Brendan: 0:37

today we've got Jeremy Smith with us. Those of you who listen to IndieRails I'm sure are familiar with him, and also I, you do lots of conference talks and also run, Blue Ridge Ruby. So, uh,

Jeremy: 0:47

Yeah.

Brendan: 0:48

Glad to

Jeremy: 0:49

Hey, thank you so much for having me. Yeah, I appreciate it

Ryan: 0:51

I, I gotta, I saw on LinkedIn you spent 140 hours preparing for your Rails World talk

Jeremy: 0:57

Yeah

Ryan: 0:58

to 300 hours for your conference, right? Like, so

Brendan: 1:02

I saw that too.

Ryan: 1:03

get, I didn't get to go to either of those unfortunately, even though I So how did, how did… Tell us how, how did it go?

Jeremy: 1:09

Uh, the talk went pretty well. Um, I am not the most amazing speaker, and it's like, it's a, it's a personal goal of mine to keep getting better at it. Um, I haven't really been speaking at conferences too much other than the last five years. So, um, basically after, uh, COVID, after my kids were old enough, uh, it felt like a time to kind of reinvest myself in the community, and I wanted to push myself into, to growing in some new ways. One of them was podcasting, one of them was speaking at conferences, and then eventually, like putting on my own conference. So I'm still getting there. My delivery is not… Like, when I watch videos of myself, I hate it. I can't, like, I see all the, the problems.

Ryan: 1:56

Don't, don't

Jeremy: 1:56

I'm also…

Brendan: 1:57

J-

Jeremy: 1:57

Yeah. I-

Brendan: 1:59

when I'm editing.

Jeremy: 2:00

I got over that. Okay, with IndieRails it drove me crazy at first. Like, the first 10 episodes it was really hard to listen to my voice. I can listen to my voice most of the time now and not be bothered by it.

Brendan: 2:11

Yeah, I, I agree with that. I'm getting

Jeremy: 2:13

yeah.

Brendan: 2:13

to my own voice, but

Jeremy: 2:14

Yeah.

Brendan: 2:15

other little ticks I have that

Jeremy: 2:16

Yes. The tics, yeah.

Brendan: 2:19

Yep.

Jeremy: 2:19

d- uh, that and then there's so much more with video because you're getting all of your mannerisms and all that that's coming with it. And so, um, I'm still working on delivery. I'm also… I prefer things like I would rather write an essay than do something off the top of my head, 'cause s- half the time it could be good, and half the time it's just gonna suck, and I can't control it. I wanna control the quality

Brendan: 2:43

Yep

Jeremy: 2:43

I hate presenting lower than the quality that I would want to be in the world, right? Um, and I am getting better at, at least on podcasts, at talking off the top of my head. Um, that's really, I think is really, it is learnable in a, in a way to, like, get not just over the, the uncertainty or fear or whatever it is that is difficult when you're starting, um, but then ac- to actually get good at just kinda talking off the top of your head. So I'm getting it better at that, but I can't do that with a, with a talk. Like, a technical talk, I don't know how to do it, so I write it, and then I practice it a bunch, and the delivery's just n- still not where I'd like it to be. But that said, I think I make f- up for it by just, like, putting way too much time into things. And this year it was, like, um, it was a, it was a talk on Hotwire, so going through Stimulus and the three parts of Turbo, leav- leaving out Native, so basically everything that you do on the front-end web. And then, uh, I wanted to talk through pitfalls that I've hit in each one of those, kind of do a tour through each one, one of them, and talk through pitfalls. And then halfway through my preparations I realized, "Oh, I could do, like… This could be like Hot Ones, where they're, like, sa- we're sampling these, like, these hot sauces, and it could be, like, the burns." And, 'cause I didn't wanna say pitfalls. And so that ended up giving me, like, a visual language to talk through, like, the, um, uh, th- through the different, uh, examples and, like, scenarios that I've run into. And I r- I'm always looking for that,'cause I realized a few talks ago that once you have, like, a story or some kind of hook or a visual metaphor, it makes everything else go easier. The scriptwriting goes easier, what to put on slides, how to style it, how d- your color choices. Like, so much of it can fall out of that, and then people… And it feels cohesive. It also gives people something to reference back to. Like, "Oh, that was the one you did, like, on Hot, on Hot," you know, like, "On the hot wings," or whatever. Um, and so I think it be- becomes stickier in people's minds when they want to reference something. So, um

Ryan: 4:50

That, that's a great tip.

Jeremy: 4:52

Yeah, thanks. Uh, and then I spent way too much time in, like, AI image generation tools, like getting d- all the Hot One sauces. Like, going to the Hot Ones videos, s- taking screen grabs of all of them, and then bringing them into this tool called Flora to then generate my own hot sauces for each one, 'cause I'd named all of the hot sauces, all the pitfalls. Um, and so I needed to have, like, the labeling right, and I decided that each flight would be a different kind of pepper, so the color in the s- of the sauce needed to be a slightly different… You know, it was orange, and then red, and then dark red, and green. So, so like, all these different colors. And, um, so then I was just, like, generating tons and tons of, uh, images for that to, like, bring into the slides. But, um, it was satisfying at the end just, like, to have all that put together. So yeah, it was good. Uh, it was too much. Too much for the summer, too much for getting into this fall, but, uh, I was… I had set… I'm the kind of person that I have to set difficult, some somewhat difficult goals for myself, um, and then just, like, as my hobby, I guess. I don't know if this, is that what it is. I don't know if it's a hobby, but, um, and then just see if I can, like, make it, pull it off, you know, without, uh, killing myself or being too terribly stressed. But a little bit, I need a little bit of stress, even in, in my side activities or whatever, um, for some reason, so.

Ryan: 6:16

Oh, it do-- it's a good motivator

Jeremy: 6:19

Yeah. Yeah, I need, I like having a fire lit under me, I guess. I don't, I don't know why that is, but, but it, uh, it gets… Like, when I look back and think,"Oh, I, I've done some fun stuff that I'm really happy about," it's like all the stuff, l- most of it's like the stuff that I did on the side, you know, that wasn't a requirement. I love my client work, but I, I lo- the, all the, the other things that I'm doing end up, like, I call them part of my business, you know? It's not really like, they're not billed hours. You could track maybe some of that back to some kind of inbound maybe for me, but mostly it's like, this is just like what I love, and so this is, I guess it's my hobby and it's not really work-life balance, you know? It's just all mixed together.

Ryan: 7:03

Right

Jeremy: 7:04

but those e- these extra things are like what I get really excited about.

Ryan: 7:08

Well, that, that's a good segue'cause, uh, my next question was gonna be, right, like you spent about, you know, two months or s- 16% of your sort of work hours in a year on

Jeremy: 7:18

Yeah.

Ryan: 7:18

things, like, a-and you sort of already answered, I think, maybe part of the question. I don't know if there's anything else to add, but like why, right? Like, okay, it's a hobby. You love doing it. You like the challenge. I mean, is there any, you know, business value for you, right? Like is that how it

Jeremy: 7:34

Yeah

Ryan: 7:35

then you just realized you loved it? Or is it like, "Hey, I needed a distraction and, you know, I, I wanna do it"?

Jeremy: 7:42

I think that if you are a freelancer or a consultant That You need, number one, you should, you should always have some kind of marketing and business dev, um, processes or things that you're doing. You need to be putting time, investing time and resources into that. Um, and it can l- I think it l- can look a bunch of different ways. But, um, when I started freelancing, that was not something I appreciated. I treated it more like this is, um, I'm just, I'm working part-time for a few different companies and, and that's it. And then if, if suddenly they needed to let me go, well, no one knew who I was, you know, no one… I hadn't connected to anybody else. Um, I hadn't built relationships. I hadn't built a reputation, all those kind of things. And then you're back in this spot where you're suddenly ne- you know, kind of going hat in hand to the next place."Will you hire me? Will you hire me?" And that's not a good place to be as a freelancer, and I don't wanna, like, oversell that 'cause there's still ti- plenty of times that I'm like, "Oh, shoot, am I gonna h- you know, will I have enough, you know, um, leads, or will I have, you know, enough clients?" Um, so there's still plenty of volatility for me. But one thing I've really learned is that you have to keep investing in building, um, a reputation and connecting to people that you would like to work with or that might like to work with you, and it's not enough just to stay at home and then create, uh, open source repos or I don't even know what we do now. Um, but you know, like, that's not, it's just not enough. But turns out I actually also really like those things. I love meeting people. I love meeting Rubyists. I love building stuff and sh- showing other people. Um, I love talking about technical topics. I like speaking. Um, I like interviewing people. I like organizing things. So all this stuff ends up just kind of feeding back on itself, and I realize, like, a lot of those things, like, that we think of, you know, like, typically developers are thinking, like, "Oh, if I want to work for myself, the thing I would hate is the marketing or the sales." And when I stopped thinking about marketing being this thing where I'm trying to get people to do something they don't wanna do or entice them to a thing that, that's not really good for them and started to realize, like, no, it's about connecting to people that actually want what I might be able to give them or just to, like, serve people and help people and do something good in the world that would be be- beneficial. Um, and, uh, when, like, k- kind of making that switch really unlocked it for me so that it, like, oh, this is where I, I can just invest in doing these things. Just do, do cool stuff with people to be helpful, um, talk to people, all the things that, I don't know, they feel very natural now. But what, you know, kind of, I don't know, at some point earlier in my career did not feel natural or didn't, I couldn't see how to do it, how you… It seemed like what you should do is send, you know- Just cold emails or just call random business owners or trade business cards or something like that. And I realized, like, I don't like that, and I'm guessing most other people don't either. But what I like to do is learn stuff from other people. I like to talk to people. Uh, I like to have people listen to me when I'm talking and sharing about my problems. Um, and I can do that for other people, and that ends up being really rewarding. So that's like just… Yeah, that's, that's now just a regular investment in my year. I don't h- I don't put exact numbers on it, but I always wanna have something going. Maybe right now I'm like,"Okay, Rails World was last week. Maybe I don't need anything going right now." I could use, I could use a peaceful October, but by November I'll be like,"Hmm, what, what, what should we do now? Like, what's next?" Um, yeah

Ryan: 11:40

I think that's a, just a healthy perspective on business in general, right? Serve people, try to help people to the best of your ability. You know, h- have a process that you enjoy doing and, you know, don't do the things that you don't. Like, uh, for, for a long time I was like, "All right, I need, I need to build my personal brand. I need to build my personal brand," right? What, what does that mean?

Jeremy: 12:01

Yeah

Ryan: 12:02

know, you can do a podcast, you could do writing in a blog, you could do LinkedIn, you can do video and, and it's like, I actually like writing, right? But it, it, it, it can be tedious, right? And it

Jeremy: 12:12

Угу

Ryan: 12:13

time-consuming. you know, so obviously, ultimately, I settled on podcast as the, the leading thing and haven't done anything beyond that. But, you know, I think it's important to remember if you're in… You find yourself in the same spot is, like, find the thing that you enjoy doing 'cause then it doesn't feel like work

Jeremy: 12:30

Yeah. I think part of the, the thing about writing too, like I have, had the same thing where I have a backlog of all these blog post ideas that I've, um, wanted to get out, and even drafts that I just haven't gotten over the finish line. One of the things that's hard about writing on your site is, or it has been for me, is that I get way less feedback about those things than I do from posting on Twitter or doing podcasts or going to conferences, um, speaking or things like that. You get that immediate, um, reward of, oh, I'm making new friends, or, oh, like I'm talking to people and seeing their reaction right now, as opposed to just like throwing words out into the ether and you don't know if someone's… Like, I've had occasional times in, like, where someone emailed me and thanked me for a blog post I wrote, but it's like not very, it's not very many times, you know? And

Ryan: 13:22

hit

Jeremy: 13:23

Yeah. It's too much of a delay that I couldn't learn the reward, you know, from that. Like, so yeah. Um-

Brendan: 13:32

Yeah, I guess one thing that I'm curious about too is, like, h- how is consulting going right now? You know, with… We're talking about the marketing and getting, uh, you know, clients. And, and by the way, like, you know, the three of us were all at Sin City Ruby, like, was that two years ago almost?

Jeremy: 13:47

Yeah

Brendan: 13:48

if I remember right, your talk was mostly about, like, how do you become a freelancer or a

Jeremy: 13:53

Yeah

Brendan: 13:54

And so there was a lot of good advice, you know, if I remember right. I haven't actioned any 'cause I haven't, you know, I still have a full-time job, but, things like that are in my head, so I'm just kind of curious, like, if I'm someone that actually would want to take that advice now, like, what would your advice be to them? Like, how are, how is consulting going for you at this point?

Jeremy: 14:16

I think, um, I'm gonna take it from the angle of the person that's thinking about going into freelancing and consulting, 'cause that's what I was hearing from you maybe.

Brendan: 14:24

Yeah, that, that's a good place to

Jeremy: 14:25

yeah. Okay. Yeah, maybe to start there. This is-- I, I realized in January, this is a really hard year to enter consulting and freelancing. I would not recommend it. And it's for a number of reasons. Like, in, in the US we, we lost Obamacare, some of the government subsidiza-ization of, affordable healthcare. Um, so that was like a big hit at the beginning of the year. And then just like AI nonstop, um, complete game change for everybody probably in how their, you know, what, what their workflow looked like a year ago. But then if you're also trying to figure out like how to adapt that to each of your clients and then also anybody coming to you now, like you're trying to figure out what are-- what's my story that I'm telling? Like, wh- what services am I offering? What do people value now? There's still like a… There's n- never an end to the h- the needs that humans have, but those needs do shift and move around. And so this is a year of major volatility where it's really hard to tell what do people need now. And I think a lot of people don't know what they need. And when people don't know, aren't certain that they're making good investments, they hold their wallets. So I think a lot of… My guess is, and this is-- I should s- also say like I'm a, I'm a freelancer. I work for myself. I'm not an agency. I don't have a huge lead flow. Um, I'm, I'm talking to a number of other independent consultants. I talk to some agency owners. Um, I'm not the most connected person in the world, but you know, I talk to a few. So from my-- I have a limited perspective here. Um, but my sense about this year has been that a lot of companies are holding their wallets tighter. They're-- It might be a… My guess is that it's a, a sort of a wait and see. Like we're not sure what's happening here. Or maybe it's a, "Hey, seems like maybe with, you know, fewer people and a few Claude Code subscriptions, we could, you know, we could bang this out ourselves and we don't need to hire anybody else." Um, and maybe they give that a quarter or two and then find out if they did or not, but that's still a quarter or two that they didn't go hire somebody else. So there's now been this gap, this period that a lot of, uh, freelancers, agencies, people doing-- offering services are having to stretch across, waiting for work to show up that maybe it'll show up. But I'm betting a lot of people were like, "Well, you know what? Actually, maybe we don't need this. Maybe we can just pay the, the max plan over here and just like, you know, churn some tokens and, uh, get the same outcome." And it could be, or maybe they discover that leads to something else that they need instead. And then what is that thing? And how do we get in front of that? And I don't know what that is yet. And so I think that's what everybody's experiencing right now, and it's just like having to work so much harder to stay in one place.

Brendan: 17:27

Yeah. That makes sense just looking at it from the outside and thinking about it a little bit too, like even, like AI consulting specifically, not Rails specific stuff. Like I've thought about that a little bit, like there's opportunities, like we talked to Colleen Schnettler a while ago,

Jeremy: 17:43

Mm-hmm

Brendan: 17:43

she was doing it. And the thing there is like, well, and it's probably the same with the Rails stuff too, it's like at a certain point, if you're doing AI stuff, like specifically, you're gonna g- the, I would expect the companies get to a point where it's like, what do we need you for anymore? You showed us how to do it, and now we

Jeremy: 18:01

Yeah.

Brendan: 18:01

it, and the AI can do it."

Jeremy: 18:03

Right

Brendan: 18:03

thing with the coding, like you were saying. It's like, well, maybe we just get a Max plan and

Jeremy: 18:07

Mm-hmm

Brendan: 18:07

see how it goes.

Jeremy: 18:08

Yeah

Brendan: 18:09

So that, yeah, obviously I haven't experienced it, but what you were saying there seems to be in line with that thinking

Jeremy: 18:16

Yeah

Ryan: 18:17

we're starting to already come out of that though, and may-maybe I'm too optimistic, right? But, you know, I think that w- the initial reaction is, "Oh, I can do it," right?"Claude can do it. Like, I can save the money." But what, at least what I'm seeing for myself is that I'm like, yeah, I'm moving a lot faster, I'm pushing a lot more code, but now my, like, my universe is bigger, right?

Jeremy: 18:40

Угу

Ryan: 18:41

that I want to do now that it's like know that I can do it,

Jeremy: 18:45

Yeah

Ryan: 18:46

I'm, I'm actually swinging back the other way. Like I'm, you know, I d- I, I just hired FastRuby to do an AI project about OmbuLabs to do

Jeremy: 18:53

Cool

Ryan: 18:53

project, right? But it's like, right, well, I, I still need capacity, right? I'm still only one person, and there's only so many hours in a day. So

Jeremy: 19:01

Yeah

Ryan: 19:01

hopeful that, that the, the wallets are gonna loosen a little bit as people, like, expand. You know, the market e- is expanding.

Jeremy: 19:09

Угу

Ryan: 19:10

ability to, like, sort of envision a product is expanding now that, like, you know, you sort of infinite possibility really. Um, projects that I never would've, you know, taken on individually or even as a company, I'm like,"Yeah, I c- I can do that." And a lot of times I can, you know, get a, you know, 80% there in, in a, in a weekend.

Jeremy: 19:33

Yeah

Ryan: 19:33

the last 20% takes another three weeks like it

Jeremy: 19:35

Yeah.

Ryan: 19:36

But, um, yeah, I, I don't know. It's, I think it's a interesting time to be here, but I, I'm still… I, I would, I would be bullish on consulting still. I mean, a-as

Jeremy: 19:50

yeah

Ryan: 19:51

as, as a long-term, know, opportunity for someone who's driven. now there's just a lot of uncertainty and yes, the market is a little bit hard. You know, I don't think that's only AI. I think there's other,

Jeremy: 20:04

Yeah. The market was already bad. Yeah, it was already bad like a year ago. Yeah. Yeah

Ryan: 20:09

we don't need to get into that, but, like, just spending is down, costs are

Jeremy: 20:12

Угу.

Ryan: 20:12

People are a little skittish. We-- It doesn't m- That, that's industry agnostic, right? Across the board, doesn't matter what you're doing. So, um, yeah, I don't know. How-- What do you, what do you think about that?

Jeremy: 20:24

I, I definitely think that I, I love working in service. Like, um, it's hard in a time like now where it's really hard to know what people need, where prior to this I'd identified a market. Okay, I work with, um, I work with CTOs of small organizations or bootstrapped, you know, SaaS, uh, founders who have decent revenue with their product, um, can't hire a full-time team, don't necessarily need it, but need somebody that is coming in, taking some ownership, filling a fractional role, taking a res- fractional responsibility in, in this product, and can help them out, you know, on, uh, you know, heavy lifting on back-end stuff usually. Um, and that may still be the place that I go, but it is already looking very different serving those people. Um, the velocity has changed. Um, the way we're, yeah, the way we're shipping has changed. Um, I've been trying to figure out what the what my relationship to responsibility is now, like to my resp- operational responsibility is because there seems to be a higher risk appetite that people have right now. Like it's… We're more willing to live with, um lower quality code or with more bugs or defects if it lets us go faster. Everyone feels like they need to go faster right now. It's been my feeling, like, and I'm not saying that everyone is doing this, but my feeling has been that there's pressure to go faster. Um, I feel it. I'm not having anyone tell me directly. Yeah. Okay.

Ryan: 22:20

I

Jeremy: 22:20

Yeah.

Ryan: 22:21

uh, you know, as you were saying it, I was thinking as like, it… When I think about the risk,

Jeremy: 22:27

Yeah

Ryan: 22:28

quality code

Jeremy: 22:30

Угу

Ryan: 22:30

as humans perceive it,

Jeremy: 22:33

Yeah

Ryan: 22:33

It's harder for us to understand. I'm not sure we're actually getting worse output. Like, it, like, you know, using an app in the web browser sort of experience. And what I can say, I think with some certainty, is when there are defects, they get, uh, fixed, you

Jeremy: 22:53

Faster

Ryan: 22:53

uh, f-faster. I mean, almost instantly, right? Like, I have an agent that just watches Honeybadger, right? It sees an error, it does the research. If it is a simple thing, right, it puts up a PR. If not, it creates a ticket for me and, and flags it. I mean, I, I'm… that maybe would've sat in the queue for weeks are done in minutes

Jeremy: 23:14

Yeah. Yeah

Ryan: 23:16

and the smaller I can get the agent to do the chunk of code, right, the smaller that PR is, the faster it moves.

Jeremy: 23:22

Угу

Ryan: 23:23

I think that's an important too. But, um, y- I forget where I saw this recently, um, and if it was, if it was you, I apologize in advance. But, right, s-someone was saying, you know, like some of the principles that, you know, we, we used to talk about like a year ago around maintainability, like don't really apply anymore with AI coding. Like DRY, for example.

Jeremy: 23:53

Yeah

Ryan: 23:54

much more willing to be less DRY with AI.

Jeremy: 23:57

Угу

Ryan: 23:57

Um, however, I, I find other programming paradigms more important, like single responsibility as,

Jeremy: 24:06

Угу

Ryan: 24:07

a big blanket statement, right? But, like, if you have good seams, it makes it easier for the AI to work. Um, you know, so just the shifting of what, you know, good software means and, you know, what m-maintainable software, it like… That's the uncertainty to me, is

Jeremy: 24:26

Угу

Ryan: 24:27

I don't think we have good definitions of any of that yet. but I found myself the other day just thinking as like, you know what? I've actually… I haven't written code by hand in months.

Jeremy: 24:38

Yeah,

Ryan: 24:39

don't actually

Jeremy: 24:39

same

Ryan: 24:40

Like, the muscle's maybe even gone at this point,

Jeremy: 24:43

Yeah.

Ryan: 24:44

disappears quickly.

Jeremy: 24:45

Your, uh, your pinky's feeling better 'cause you're not hitting, like, c- control anymore? Like, it's like, all, all of a suddenly are not-- or symptoms are reversing.

Ryan: 24:55

Um, but I, I also find myself, I, I, uh, I distinctly remember this the other night. I was sitting there and I was like: I don't read code anymore.

Jeremy: 25:03

Yeah

Ryan: 25:03

not, not like I used to. Like, obviously I have to look at PR reviews and, you know, I scan stuff. But I, I used to, like, study code bases and try to figure out, like, how… You know, what is my style? Like, how do I wanna put this together so it's maintainable? I don't do any of that anymore, and I don't… You know, I enjoyed that part of the job, but it's completely r- you know, it's almost irrelevant now, uh, in my opinion. know, maybe I'm too… Maybe I, I've swung too far in that direction. But I'm curious what either of you, you know, feel about any of that.

Jeremy: 25:40

I think, well, where I am right now is I haven't written a line of code since probably January. Um, I'm still a pretty, um, uh, like, what's the word? Micromanagy, like manager over my agents, you know? Like, "Do it this way. I don't like that. That looks dumb."

Ryan: 26:03

do that too.

Jeremy: 26:04

Um, yeah, so I'm sort of like that. Um, I would say I read all of the app code and scan test code, um, which I can't decide if that's okay, if I'm okay with that or, you know, with doing that at this point or not, but that's what I am doing.

Ryan: 26:22

I, I'm glad you brought that particular piece up 'cause I feel

Jeremy: 26:25

Okay.

Ryan: 26:25

same way.

Jeremy: 26:26

Yeah? Okay.

Ryan: 26:27

it, it, just a, as a, a feedback point, um, I wanna read test code more.

Jeremy: 26:33

Угу

Ryan: 26:35

Um, I have a suspicion that it's not great

Jeremy: 26:38

Yeah

Ryan: 26:39

uh, in a lot of cases, but I don't know if I have the mental capacity yet to dive deep on that.

Jeremy: 26:45

Right. I'll tell you one thing that's maybe spicy. I think that if we're spending so much compute now on LLMs and automated verification, I would like to stop using fixtures and just do factories, because if I'm looking at tests, I wanna look at one test by itself and not have to think about the Harry Potter world system, world-building that's in all these YAM- YAML files and all these characters and places to understand if this one little test of five lines actually makes sense. Um, and if AI is generating more and more of the code you already have, and you're already using fixtures, I'd rather use factories and have all the, the world-building done right inside that, um, right inside that single test, so I can just dial into that and don't have to go look at agentic, um, YAML files

Ryan: 27:44

I'm, I'm gonna show my testing, uh, naivete,

Jeremy: 27:48

Ага

Ryan: 27:48

with this question, right? But, like, do we need either? Right? Like, with AI doing it, I mean, why not just build it in the setup, right? Like, e- especially if we're not gonna, like, read the code. Um, at least if you do have to read it, it's all in one place, and you don't have to then, you know, when you're reviewing it, you don't have to reconstruct, you know, the entire worldview, as you put it. Um,

Jeremy: 28:12

I think the one benefit there is that with, with one or the other, you can, um, you can establish which attributes of models are irrelevant and can be just defaults, where the test can then just specify the attributes that matter for that test. I think that's still beneficial, where it's like, it's nice to be able to say,"Oh, it set the, um, expiration date on this attribute of this model," because in the test, that's what we're looking at. If the, if the expiration at, you know, expires at is nil, then it means this, but if, when it's set to a date in the future, it means this, and that's relevant. But all the other attributes that have to be s- you know, added to save that model are irrelevant to this particular test.

Ryan: 29:03

Yeah, right. So it's more about getting the base case right,

Jeremy: 29:07

Yeah. I, and that,

Ryan: 29:08

off point. Yeah, that

Jeremy: 29:09

and that to me, that, that's my argument then for factories,'cause then you can just do that that way

Ryan: 29:15

So would you still use like the different, you know, uh, you know, traits and things that are common in factories? Or would you just sort of do

Jeremy: 29:22

yeah.

Ryan: 29:23

you

Jeremy: 29:23

I've done… Yeah, I mean, I've gone that route and then I think it's like, ends up being too far. I'd rather, I'd rather have a little bit more verbosity in the test itself. I wanna do, I wanna be able to get almost everything I can out of the, that one test I'm looking at. Uh, if I have to, if I have to read it, I don't wanna have to go look through 15 YAML files to build up the world in my head about what is happening here and why this is happening, you know? Um, and so I don't mind looking at maybe a, a factory if I have to. Um, but I'd rather be, I'd rather just be looking at that one test,

Brendan: 29:59

Well

Jeremy: 29:59

or maybe like a private method at the bottom

Brendan: 30:01

yeah, it's funny you say that 'cause that's kinda how I always felt about tests. I just

Jeremy: 30:05

Yeah

Brendan: 30:06

there anyways because you'd always get lost. Even, you know, with lets and things like that,

Jeremy: 30:10

Oh, no doubt.

Brendan: 30:11

like yeah,

Jeremy: 30:12

That's, yeah, that's our, our spec, the bad parts. Um

Brendan: 30:17

Yep. Right

Ryan: 30:19

are you in the, uh, you know, uh, uh, feature spec versus unit test, uh, debate at all n-now that you're in AI world? Like, do you favor one over the other? Has your view on that changed at all?

Jeremy: 30:34

I think I'm in so many different code bases, I'd forgotten, like, in some ways some things that I maybe used to had opinions about. Like,

Ryan: 30:43

Hmm.

Jeremy: 30:43

if you play in other people's, you know, um, backyards enough, you kind of forget what yours looks like. Um, it's… I think it's… Yeah, I don't even know anymore. I guess I tend to be more on the, like, yeah, you wanna have really good unit test coverage, and then, um, I, I'm definitely not a let's abandon system tests and things like that. I, I think they're really useful personally. I would still prefer having them. Um-

Ryan: 31:15

Yeah, I think that's one of those things right now that has just sort of take on, taken on a world of its own in the Twitter sphere.

Jeremy: 31:21

Yeah

Ryan: 31:22

like, "Hey, we don't need unit tests anymore. Everything's browser-based." it's like at face, at face value, like when you're thinking about, you know, AI doing it, like I, I can get, you know, the idea of it. I'm not sure it's a good one, but I

Jeremy: 31:35

Yeah, and I'm just not sure you can actually cover the entire surface area from that level. It's just the m- it is the most expensive way to do it too. Um,

Ryan: 31:45

T- totally.

Jeremy: 31:46

it feels like people are throwing so many dollars now at automate, the automatable parts of the system, and eventually that has to swing. I j- just think it's gotta swing. Not that we'll ever go back to a place where we're

Ryan: 32:00

have to pay like a going rate for AI credits, and then it will swing back

Jeremy: 32:04

Yeah, maybe.

Ryan: 32:06

Like this is

Jeremy: 32:07

I think people are pushing the, they're pushing the envelope on how much of everything we can automate, and I think that makes a lot of sense, but it's just funny compared to a year ago where, or two years ago, where everyone's like, you know, crying about their CI bills and it's just like, okay, you're just like, this is more compute in a different area. Um, I, I just feel like there's a ton of that. And now we're, you know, like all the other ways, like, you know, like, uh, when you're done with this PR, spin up, you know, 15 instances of Chrome and test, you know, like test, uh, every, uh, responsive breakpoint in the, in the browser, you know, for every happy path in the app or something. You could do that, yeah. You could do that before. Um, you know, you could have made a way to do that. Um, but now everyone has the appetite to do it. Yeah. Yeah.

Brendan: 33:00

Yeah, th-that is interesting too, and I think, Ryan, you and I have talked about this a few times as well, just like the, the line between AI and the automation. It's like, one, AI is non-deterministic, so, like, if you're gonna do something like you just described, Jeremy, probably gonna want that to mostly be an automation. And maybe AI helps write the code or the, you know, framework or whatever it is. But realistically, that, that is just an automation, so, like, to your point, we were doing that before. kinda funny though that people are now pushing on those things more

Jeremy: 33:33

Mm-hmm

Brendan: 33:34

In

Jeremy: 33:35

Yeah

Brendan: 33:35

make sense to me , but I, I guess I also get it. But

Ryan: 33:38

it's way easier to figure it out.

Brendan: 33:39

yeah, th-that… I think that's it

Ryan: 33:41

up like a, you know,

Jeremy: 33:42

Yeah.

Ryan: 33:43

pipeline before,

Jeremy: 33:44

Yeah

Brendan: 33:45

Right

Ryan: 33:46

GitHub actions, and like the only way to tell if it fails is to let it run for 30 minutes before it fails, right? And then

Brendan: 33:51

It's a fair

Jeremy: 33:52

Yeah.

Ryan: 33:53

and now you can just give it to an agent and go to sleep. the morning it's working,

Brendan: 33:56

Yep

Ryan: 33:57

I mean, some of it makes sense, right? That or

Jeremy: 33:59

Yeah

Ryan: 34:00

on it because the mental load is, is significantly lower,

Jeremy: 34:05

Mm-hmm

Ryan: 34:05

than it was before.

Jeremy: 34:07

and it, it does feel like, uh, kinda like everyone had with, um, what's it called? Like, uh, OpenClaw or whatever, um,

Brendan: 34:19

Oh, yep.

Jeremy: 34:19

where it was like everyone was test running that, like the, "Oh, I could have my own, um, Jarvis," or whatever. Um, and then that was having a moment, and then it just died. Um, I wonder if people are having this moment, like, to see how far they can push the idea of a factory, and if the e- seeing if the economics have tipped enough to get to that point where your app just builds itself and while you're asleep. Um, and then all the automation and becomes worth it at that point'cause you're not having to… You'd rather, you know, just spend, pay compute and tokens instead of paying people.

Brendan: 34:56

But,

Jeremy: 34:57

And maybe we find out it works and maybe we don't. I don't know.

Brendan: 34:59

Yeah. The thing I always push back on that, like I think there is some semblance of the factory that makes sense, right? Like

Jeremy: 35:06

Mm-hmm

Brendan: 35:06

gonna do a feature, that'd be pretty awesome if I could just tell the AI, "Do this," and then it goes all the way through, and it gives me that feature. at a certain point, like this is software for humans still. It's not s- I mean, sure, there is software for agents, but primarily it's software for humans. So

Jeremy: 35:22

Yeah

Brendan: 35:22

be a human in the loop at some point to say, "Yeah, that was actually what I wanted," not just based on the, you know, input and the, the AI saying,"Yep, that was what he told me to do."

Jeremy: 35:32

Yeah

Brendan: 35:32

that's where I lose the, the thread of like, okay, just let it run for the entire week, and then I come

Jeremy: 35:37

Yeah.

Brendan: 35:38

like what, what's

Jeremy: 35:39

You just hook it up to your customer support. That's the human in the loop.

Brendan: 35:42

I mean, true, if you have enough customers, I guess. But,

Jeremy: 35:46

Yeah

Brendan: 35:46

at the same time, like I would want a little more quality assurance on like

Jeremy: 35:49

Yeah

Brendan: 35:50

the right things before I get these signals back from customers like, "What the heck is this?"

Jeremy: 35:54

Yeah,

Ryan: 35:54

but isn't,

Jeremy: 35:55

no doubt

Ryan: 35:55

whole point of pushing the factory, right? Like, people are pushing because that's, that's the risk and that's, like, the limitation, right? Like, we, we don't yet know how to put guardrails on this thing or create a fully autonomous loop, so we're pushing on that thread, right? And we're getting better and we're getting better. I think ultimately what's gonna happen, in my opinion, is, like, that is gonna get really good. We're just gonna-- You're still, still gonna have human in the loop. They're just gonna be at a different point in the loop,

Brendan: 36:23

Yeah

Ryan: 36:24

I think you can get a, a long way with it,

Jeremy: 36:26

Угу

Ryan: 36:27

then, you know, the next, you know, everyone levels up at that point, and then hiring will loosen because, you know, bringing another person to guide a team of 100 agents or whatever, you know, the number is in the future, right? Like, that's a, you know, 1,000X, you know, 10,000X, like, you know, improvement, right? If you have one human that can run 1,000 agents, you know, and each agent is like, you know, 10 to 1,000X better than, you know, the human doing the tedious work, mean, that's huge returns for business, right? Like, you know, we got a long way to go, but I think that's why we're pushing so hard on, you know, the software factory and the, you know, the notion of, like, what automation-- what, you know, what CI and, like, guardrails can I actually put on this thing to make it work, right? Like, and,

Jeremy: 37:12

Yeah

Ryan: 37:12

far we don't… You know, may-maybe Shopify, you know, sounds like they figured it out to, to some extent, and

Jeremy: 37:19

It sounds like Fin. Yeah. Yeah. Mm-hmm.

Ryan: 37:22

has, has done the same thing. But, you know, how many man-hours did they put into that, right? They're huge companies with whole teams,

Jeremy: 37:29

Yeah

Ryan: 37:29

So yeah, maybe they figured it out, but the rest of us are still, you know, running around with our heads cut off.

Jeremy: 37:36

Yeah Yeah, and I, I guess I do wonder about like the, how many, when we get on the other side of this, how many, um, agents can one human manage in a day? I'm, you know, if I've got three work streams going with, you know, different Claude sessions right now, I'm burn- I'm burning out by the end of the day. You know, I'm like, I feel toast. My brain's like so toasted every, every day. Um, so

Ryan: 38:04

but is that because you don't trust it and you're like really having to babysit, you know, every agent and make sure that it's doing the thing that you think it is? Right? So like what I've found is the mental load of like, you know, it used to be I focused on one thing and I did that all day. Well, now I, I got three agents, maybe more running, and all I'm doing all day is one, two, three, one, two, three, one, two, three, right? But I'm trying

Jeremy: 38:26

Yeah

Ryan: 38:27

the same amount of information on each task in my head, right? So that for me, that's why I get… I'm, I feel the same way at the end of the day, I'm exhausted.

Jeremy: 38:34

Yeah, the one, two, three is what's so ti- tiring. Yeah.

Ryan: 38:37

Um,

Jeremy: 38:38

very satisfying. It's not very rewarding. I can do it, but it's tiring. So that's… I can do it. I can do three. I can't… I don't think I could do six

Ryan: 38:50

Uh, yeah, three. I'm trying to think. The other day I had four going, and that was a little bit tough, and then it, like, at one point it, it screwed something up and I was like, "All

Jeremy: 38:59

Yeah.

Ryan: 38:59

untangling four work trees," and I was like,

Jeremy: 39:01

Yeah

Ryan: 39:02

Uh, I, forget it."

Jeremy: 39:03

There's a point where I get where I'm like,"Okay, what were we doing here? Just please tell me what we- I can't remember. I can't even remember what we were even doing here."

Brendan: 39:12

Yeah.

Jeremy: 39:12

I can't even read anymore

Brendan: 39:14

I, I run into that a lot, the whole, not AI context, but, like, the context of what it was asked to do.

Jeremy: 39:20

Yeah

Brendan: 39:20

that. And then, like, I feel like I'm constantly asking the bot, or the agent I guess, to, like, what was this thing again? What are we

Jeremy: 39:27

Yeah.

Brendan: 39:28

Like,

Jeremy: 39:29

Yes.

Brendan: 39:30

hard.

Jeremy: 39:31

Yeah. I've never, like, never experienced that, like, where you're, like, losing the plot multiple times a day. I know people have that and, like, experience life that way, but I haven't, not until this year. You know? Like, that's not been my experiences with development that I like losing the plot multiple times a day on what I'm doing. Ugh. I don't love that. Um

Ryan: 39:54

I, I will say though, on Monday I had maybe one of, like, the best feeling days I've had in a long time.

Jeremy: 40:00

Угу

Ryan: 40:01

pushed, I don't know, 17 or 18 PRs into production over the course of a six-hour period,

Jeremy: 40:09

Nice

Ryan: 40:10

all in this sort of agent, you know, multi-agent

Jeremy: 40:12

Yeah

Ryan: 40:13

The only way that was possible for me, like mentally,

Jeremy: 40:17

Mm-hmm.

Ryan: 40:17

were all small,

Jeremy: 40:19

Yes. Yeah,

Ryan: 40:20

that

Jeremy: 40:21

I buy that.

Ryan: 40:21

Like,

Jeremy: 40:22

I bet that

Ryan: 40:22

it's like the opposite of what AI wants to do. It wants to give you the, the 10,000 line commit, right,

Jeremy: 40:28

Yeah

Ryan: 40:28

and like do it. But like, if you just like do the next thing, do the next thing, do the next thing, you get that… We were talking about dopamine hit earlier, right? Like you get that immediate feedback. Yes, it's in production, right? Observability is good. It, it felt like I did something, right?

Jeremy: 40:43

Yeah

Ryan: 40:43

I feel like slowly every day I'm getting a little bit better at figuring out how to like slice and dice this huge feature project

Jeremy: 40:50

Mm-hmm

Ryan: 40:51

into pieces that are actually deployable. And I think that's part of the software factory too, is like there's the mental, human mental side of it that we haven't figured out yet. And, you know, I think we all sort of feel like we're ADHD, right? Like jumping all over the place, can't focus.

Jeremy: 41:06

Mm-hmm.

Ryan: 41:06

At least, at least I do. I don't wanna put words in

Jeremy: 41:08

Yeah.

Ryan: 41:08

but that's, that's sometimes

Brendan: 41:09

Yep.

Ryan: 41:10

the day feels right now with, with AI assisted coding. Um, and it is not maintainable, I can say that. Uh, long term I, I'd probably go nuts before. So some- something's gotta change there, but

Jeremy: 41:23

Yeah

Ryan: 41:24

I still enjoy, you know, the end result, from a business perspective for me, right? Like I'm… You know, I, I code because I like building businesses, not necessarily the other way around. So if I can get a good business result out of something I do for a week, I'm feeling good no matter what, and that still drives most of my, uh, energy.

Jeremy: 41:45

Yeah. That's-- I think that's, uh… I wanna cue off that 'cause it's like one of the things that I think people have been saying about burnout recently is that it's not just about working too hard, it's like working hard without, um, knowing what it's for or having the reward, like a, the, you know, like realizing an out- the outcome that you wanted. Um, and I think for me, two things that I love about my work have been historically, um, making choices or building things that serve a person, that makes another person like, "Oh, thank you for fixing that for me," or, "That looks so good," or,"Yes, we finally got that feature out." Or I'm, or I'm learning something."Oh, let me try that. Oh, I wonder what will happen if this hap- you know, if I build this out, will people use it? How will this be used?" Or like, "Well, what'll I learn if I tried this technique?" And now I'm not doing-- I'm not getting much of the learning other than how do I, like, work with Claude better? Um, how do I prompt better? And that's like somewhat satisfying, a little bit. But because of this pressure to produce more faster, the appetite for getting, like, getting more done has gone up by, I think, by business owners, um, so that it's like,"Okay, great, you got this done." N-no one's saying this to me, but the way I feel is j- "Okay, now the next thing, and the next thing. Okay, awesome. Got that one done. Now the next thing, and the next thing." And it's not like ever like, um… I, I mean, it's never been a,"Well, it's always a new thing."

Ryan: 43:27

that way, right?

Jeremy: 43:28

Yeah, but-

Ryan: 43:29

an employee is. It just, it's just accelerated and it's more obvious now,

Jeremy: 43:33

I th- I think it's accelerated, yeah, and, and the, the enjoym- like the celebration of getting over certain milestones feels diminished by the, you know, ever-increasing d- you know, need to like get more done or something. I don't, I don't know, like this, the race.

Ryan: 43:53

thing in a day or two,

Jeremy: 43:55

Yeah.

Ryan: 43:55

feel like it needs to be celebrated as much as the big

Jeremy: 43:58

Yeah.

Ryan: 43:58

six months,

Jeremy: 43:59

Yeah,

Ryan: 44:00

you get a

Jeremy: 44:00

that's true.

Ryan: 44:01

done, it's like,"Hey, we're having… You know, we're going to Hawaii."

Jeremy: 44:04

Yeah. Yeah.

Ryan: 44:06

right? It's like,"Hey, it's time for vacation."

Jeremy: 44:08

Yeah

Ryan: 44:08

You know, you get a, a big project done in a weekend, which would've used to take six months, and it's like, No, you did your job. Great. Cool."

Jeremy: 44:16

Yeah

Ryan: 44:16

Right?

Brendan: 44:17

f- feels more like a treadmill now. You're, you're

Jeremy: 44:19

Угу

Ryan: 44:19

it

Brendan: 44:19

constantly onto the next big thing

Jeremy: 44:21

Mm-hmm. Yeah. Which, I mean, I don't, I don't on one hand don't wanna, um, sound like I'm complaining because there are plenty of jobs that are like treadmills where you, you know, you show up, you do a really hard thing, you go home, and then you show up and do the really hard thing again, day after day, and there's no letup. So I get that. It's been har-hard to realize that that may be what this job is now. Um, I would love it if it wasn't. Um, if it is, there are still enough things about it that I think I like that can make, maybe make it worth it, but it's not what it was, and that's still hard. So I'm still, I'm still dealing with that.

Brendan: 45:02

Yeah, that makes sense. Yeah, I mean, we- we've touched on this. I feel like we've kinda, we've gotten away from the consulting thing a little bit,

Jeremy: 45:09

Mm.

Brendan: 45:11

steer back towards where we were a little bit ago. Like, how, how is your business overall? You've touched on things here and there as we've been talking. you're managing three agents or whatever. Like, is that, is that three agents per client? Is that, like, each client, uh, y- you're running? Like, maybe

Jeremy: 45:26

Yeah

Brendan: 45:27

a little bit.

Jeremy: 45:29

So Usually I am working with two to three clients at a time, um, and any client engagement might be three months to three years, usually toward the years end of things. Um, so I'm usually embedded, taking on a role, not a specific deliverable or project, um, and working fractionally anywhere from 10 to 20 to maybe 30 hours. Um, I'm usually billing my capacity. I tell people it's around 30 total, um, maybe a little bit higher, uh, for certain periods, but not very much. And for me, billing 30 hours is working like 45 probably total when I count all the other things that I consider business activities. Um, so that's kinda like how I work right now with, with clients. Um, this year has been fairly stable. I've got two clients. One… O- yeah, two clients and a smaller project for a third client. One has kinda been-- I've been with them for a year and a half. Things are going well. I know the code base well, the te- you know, like the, the, the, um, the team and, uh, feel like I know what I'm doing. In the other one, I just, just started it, like just onboarded this week. Um, so brand new. Uh, often for me, I'm coming in and trying to form fit myself to the client and their process. You know, like I'm not coming with,"Here's how I do things," or, "Here's the, you know, project management tools I use, and you're gonna have to adapt to me." I'm adapting to whatever the situation is for the client, um, learning their systems, trying to become an embedded member of the, the team just in a fractional role so that they feel like they're, um, they're not having to do… I don't want them to have to do extra thinking to like adapt to my way of doing things. That's my hope at least. Um, so depending on the team, it can look different, yeah, like whether I'm in Basecamp or, uh, Trello or, you know, GitHub Issues. Uh, from that standpoint to whether I'm involved in operational responsibility with like Heroku or, uh, you know, do I have ability to deploy or not? Do you want me like watching the server over the weekend? That kind of stuff. So that can vary. Um, and the processes… Process used to be more straightforward, and now it's like everyone's coming up with very different processes for… And they're-- we're all like morphing our processes."Okay, here's a dif- here's a different way we could do it. Oh, what if we did this instead? What if we automated this? Actually, I don't like that. What if we used this code review tool?" You know."Oh, don't like that one. Let's try this other one." So all that stuff is just like changing a lot. But I'm trying to adapt to like what my clients are doing and their, their, uh, approaches to things. Um, I don't know if that's painting enough of a picture. Uh, in terms of like The day-to-day work, it r- it can vary. Like I'm o- I'm often, um The most senior or close to the most senior Rails developer on, if there is a team, you know, sometimes I'm just working solo, uh, with no one else technical.

Brendan: 48:48

Sure

Jeremy: 48:48

is another technical person, I'm often more, maybe more senior than they are. So it might be I'm helping to improve code quality or help with architecture challenges or solve, you know, um, Honeybadger's like got tons of exceptions, let's work that down, things like that. Um It c- it could also be like finishing migrations, like finishing upgrades, finishing, you know, dependency migrations. A lot of times it's feature development. Um, so it can be a, a real mix there. Sometimes they even do like, um, more roles like, um, more on the front-end design side of things or UX or even PM kind of roles. Um, but typically it's back-end Rails, well, or like full stack, you know, own this feature all the way through,

Brendan: 49:45

Yeah

Jeremy: 49:46

h- handle this major migration that we're doing, that ki- that kind of thing.

Brendan: 49:50

Gotcha.

Ryan: 49:51

stuff

Jeremy: 49:52

Yeah. I mean, I like, I like variety, so, uh, I wanna kinda do it all. I also wanna be really close to the business, close to the metal. Um, I don't wanna run my own VPSs, but close to the operations, I guess,

Ryan: 50:08

Yeah.

Jeremy: 50:08

I can be. And,

Ryan: 50:10

Well, it helps with motivation, like you said before, right?

Jeremy: 50:12

yeah

Ryan: 50:13

part of the vision of what you're trying to build, it's different than just delivering a piece of code.

Jeremy: 50:19

Yeah. I don't, I don't wanna just solve a puzzle and push it through, you know, a dark doorway and move on to the next puzzle. I don't… Like, I need to see somebody's face, like,"That's a good, you know, a good job." That's what I feel like I need.

Brendan: 50:34

Sure.

Jeremy: 50:35

yeah. And we were talking about that, like, being like maybe not about the consulting, but honestly it is, 'cause, like, if the engine of my business is me and my mental state and the-- my ability, you know, like, my business' ability to, like, do well for myself and my dependents, like, is on, like, my-- how I'm doing mentally, my excitement, my, you know, ability to sell s- my ability to be confident, to sell services, to, uh, show up with energy for whatever situation there is. Like, my mental state is super important.

Brendan: 51:12

Yep

Jeremy: 51:12

Um, and I don't have a manager that's like, "Oh, you should take a mental health day." You know? Like, there's not a person, there's not a per- y- clients, clients can be wonderful people, but they're not your boss. They're not, or they're not your manager. They're not there to do one-on-ones with you. That's not, there's… You should never have that expectation. So it's you doing the one-on-one with yourself or figuring out how to get what you need, you know, for you so that you can show up for the client the next day or, you know, uh, to keep moving your business for-forward. And that takes more, um, it takes more, uh… What's the word? deliberate choi-- you know, action or choice than I wa- was thinking it would. I thought it would just take care of itself, but it actually is a, it's takes more deliberate choice to like say, "No, I, I need to do this for me." It's not enough to… Like this is, this is a job to be done in the jobs of

Brendan: 52:13

Yeah

Jeremy: 52:14

running a one-person freelance business

Ryan: 52:16

here's your permission to, uh, take a mental health day every once in a while. It's good for you

Jeremy: 52:20

Yeah. Yeah. It's been a little while since I've taken one. I could use one. Uh, maybe October, maybe October m- is my month. That'd be good.

Ryan: 52:29

It could be a m-m-mental health afternoon too. I,

Jeremy: 52:32

Yeah. Yes.

Ryan: 52:33

just, the other day I

Jeremy: 52:33

Yeah

Ryan: 52:34

2 o'clock, I went and got a beer and a, a, a, a burrito and came home and took a nap before the

Jeremy: 52:39

Yeah.

Ryan: 52:40

Like, it

Jeremy: 52:41

no, it's…

Ryan: 52:42

hours, but I

Jeremy: 52:42

Yeah.

Ryan: 52:43

know what? I needed that today."

Jeremy: 52:46

I- I'm the kind of person that has a really hard time with that. Like, yesterday I had, like, so many things going on, and then I'm recovering from… I got sick at Rails World, so like got s- like trying to get over that, but had like onboarding with this new client, and by like 3:00, I- my brain was just shutting down and like I had the worst headache, and I'm like, "I just, I gotta tap out." I never tap out. I, I sl- I just like laid down and like zonked for two hours, and I never do that. I still felt guilty about it. And I was like getting over sickness, so I like, I really struggle with that, um, 'cause I'm like, "Oh no, I need… I didn't hit my build, you know, my build, hourl- hourly build goal for the day." You know, like, "Ah, it's gonna mess up my week," you know? Um, uh, so I have high expectations on myself about that. But like, yeah, your, like your body's gonna respond. Eventually, if you don't take care of it, it's gonna come out.

Brendan: 53:45

Yep

Jeremy: 53:45

will have to be, and it'll, it'll be n- often not in ways that you wish, you know. Like it would be easier just to manage yourself earlier, you know? Uh, yeah.

Brendan: 53:56

Yeah, no, I mean, those are important things in general, for sure. And, and yeah, I mean, just to tie it back into what you were talking about earlier, like all, all the strain on all of our mental health with the way AI is accelerating things, you, you do need to be aware of that stuff, for sure.

Jeremy: 54:09

Yeah

Brendan: 54:11

I'll, I'll, I'll make sure I send you messages from time to time and be like, "Jeremy, did you take a mental health day yet?"

Jeremy: 54:16

Yeah. Oh, yeah

Brendan: 54:19

Well, speaking of, uh, mental health and, and resting and such, you guys are both on the East Coast, and it's 5:00 on a Friday, so m-

Jeremy: 54:27

That's right

Brendan: 54:27

not a bad time to just start wrapping up. Um, anything else you wanted to talk about or plug before we do kinda jump off here?

Jeremy: 54:36

Oh man, I'm trying to think I am, uh, I wanna go back to what Ryan was saying. Like, I am still very positive about the future of consulting. I, I really do believe that there's no end to the needs that humans have, which means there's always capacity to come in and serve needs. It might look very different, but I like finding ways to s- meet, you know, to solve people's problems, to meet their needs with computers. That's always been fun to me. I believe there's always a world where that will be needed. It's hard to figure it out right now, but I, I don't think that goes away. Um, I think this has been a super volatile year, so it's hard for people who might wanna jump in this year, but I don't think that will always be the case. So my hope is maybe it's next year, maybe the year after that, things settle down a bit and then it's more clear. One thing I'd be really excited about is I've been working fractionally for 13 years. I really like the style of work where you're working long term with a few different clients, and there's benefits on two sides. You-- On my side, I get to spread my risk and get more variability, um, in my day-to-day work by working with multiple teams. On their side of it, they get to work with someone who is fairly senior, um, but don't have to pay a full salary for it. Um, so I think there's benefits both ways, and I really, for years I've thought, I think more should-- people should do this. I feel like you could do this with, um, you know, content marketing and design and UX and customer support, and so many other things could be this way. I mean, your account, uh, your-- most people's accountants are this way, like you, or your lawyer. Like, these are things that are typically on, you know, their retainer or their, you know, regular services. They're, you're paying them, you know, a few thousand dollars a year potentially, um, for these services. I feel like there's more opportunity here and with, with all the stuff that's come with AI, I feel like the opportunity for fractional businesses and, uh, solopreneurs, uh, just grows dramatically, what they could achieve on their own with the systems they develop that they could bring to, uh, other businesses. And I think people, businesses, especially right now with the, just the way the economy's been, are looking to save money and are looking to, um, reduce their own risk. And one way to do that is to reduce headcount, but then work with more, like with a network. So one thing I'd love to see is just more people doing this kind of stuff, and it's something I've wished to make popular at some point. I, I don't know how to do this, uh, but I would love to, um… I'm always, I'm always up for talking to people about this stuff, but I'd love to at some point, like figure out, like, how could I help more people build fractional businesses?'Cause I, I love it and I wanna see more people. I think there's just so many services you could offer that com- like companies could be like, "Oh, yes, we would pay two grand a year to have someone that did this, that filled this role, and it's a role we always need filled. It's not just like a one-off, you know, problem solved, but we need someone who's thinking about this. We'll pay this much for our business, you know, just thinking about this part of our business permanently." Um, I like that kind of stuff. I like the long-term relationships, and I feel like it's also one of the easier ways to be in business for yourself, where you're not having to work on an endless, um, sales cycle, um, sort of never-ending pipeline, um, fini-finishing one project every, after the next. Um, so that's a totally different topic, but one I'm, one I'm very interested in and I feel pos-- have po- you know, sort of positive, um, feelings about for what, what the future, uh, could look like there.

Ryan: 58:33

Well,

Brendan: 58:33

Cool

Ryan: 58:33

to come back on and 'cause that's a whole nother conversation that I'd love to

Jeremy: 58:37

Yeah.

Ryan: 58:37

Um, it's intriguing. But in the meantime, you wanna direct, people anywhere specific to find you, IndieRails

Jeremy: 58:45

Yeah. Uh, so my friend Jess Brown and I, we, uh, host the IndieRails podcast. So, yeah, you can check that out. Uh, I have put on the Blue Ridge Ruby Conference in Asheville, North C- Carolina. That's coming back, tentatively. We haven't announced any dates, but, uh, keep it, you know, keep this hush-hush. But, uh, I think it's coming back in 2027, which is pretty cool. I'm-- I'll probably be in, like, an advisor role this year, but the team is, like, gung-ho to take it forward, so I'm really excited about that. Um, let's see. You can find me on most places at, my handle's jeremysmithco. I still call it Twitter, so you can find me on Twitter. And, uh,, what else? Oh, and then, my tiny little, you know, one-person, studio is called Hybrd. It's H-Y-B-R-D dot C-O. Um, yeah, I think that's it.

Ryan: 59:34

Awesome. Well, thanks for coming on. Really enjoyed the, uh, conversation, and hope you have a great weekend. Thank you

Jeremy: 59:40

Thanks. Thanks to you both. I appreciate it. It's been great.

Brendan: 59:42

Thanks,

Jeremy: 59:43

Mm-hmm.