to adopt new processes, software developers in this industry is, uh, these are folks that love doing that almost. Yeah.

216
00:28:34,200 --> 00:28:45,920
Matt: And, and certainly, you know, uh, a more regular business person is not, is not, you know, that hyped on it. I'm sure if you go talk to them, they're like, Oh, are you using AI in your business? And it's like, Oh yeah.

217
00:28:46,080 --> 00:28:58,060
Matt: I searched with, with Claude or I, I, I chat with, you know, ChatGPT or something in a way you're, we're sort of privileged in that we are almost like, you know, two years ahead of some of the other, the other fields.

218
00:28:58,060 --> 00:29:10,600
Robby: And so maybe that's part of, uh, what your advantage can be is, it's bringing that, yeah, bringing that back in my, in my space, the, uh, because we, we work on a lot of, uh, retainers with clients.

219
00:29:10,600 --> 00:29:19,220
Robby: So we kind of like a approximate budget we work with them each month. And so we've usually been able to kind of like scope that in a way that that's kind of based off of the capacity of our clients.

220
00:29:19,380 --> 00:29:27,820
Robby: Like when I say capacity of like their, not just budget, but like mental capacity and attention because quite often clients end up being a bottleneck and that's historically been an issue for us.

221
00:29:27,820 --> 00:29:33,680
Robby: Like, like we can, we, yeah, we could build, we could bill you more, but you're going to be a bottleneck unless you have some more people.

222
00:29:33,680 --> 00:29:44,000
Robby: But we know that you're an important person in your company and you can only respond to so many questions and tickets and be part of conversations with us or review something or be a final sign off or something.

223
00:29:44,000 --> 00:29:51,140
Robby: And so I think we've had to learn how to remove them a little bit from needing to be involved in all the approval process.

224
00:29:51,260 --> 00:30:00,120
Robby: So we have to build that trust, which relies on a lot of human letting them know that we understand where their goals are. So be like, Hey, we ship this, it's out there.

225
00:30:00,700 --> 00:30:10,140
Robby: Just want to let you know, FYI, you don't have to do anything like, so we're, we're trying to get our team to think more about like, let's show that it's working in production, take that little video. Yeah. Like, look, it's working. You don't have to think about it.

226
00:30:11,200 --> 00:30:14,140
Robby: Like what, what's, so they don't have to be a final sign off on some of that work.

227
00:30:14,460 --> 00:30:23,960
Robby: But as far as the, the throughput or being able to do more, uh, the, the, how we're using that extra time or capacity with the budgets.

228
00:30:23,960 --> 00:30:35,760
Robby: So if we're able to ship features a little bit faster with our little AI agent buddies, then the questions always end up being more like, okay, how can we, what are the things did we not get a chance to really take care of here? How can we do a more thorough job with this?

229
00:30:35,760 --> 00:30:48,540
Robby: Like how, like, let's go back and our, another question I was talking about, I was mentioning this earlier in a conversation, but I'm always like, if someone else had to come in and review this code base, what things would they highlight that we've, we've been kind of negligent about?

230
00:30:48,540 --> 00:30:55,780
Robby: Because there's probably things where things were our workarounds and I kind of touched on this in my talk a little bit, but I think we have time to do that kind of stuff a little bit more right now.

231
00:30:55,780 --> 00:31:06,300
Robby: But I, I do hear from folks where they're like, great, we got our engineering team. They can, they can push through a lot of things, but we don't know how to market, sell this stuff to people.

232
00:31:06,300 --> 00:31:15,700
Robby: We have a one client that like super, they've got a nice, big, healthy budget and we've built a bunch of stuff with them on a new, we don't do a lot of greenfield projects, but we did work on one this year.

233
00:31:16,040 --> 00:31:24,100
Robby: Um, and which is a fun story because like that was a project that we almost weren't going to get because they weren't sure about moving forward on this project.

234
00:31:24,100 --> 00:31:30,620
Robby: And then our project manager vibe coded an MVP for them and they sent it back and they're like, that sounds exactly like what we want.

235
00:31:30,800 --> 00:31:42,640
Robby: Like our project manager just took that extra initiative and spent like a couple hours, working on a MVP thing and it's been like, I think we made $150,000 on that project so far. So thank you, Michael for that, uh, clever idea there.

236
00:31:43,700 --> 00:31:46,120
Robby: I think the sales process is a little bit different for us now too.

237
00:31:50,040 --> 00:31:50,700
Matt: I think it's been a little bit different for us now.

238
00:31:51,340 --> 00:31:51,580
Robby: Yeah.

239
00:31:51,580 --> 00:31:51,860
Jeremy: Yeah. Yeah.

240
00:31:51,860 --> 00:31:59,900
Robby: And I'm like, well, we can't solve that with AI necessarily. Like, well, where's the customer base? Like, yeah, we can build things really quickly. Yeah. You've got all this stuff there, but like you need to find your customer.

241
00:31:59,900 --> 00:32:05,400
Robby: Like you've got all this stuff there, but like you need to find your customer and that takes time.

242
00:32:05,860 --> 00:32:16,120
Jess: And I think it's going to take a while for, uh, businesses and managers to sort of catch up because they're used to this backlog of things that takes six months to build.

243
00:32:16,520 --> 00:32:20,400
Matt: And, and now there's a different velocity. Yeah.

244
00:32:20,440 --> 00:32:34,320
Matt: Well, and it's part of the job as a software product builder or a business owner is show the value to them in the world and their language that they understand because they ultimately, they didn't care about the technology before.

245
00:32:34,460 --> 00:32:42,640
Matt: They don't really care, but maybe they slightly care if it is AI because that's a hot buzzword that they want to make sure they're not getting left behind.

246
00:32:42,680 --> 00:32:54,060
Matt: But certainly none of your customers care about what model you're using, what harness you're using, what skills you're using. And that's always been the case. Like they don't care what version of a database you're using or what language or framework.

247
00:32:54,060 --> 00:33:05,180
Matt: And, and, and, you know, that, and I think that's, uh, just worth, worth keeping in mind that like, that's part of our job is, is to translate that back. And, you know, this is a transformational technology.

248
00:33:05,340 --> 00:33:12,420
Matt: So of course that means that the, the answers are not well-written, right? I think in the past we could, we could all probably off the cuff.

249
00:33:12,660 --> 00:33:17,560
Matt: You know, describe why a web app might be better than a spreadsheet for some business process.

250
00:33:17,560 --> 00:33:18,320
Robby: Is it?

251
00:33:18,420 --> 00:33:26,340
Matt: Well, depends. Does the spreadsheet have an AI chatbot in it? Yeah. Right.

252
00:33:26,460 --> 00:33:35,640
Matt: But I just think going forward, it's like, it's not yet clear necessarily, like how, how do we best use this technology to help business outcomes and where does all that fit in?

253
00:33:36,080 --> 00:33:46,280
Matt: And so that's just, I think adding a little bit to the, to the uncertainty or the feeling like, Oh, I used to have the answers for how to sell consulting services or to, to, to know what tool to pick.

254
00:33:46,280 --> 00:33:54,680
Jeremy: And now it's just a little fuzzy. Robby, uh, you, um, highlight a new framework in your talk called WUTS I think it's called, is that right?

255
00:33:54,840 --> 00:34:06,480
Robby: Did I pronounce that right? Yes. You got to say it loudly and quickly. WUTS, that's, uh, kind of my need to give a silly name to a sequence of otherwise normal human behaviors.

256
00:34:06,480 --> 00:34:17,160
Robby: Uh, so WUTS is W-U-T-S and it stands for wonder unpack tinker and share. And you can find out more information at wuts.wtf.

257
00:34:17,420 --> 00:34:28,860
Robby: I bought my first .wtf domain name. Nice. And so I'll be sharing some stuff there. I actually was working on some, uh, like Claude skills and stuff for it, but I decided I wasn't going to publish them immediately.

258
00:34:28,860 --> 00:34:40,040
Robby: But I want to test them a little bit more, but the idea is just more about scratching that curiosity it's like when you're working on things, like let's learn more about our infrastructure. Let's learn more about why that one weird thing is doing the thing. Just having a little bit more of a framework.

259
00:34:40,040 --> 00:34:50,000
Robby: That's not tech stack specific. And, um, because I have a tendency just to like jump into solving problems. I like, I love the endorphins of that.

260
00:34:50,040 --> 00:34:56,220
Robby: And I don't always care to understand why things are the way they are. And I think that's also been detrimental for me at times.

261
00:34:56,220 --> 00:35:03,400
Robby: And so I've been, been thinking a lot more about that over the last many months and wanted to try to like give myself a little bit more structure for that.

262
00:35:03,560 --> 00:35:13,960
Jeremy: And I thought I would share it and see if it resonated with people. Yeah. That's cool. Yeah. That's interesting. That is like my go-to is to like, try to understand why things are the way.

263
00:35:13,980 --> 00:35:20,140
Jeremy: That's like, just like, why are things this way? Like, uh, yeah, I think like doing tons.

264
00:35:20,160 --> 00:35:32,040
Jeremy: Like I'm the kind of person that before building a feature, I would like do research on like, how, how has everyone solved this, you know, in history before, before even sitting down to implement, that's always like been my approach.

265
00:35:32,100 --> 00:35:33,700
Jeremy: But I know it's not, not everybody.

266
00:35:33,740 --> 00:35:39,580
Robby: Um, and, uh, yeah, just having like, yeah, like being curious and asking those questions and then tracking that stuff down.

267
00:35:39,580 --> 00:35:48,460
Robby: Just, and I, and I was in for, because in the talk, it was more focused towards people that are inheriting code bases or inheriting other people's decisions.

268
00:35:48,820 --> 00:36:00,640
Robby: And that, and a lot of people joining projects don't understand why things are the way they are. They assume that they're the way they are because they should be the way they are. Yeah. And that's not necessarily, uh, uh, I think a healthy way to look at it.

269
00:36:01,660 --> 00:36:07,580
Robby: I think we should be able to question things. You can also not be that annoying question person on a team. That's why did we do it this way?

270
00:36:07,580 --> 00:36:08,420
Jeremy: Why did we do it this way?

271
00:36:08,460 --> 00:36:15,300
Robby: Like, that's not super. So do want to encourage people to like go answer those, some of those questions themselves and not subject your teammate to that.

272
00:36:15,400 --> 00:36:22,520
Robby: Uh, but I do think the, the, the tendency is just to like not ask or to not inquire more.

273
00:36:22,520 --> 00:36:32,660
Robby: And I think, I think right now, because, because AI, I think there's, there's this interesting thing where I think we're, we're going to become less curious. I worry that we'll become less curious and I, I don't think I want to give up that.

274
00:36:32,940 --> 00:36:41,560
Robby: I don't know if that's that skill or value. I think that teams could benefit from right now. Cause I think that will then allow you also just to readapt and think about how your job's going to evolve.

275
00:36:41,560 --> 00:36:47,160
Robby: And again, like be the person that's asking or inquiring or sharing their, that was a big part of it.

276
00:36:47,160 --> 00:36:58,700
Robby: It's not just asking questions, but also then sharing maybe like a tool or useful resources or documentation or play playbook or something like doing a little bit more than just being like, Oh, I figured this out for myself.

277
00:36:58,700 --> 00:37:05,520
Robby: And then I'm going to just going to keep that information in my head. I don't think that's as valuable as just taking that extra step. And like, does everybody else know how to do this now?

278
00:37:05,660 --> 00:37:17,460
Robby: Now our agents can also maybe take advantage of this new tool that we, I built that simplifies things or been set up actually does what expect like, yeah. I know that people are like, what do you use been set up? They're like, no, it doesn't really work the way we work at our office.

279
00:37:17,460 --> 00:37:26,540
Robby: I'm like, you know, it's a bash script, right? Like you can make it do whatever you want. You have complete control. Like why can't you do this in one command? That's true.

280
00:37:26,560 --> 00:37:37,640
Robby: And that just seems like such low hanging fruit that I, I don't know. So I've used that as an illustrative example there, but I, you know, so I just want to encourage people like simplify it for yourself and for your coworkers and for your agents.

281
00:37:38,060 --> 00:37:42,920
Jeremy: Yeah. One thing I've been thinking about a lot, cause I've worked on a lot of legacy code bases for clients.

282
00:37:42,960 --> 00:37:52,280
Jeremy: A lot of times, uh, we're changing direction architecturally, like, and with, especially with Claude and stuff, it's going to keep doing more of what see it already.

283
00:37:52,420 --> 00:38:01,920
Jeremy: And so I've been trying to figure out like, okay, I need to, I've already identified several things that I know we're not going to go with this old pattern, but it exists all these other places in the code base.

284
00:38:02,040 --> 00:38:05,640
Matt: And I need to sort of audit that and then write. No, we're not doing this, but it's also for people.

285
00:38:06,340 --> 00:38:11,280
Jeremy: Like we're not, we've decided not to do things this way anymore.

286
00:38:11,580 --> 00:38:21,500
Jeremy: And we had that conversation, you know, over zoom or whatever, but it never made it into documentation or whatever to say, like, this is, don't look here. Like if you see this, don't copy that. We're not doing any more of that.

287
00:38:21,500 --> 00:38:25,200
Matt: We're not going to remove it right now. Um, but don't do more of it. Yeah.

288
00:38:25,480 --> 00:38:36,560
Matt: I think the interesting thing there is like in the past you had to have a certain team or be a certain size to really justify like automating or writing your own custom dev tooling to do that. Yeah.

289
00:38:36,740 --> 00:38:46,960
Matt: Now a lot of time and energy. Yeah. There's the classic joke of, you know, I, I, I spent eight hours writing the script to save me, you know, 15 minutes. And you know, how many times do you have to run it?

290
00:38:47,140 --> 00:38:48,040
Jess: Right. To pay off.

291
00:38:48,040 --> 00:38:56,380
Matt: Realize that benefit. Yeah. Calculus just changes a lot on that. So like when I hear that, that example, Jeremy, I'm thinking like, yeah, that's great.

292
00:38:56,560 --> 00:39:01,260
Matt: And in the past you would have just told people like, don't like, you know, which are the bad files.

293
00:39:01,820 --> 00:39:10,600
Matt: And, but I don't think we would have really said like, does it really make sense to write an automated script that will reject a commit that if it does it that way, or, or it was simply too difficult. Right.

294
00:39:10,780 --> 00:39:20,640
Matt: Because you didn't have the ability to just, you know, you could have that, you could have some, something run a cheap model check to say like, are you using the old pattern before you, how do you, how do you hard code that in?

295
00:39:21,780 --> 00:39:33,700
Matt: Especially in a language like Ruby where you have, you know, dynamic metaprogramming stuff involved. So I don't know. I think that's just like a good example of like, that is still engineering work. That is still value and judgment in your experience.

296
00:39:33,780 --> 00:39:41,060
Matt: That is still, you're making a decision to do something. You're still applying, you know, the strangler fig pattern or something.

297
00:39:41,080 --> 00:39:51,540
Matt: You're still saying like, oh, I see the value of CI and linting and how can we automate a check so that we remove the human judgment from it. So to me, that's just like, it's more exciting.

298
00:39:51,540 --> 00:40:02,920
Matt: It's definitely different than the day to day, but I don't, I don't think it goes away. I think, you know, for me, the biggest realization of the past several months is that, you know, software is more than just the software construction.

299
00:40:03,040 --> 00:40:12,200
Matt: There's, you know, on one end, there's the product and the user and the user experience. And then on the back end, there's like quality assurance, there's verification, there's rollout, monitoring, all that.

300
00:40:12,200 --> 00:40:21,480
Matt: And, you know, while the software construction in the middle is getting kind of squished, I think engineers will have to choose which direction they head. Yeah.

301
00:40:21,600 --> 00:40:29,640
Matt: And I think, you know, for people that were more into the craft of the code, the technical, like I just, I just like the puzzle solving nature. Like maybe you shift toward verification.

302
00:40:30,500 --> 00:40:41,800
Matt: How do we, how do we, you know, how do I spend my time writing a simulator that can test, you know, this external API integration in an automated way that before we would never do.

303
00:40:41,800 --> 00:40:52,640
Matt: You know, those people can go that way. And people that are more want to focus on the business outcomes or the product thinking, or I don't care about the code. I don't want to look at the code. I can look at the finished app and I can polish the last little bits.

304
00:40:52,740 --> 00:40:59,020
Matt: Like maybe they're shifting more toward. More to the human side. More toward the product. You know, what typically would be a product manager or something like that.

305
00:40:59,020 --> 00:41:10,640
Matt: And, you know, I think that's, that's been my realization and my hope that people will sort of take away from some of this is like the software construction part is maybe going away, but that doesn't mean that the role of software is going away.

306
00:41:10,820 --> 00:41:12,780
Matt: You just need to sort of find something else.

307
00:41:12,780 --> 00:41:26,020
Matt: And, you know, I think once you find that you can achieve some, you know, similar flow state in one of those two domains, I think it becomes a little bit less scary and you can sort of see that the future is, is maybe not as, not as uncertain as you might think.

308
00:41:26,120 --> 00:41:26,720
Robby: Flow state.

309
00:41:26,720 --> 00:41:37,340
Jeremy: That's the thing that's killing me the most. I'm a flow junkie and I lost it this year. Yeah. You know, like my flow's gone. I'm burned out every single day. Like, just like toast. My brain's fried.

310
00:41:37,340 --> 00:41:43,380
Matt: Yeah. Like from so many work streams and stuff. And I miss flow so much. Yeah. And I struggled with that too.

311
00:41:43,380 --> 00:41:52,860
Matt: And like, it's only in the, you know, recent weeks that I've started to have more sessions that are like that, where you can, where you can see, or you get the satisfaction of like, oh, this is cool.

312
00:41:52,860 --> 00:42:02,560
Matt: When I can spin up a cloud agent that records a video of its, of its thing and does a better job, you know, testing and documenting than I, you know, I think eventually something will click.

313
00:42:02,560 --> 00:42:17,640
Matt: I mean, for me, another good example was like, if you've ever tried to debug a customer issue and you've had like a fully realized setup where you give your agent a read-only production database access, you give it access to your, you know, to your AppSignal errors or your, your logs.

314
00:42:18,320 --> 00:42:22,000
Matt: Like just seeing that go, it's hard to go back to it. Yeah.

315
00:42:22,720 --> 00:42:30,760
Matt: I kind of felt like some sense of flow of like, I set up all the pieces in motion and then I watched it go and it was like, it was satisfying my brain in the same way.

316
00:42:30,760 --> 00:42:40,380
Matt: And so hopefully you'll find, you know, something like that, because I think, you know, that, that will be important going forward because, you know, so many people did get into a flow state.

317
00:42:40,560 --> 00:42:51,860
Matt: You're in the headphones, you're, you're, you're, you've got your beautiful editor with your code color scheme, your keyboard, your keyboard switches that you've dialed into perfection and you just go and, and, you know.

318
00:42:51,860 --> 00:42:54,800
Jess: My VIM bindings. Exactly. Yeah.

319
00:42:54,800 --> 00:43:05,940
Jess: I, that reminds me of someone I've read was talking about the term of an architect and how an architect designs a building, they plan the building, but they don't, they're not on the ground doing the work.

320
00:43:06,140 --> 00:43:08,440
Robby: Yeah. And I feel like that's, that's where a lot of us are.

321
00:43:08,880 --> 00:43:12,580
Jess: But they get to put the cool photos on their, their portfolio.

322
00:43:12,680 --> 00:43:15,480
Robby: Exactly. Look what I built. Yeah. Look what I designed.

323
00:43:15,480 --> 00:43:20,640
Matt: The finished product. Yeah. So they take it, you get to take their pride. Yeah. But it's not too different now.

324
00:43:21,080 --> 00:43:31,380
Matt: It's the same, you know, I shipped this feature and you know, whether I wrote the code or not, you're, you're, you're very rarely reporting the code to customers in a, in a release or even to your own team.

325
00:43:31,380 --> 00:43:34,440
Jess: Like we have a release channel where like, Hey, this feature is now live.

326
00:43:34,840 --> 00:43:38,240
Jeremy: You don't tell them what class you wrote to do it. Exactly. Yeah. So. True.

327
00:43:38,720 --> 00:43:40,460
Jess: Well guys, this has been a great conversation.

328
00:43:40,600 --> 00:43:45,040
Robby: Thanks so much. Robby, Matt. Going in the spur of the moment. Yeah.

329
00:43:45,080 --> 00:43:48,360
Matt: Thank you, Jess. Thank you, Jeremy. And thank you, Matt. It's been a delight to chat with you both.

330
00:43:49,700 --> 00:43:54,640
Robby: Buzzsprout for putting the podcast together. Yeah. They always come through. Yeah. They do.