[00:00:05] John: Hello. You are listening to standing in the fire where Kris, John, and Garrett again or Garrett and I might just start a new podcast, and then he'll be on that one every time. We'll see. The owners of fireside.fm and a couple of other SaaS talk about what it's like to be in the heat of SaaS growth. So today, we are we're gonna go positive from the beginning. So we're not gonna jump into what's been hard and messy, but we're gonna jump into, like, what's been, like, easy or great or what's been fun. So anything been fun recently?

[00:00:36] Kris: I got asked that last night for my mom.

[00:00:40] John: She's like, what's good that you can set? What's oh, no. Is mom listening to the podcast now? Oh, gonna watch the language.

[00:00:49] Kris: So what what's been fun? Don't Do do anything on top of mind?

[00:00:54] John: Yeah. I mean, for me, what's been fun is I I kind of like, mentioned it a few episodes back that I was going through, like, a little bit of, like, is AI, like, bothering me? Am I getting more stressed because of it? Stuff like that. And, like, I I basically realized, like, it was not that at all. So literally, I'm like the title of this episode is like, it was Flonase. Like, for whatever reason, yeah, my my body just reacts strongly to, like, any kind of medicine. Like, taking over the counter medicine, I wouldn't think, and was definitely taking more than I was supposed to. And, like Yeah. That was basically it's a steroid. So it was it was juicing me. And I I'm not one for steroids as you can tell from my arms and and various, you know, sizes. So so, like, I've been doing a lot more. I'm kinda back on the train.

[00:01:36] Kris: Cool.

[00:01:36] John: Running, like, you know, anywhere from, like, I would say two, three, four things at a time and pushing them forward and then, like, you know, three or four things that are back burners that, like, if all of those are busy and I'm just sitting there and, you know, ready to go, I'll I'll kinda tinker with those. Mhmm. And then and that's worked really well, I feel like, just getting, like, back to that rhythm. And and I've also like, I I posted online the other day, like, my GitHub contributions were, like, this, like, dark green for, like, three months. And now they're, like, a little more, like, light. Yeah. And, like, there's space on the weekends and stuff. So, I you know, Flonies is not the only thing I changed, but that that was the you I cut that in half and immediately felt, you know, that much better Yeah. And then cut it all the way down. Now my nose is stuffy, but I feel way better. So I'm like, I can deal with a stuffy nose. That's fine. So that's been one of the great things for me is just getting figuring out what was bother what was bothering me physically and making me feel, like, not right. And now I feel right again. So I've been playing basketball and doing all the other things. Oh, nice. That's been fun. So

[00:02:39] Kris: Yeah. I was thinking of now, actually, I'm gonna pull it up to read. Except because so I was trying to say, like, sometimes I'm I'm good at just focusing on the problems. But Mhmm. I got an email yesterday from we'll say Alice is her name, and this is from one of the companies we have. She just wanna say thank you for the great job on these shirts. Kids love them. They love the design. Someone apparently told the kids that out of all the shirts they saw at the park, they liked ours the best. Just saying thanks. You guys rock. So I think, like, getting those types of messages, I'm like, oh, like, it's it's to make other people's day Yeah. And to make, like, their trip or what you know, whatever event they're they're doing. I think those are always, like, helpful to keep, I don't know, fighting the good fight.

[00:03:27] John: Yeah. Well, I mean, I think where that comes from is, like, everyone wants validation. That's what you want. But, like, what you typically what everyone chases is typically revenue, money, profit, things like that because that's the easiest thing to validate. Right. You know? And I think we're less good typically as entrepreneurs at, like, collecting that kind of stuff. And then when you're having, like, a rough moment, just, look back through it. Yeah. I I don't have a single, like, I love your piece of software saved. Oh, yeah. Yeah. Yeah. But, like, every time I, you know, like, do something to improve performance and get a drop off graph where, you know, time consumes way down, I save that. I'm like, Dexter, you know, I've got my little rolodex of like graphs where I've killed milliseconds. And I'm like, but I should I just when you said that, I'm like, I should be Yeah. Screenshotting those and putting that kind of stuff in a folder too because it happens.

[00:04:15] Kris: And

[00:04:15] John: and the there's one this morning. I think you met Josh. Have you met Josh? I know a lot of Josh's.

[00:04:21] Kris: From here, our new our new tenant here.

[00:04:23] John: Oh, yeah. I think I met him.

[00:04:24] Kris: Yes. Yeah. Yeah. You know, they were walking in and

[00:04:26] John: Oh, yes. Yeah. I know him.

[00:04:27] Kris: Yep. And and so I said, oh, like, how you liking your oh, like, we love it. Like, it's so awesome. Like, ah, like, it's I I I maybe I don't know if it's ego that you need some of that validation. I think this is probably it's normal.

[00:04:38] John: But that's why you do it. You do it for that. So I I think that's totally normal. Don't even think it's ego. It's like, why would you do something if ever if it was horrible all the time? You know? Like because there's it's not easy. Mhmm. Like, you're not going through easy things. Yeah. You know, right now with Momentum, with other stuff. It you know, it's not easy. So those those are the things that yeah.

[00:04:56] Kris: So so I think, like, those two are just like, oh, like, cool. And then then if we can knock off a couple more, like, projects, tasks, that'd good. But, yeah, I I think getting in the last couple days, like, getting that validation from people has been super helpful.

[00:05:11] John: Yeah.

[00:05:11] Kris: That's awesome.

[00:05:12] John: Positive. Love it. And, also, that's access when you're talking about the t shirts. Momentum does not also do t shirts.

[00:05:19] Kris: Yeah. But if you need t shirts, it's

[00:05:20] John: screen printing or embroidery. Boom.

[00:05:22] Kris: Accessimprints.com. Yeah.

[00:05:24] John: We'll we'll do like a YouTube link or something. We'll figure that out. So Nice.

[00:05:28] Kris: We can figure out sponsors. Man, they should we can do it every There we go. Yeah.

[00:05:31] John: Yeah. We'll be like, brought to you by Speaker Deck. Share your presentations without the mess. Yeah. Actually, caught me kinda funny because I was listening to Moneywise, which is Sam's other podcast. And it cracks me up because, like, everything on that podcast is not even necessarily related to money. Yeah. But, like, he'll kind of make it for the ad rates and stuff. Because it's like like My Body Tutor or My Body whatever. Yeah. My Body Tutor, think is what it And like Kubera. It's like all these apps that he uses. He basically was obviously just went out and was like, look, I love your stuff. Let me rep it and, you know, probably work out some kind of deal. And then Hampton is the the one that gets interweaved

[00:06:14] Kris: Yeah. Everywhere as well. See, that's still I mean, we should maybe we should think about doing that with our own stuff. We can. We can do some ad reads.

[00:06:20] John: I mean, now we're getting, you know, thousands of downloads into that level. Pretty soon, we're gonna get like a little little pin. So Nice. Big time. Nice.

[00:06:29] Kris: One of our own pins?

[00:06:30] John: That would be good.

[00:06:31] Kris: Like, not in Fireside.

[00:06:32] John: Yeah. When you when you log in, I don't know if

[00:06:33] Kris: do

[00:06:33] John: you remember? Do we still have those? I I don't know actually if we do or not. Yeah. But I know we did.

[00:06:38] Kris: We did we did have them. Don't know with the new studio if we have them. Yeah.

[00:06:42] John: Well, that's actually a a good thing that we should probably talk about what we haven't yet is the new studio. Oh, yeah. Like, it's out and Yeah. It's working. I think everybody's on it. What are your thoughts now that you're using the new studio? Obviously, not perfect. Well Yeah. But how's it going?

[00:06:56] Kris: So I mostly started using it with the new studio versus the old one. And so I I'm both learning the trials and tribulations of, like Yeah. Podcast people and producers and publishers, as well as just, I wanna get this out quickly. Yeah. So I think there's lots of tweaks that as a if if you think about it from someone who's not technical and not an expert, like, that's part of our audience in in the software

[00:07:25] John: Yep.

[00:07:26] Kris: I think we have a good structure to get it to where to make it super easy. Yeah. But as you've noted, or I've I've noted to you, I have my laundry list of things like, I'll give this was tweaked, this was tweaked, this was tweaked. It would make it more intuitive for someone who is not, like, used to the workarounds for Yeah. How something would work.

[00:07:46] John: Yeah. Agreed.

[00:07:47] Kris: But I think it's yeah. I think it's gonna hopefully help people streamline and save them time with with it. How about you? Do I don't I don't know if you've used it as much.

[00:07:56] John: I've used both.

[00:07:57] Kris: Yeah. Yeah.

[00:07:57] John: Yeah. Because I I was so I previously did, like, all the show uploads and stuff. Kris started doing it the last, I don't know, three or four or something. And so so if the titles are getting better, that's why.

[00:08:07] Kris: Or worse.

[00:08:07] John: Or worse. Yeah. And so, like, we so we kinda handed that off. So and that was right around when the new studio came out. And for those who don't know, the new studio is just like, that's where you go to, like, manage your podcast. We're calling it the studio, blah blah blah. I think, actually, YouTube calls it that as well, the YouTube studio. They do. Yeah. I don't know that we intentionally Yeah. Moved it because of that, but that's kinda funny. So I've used both. I like the new one dramatically better. Obviously, there's we didn't change any workflow. We did we kind of did like, there's a lot of improvements that improve workflow Yep. In other areas for like setup, like domains, connections to, you know, the various Mhmm. Podcast directories and things like that. I think that stuff's a lot more intuitive and better now. But there's still a lot of things like in just like the new episode flow Mhmm. That's kinda overwhelming. Yeah. So we're look what we're our goal is, again, we're we're like workflow people. So the goal is to just slim that down. But you don't wanna slim that down when you're running two. For people that don't know, we have also have a feature flag product, and that allows us to roll things out like this pretty slowly and make sure everybody's really comfortable. And we did. There's still gonna be a couple holdouts that are just no matter what you do, they'll never be happy with change. Mhmm. And so you don't really build for them. You build for the people, everyone else. And so the feedback from other people has been, I think, really good. But for a while, we ran both the old studio and the new studio, which is a little bit of a development wait. Yeah. And so, you know, got everybody comfortable, which I think it was probably pretty good because then a few people hit bugs instead of everybody hitting bugs. And around the same time, Garrett took over primary support as well. Mhmm. I think that was actually a good timing thing too because for for two purposes. One, like, you know, you started using the new studio, which means you have feedback and stuff like that. And that's a good thing. And then also Garrett started doing support, means he started seeing what were the common issues and going after those. And so that's I think those two things both are gonna make fireside dramatically better long term. So Yeah. Yeah. I do. Nice.

[00:10:04] Kris: And it's been hard lately. I I feel like it's we we got similar challenges that we had last time. Just, I I would say, trying to be better at I'll I'll I'll summarize it maybe differently of being better

[00:10:23] John: delegating

[00:10:24] Kris: and trusting. Yeah. And it is not that I don't trust people. Like, people that I work with, trust. But it's it's that the the thought of that it's not gonna be done the same way that I would do it Yep. Or the with the urgency, maybe, that I would have it. But needing to, like, just keep trying to give other people more responsibility so that I can do the other things that I should be doing, but I haven't been doing. So that in three months from now, it's not sitting here to having the same conversation, and we're in a worse spot because I was trying to do everything. Yeah. Yeah. I mean, I think we just naturally used to doing everything, problem solver, and you know exactly how you wanna do it. And it's in your head, but like somebody else, it's not in their head the exact same way. It's not in the same little different. You know?

[00:11:11] John: And you just have to be okay with that. And if there's if there's something that's not negotiable, then that just needs to be clear from the front. And I think that even goes back to, like, dare I say our most popular clip. I don't know. We were just talking about a second ago of, like, the the verification for AI. I think it's the same thing for humans. You know? And so if there's some of those that are that are things that, in your head, you're like, this is definitely how I want it to work or how it should work or something like that. Or here's the end result. We definitely want, you know, getting that stuff transferred to the other person Mhmm. Is just that's really important. So I think that's definitely something that I've actually been spending a lot of time on the last month or so as well. It's just like delegating more things out, not because I'm high and mighty or fancy, but just because a lot of the stuff I was doing was like, I was kinda just doing all of the maintenance for everything. Mhmm. It's like every exception, problem, customers, work, things like that.

[00:12:05] Kris: And is that your best I I was talking with someone of like, what's my there's, a phrase used, but, like, what's my best use of time? But if there's another adjective you Like, put in there

[00:12:16] John: your core, like, your

[00:12:17] Kris: Yeah. Like like like, best and highest use or something, and Yeah. It's it's not I would say, if I tracked, which I know it's probably a good thing if I did, how I spend my days, I bet half of the stuff I shouldn't maybe be doing. Yeah.

[00:12:33] John: Yeah. Yep. Agreed. Do we have we have we have a new segment. I don't know if either of us have any today, but it's called sparks fly. So this is hot takes. Mhmm. And so it's like, you know, what I mean, I think hot takes pretty clear. So Yeah. I I have some hot day. We should okay. Hot takes should be totally sponsored by, like, the Dave's fire chicken or whatever, like Dave's hot chicken. Yeah. Hot chicken. There you go. Spicy. Whatever. Yeah. That they just popped in my head when I said that. But yeah. Do you have I have a hot takes, so go ahead. Yeah. You got one. Go ahead.

[00:13:05] Kris: The the one I'm thinking about is and I I think there's there's a lot of designers that might agree with me. If they don't, that's okay. But AI art Oh. Ugh. Like, it is for some of the stuff is like my nemesis right now is because everybody thinks they're a designer and then they for example, like the teacher shop, they're like, oh, here's my design, and it's a rasterized JPEG Yeah. That has a background Yeah. With a million colors Yeah. In the wrong resolution. Yeah. And they're like, oh, like, make this make this into a shirt for me. Yeah. And it's it's I I I've seen some, you know, videos on Instagram where it's just like like comedy ones where there's like no AI art and this guy, like, made a song about it because it's so terrible for, like, printers or, like, apparel folks. But and and then they think so not only will they do use AI for art, they'll send, like, 40 different things and be like, I want 40 t shirts, all of them with a different design. You know, and, that was never, like, three months ago, like, that's not things we saw. Yeah. Right? Because you would get you would work typically if you got now, often you'll get pixelated PNG.

[00:14:26] John: But if it's actual, like,

[00:14:28] Kris: logos and stuff, you just get it vectorized and, like, move on, like, that's an easy thing to take care of. But, yeah, this stuff, it's caused us just, a lot more work, and it's not worth it's not high value customers necessarily.

[00:14:41] John: Yeah.

[00:14:42] Kris: So, yeah, my hot take is if you're don't you're not an artist just because you used AI for art.

[00:14:49] John: Yeah. Well, and it sounds to me I I hear a couple things in that too that are interesting. Like, one is there's tedious work around, like, productionizing an asset that was given to you

[00:15:01] Kris: Mhmm.

[00:15:01] John: That's not ready for print because, again, AI has lowered the barrier. Yep. It's letting more people in to create things like they want. And so now you're getting less ready things to go. Yes. More it might have been before it might have been like, here's some text and here's a logo, like, make it happen. Now they're like, oh, no. I've designed it. You know? And so what what I would what what I immediately think about that is like, Claude has this idea of, like, skills. Mhmm. And so you can make skills that will do specific things. And, like, you probably have steps to take a thing like that and then turn it into something. So, like, the first step might be to, like, vectorize it, you know. And, like, there's some pretty good tools out there that have APIs and Claude could probably walk someone through, like, here's how you sign up for the API, here's where you put it, the key, and all this kind of stuff so that you could use, like, know, the Mac desktop app, but Claude code. Yeah. And it could write a thing that, like, you know, you do it, run a skill. Like, Claude help help me make a skill to vectorize this. And then help me once it's vectorized, like, here are rules that we go through to to say, like, reduce the number of colors. Like, how can I reduce the number of colors? Yep. Reduce, you know, whatever. Get the right dimension. Well, vectorize, you got the right dimension. But but but and and even, like, removing

[00:16:12] Kris: like, what what we're seeing lately is, like, just, like, an eight and a half by 11 size of pure design. Yeah. So there's no negative space. Yeah. So if you do, like, a a DGF transfer, it's just gonna feel like a sheet of paper.

[00:16:28] John: Yeah.

[00:16:28] Kris: Because there's no there's no breathability

[00:16:30] John: on it.

[00:16:30] Kris: And

[00:16:30] John: so It's all like that plastic Yeah. It's it's good.

[00:16:34] Kris: Like that because Yeah. It's like it can that can feel good, but it's when there's outline when they're using the shirt as well as like part of the negative space. Yeah. Well, a good example, like

[00:16:43] John: this this is I'm taking my top off, so I've shaved something in my chest. No. A good example is like this. Like Yeah. You've got negative space in in all

[00:16:53] Kris: of our information. But even in the red. So, like, if you if you look at the red, you got the the white part.

[00:16:58] John: A little bit of readability in there.

[00:16:59] Kris: And I and I think, like, that's not what's happening with AI art. Yeah. And I would say it's it's not coming from, you know, designers or from companies. You know, it's it's more from individuals. And then, you know, separating out, like, company work versus individual work. And so, you know, Mike was saying, like, kinda like, we should we should just say no to all this stuff. And, well Or just raise the prices. Yeah. We we could raise the price. And, you know, if if there should be more of that, can we figure out tools to help like, this is what we can offer if if this is what's gonna come in. But I think raising the prices for sure, a part of it because yeah. Like, we we had we had some of it, like, maybe a week and a half ago, and I was spending and it was for, you know, like, 40 or 50 shirts, and, like, all these different artworks. The amount of time I spent, I was like, I'm not I can't I can't this is not worth our time. Yeah. Because at the same time, we had 200 shirts with a little bit of artwork that needed to be like done.

[00:17:59] John: Slap it on.

[00:18:00] Kris: And and you just like and you just crank it out. I'm like, this is not this this is not profitable.

[00:18:04] John: Yeah. And I think but that that's definitely something that you can factor into the price. Yeah. And then it's so it's like, look, I I can't just give you a I can give you a standard quote for, like, a one color, two color, whatever. Or if you have, you know, something custom, like, maybe that's a checkbox on the intake form. It's like, did you use AI to make this on your own? And it's an automatic It goes it goes to the trash. Like because I'm like, that's a real thing. But I also think if you if you took here's the last five that come in that were like that. And here's what they started as, and here's what they ended as. And then you think through what are the steps or ask Claude to ask you the steps or GPT or whatever. Yeah. And then you you turn those into it's just an SOP. You turn those into like an SOP for for like either a person or a tool Yeah. To do that. Because then, once you have that, I mean, I feel like you could probably just do, you know, slash, like, t shirtify this for me. Yeah. And it

[00:18:56] Kris: And you and you could do now we have to keep like, there's other applications that if you wanted to make it easier, you could do, but some of it's like new machines. Like, if you wanna do direct to garment, you could probably take the artwork and it's it's it's essentially printing it like a printer on a shirt. Yeah. But that's like, I don't know, I'm making up like 10,000 Yeah. Piece of equipment that you're also doing like one off shirts for, so, that doesn't make any sense. Sublimation, we do have that sublimation machine, which we haven't used in a long time, but I wonder because, see, people don't want like, sublimation works good on polyester and, like, white shirts, and so that would be the only like, that's you're essentially printing on a piece of paper and then adhering it to it Okay. Without the paper. It's it's Oh, gotcha. The ink is just the part that stays. Yeah. But most they're like, oh, I want a gray shirt and a black shirt and a yellow shirt. Yeah. You know, they want like one off custom stuff. It's kinda like, go to the mall.

[00:19:50] John: Nobody uses bleach anymore. Yeah. Like, it it like, that's the only way to have a white shirt basically. It's some kind of like, you know, this is sponsored by Tide. Yeah. Yeah.

[00:19:59] Kris: Do you have any hot takes?

[00:20:01] John: Oh, yeah. My hot take is

[00:20:02] Kris: I went on I went on a long rant on that one. So maybe maybe this is a good segment.

[00:20:06] John: Yeah. This is really the it's we're standing in the fire. Our listeners are sitting in the complaints or whatever, the rants, sitting in the rants. So my my hot take is I I've been totally anti, you know, Codex and and OpenAI the last while, and but I've started to use it a little bit. Matt Belcher was talking about how he explained

[00:20:27] Kris: Yeah. Explain maybe a little bit the Codex and OpenAI versus what you've been doing.

[00:20:31] John: So, you know, you have OpenAI and you have Anthropic, and then you have Codex and you have Claude. So the the two, you know, AI companies and they're like, here's how you write code with our models. So what most people talk about the models is they say inference. So it's a way of like it's like, this is the brain and then the harness is codex or it's Claude. And that is how you use the brain to achieve a task which would be like coding and stuff like that. So I have historically used Claude code which is the harness and Claude which is, you know, the the models and stuff. But I got a free open source, like, six months of that, of the max account. So I was like, I'm basically not paying $200 a month that I was gonna pay, so now is the time I should try GPT as well. And so I started kind of kicking the tires on that as well. And, again, like, they all they're good at certain things. They're bad at certain things. Like, I think the key thing is what you wanna do is you wanna use all of them as much as you can, and then you know what they're good for and what they're not, at least every, like, month or so. Because they change pretty rapidly.

[00:21:33] Kris: Okay.

[00:21:34] John: Yeah. Yeah. So what I've been finding is my a hot take before was, like, I hated Codex for everything. It just felt like the nerd in the corner that everybody hates but can solve the hard problems, you know? And but so I just kind of ignored it. Yeah. But I've had a couple issues where like I couldn't quite I felt like Claude was suggesting things that were not really the root problem. And I was like, don't think it's the root problem. And so it would go deeper and it'd be like Yeah. I'd be like, no, it's not that. And then it would go deeper and I'd like, it's not that. But like we weren't getting to where like something that I felt like it's like yesterday. It's like these weird Facebook and Instagram working with them via Box Out to, like, share posts and stuff like that is probably the hardest part of Box Out. They just fail for random I swear they just do, like, random five digit number error code. Like, it's like, you know, error 9,004. And I'm like, do you really have 9,000 unique errors? I don't that's insane. So yeah. That seems like a problem. And so, you know, we've been getting these weird and people just spots. You know, people have problems. They disconnect, reconnect. Sometimes that fixes it. Sometimes it doesn't. So it's just these weird things. And so Claude suggested a couple things, and I was like, those seem like they could cause some issues, but there's also ones that succeeded that were that had those issues. Yeah. So I'm like, it wasn't that. Yeah. So I ran it through GPT with the same context that I had given Claude. And it was like, these are a 100% temporary failures on Facebook's side. So the Instagram API is through Facebook as well. And so they were like, you need to retry those. Like that's because we we intentionally don't retry anything because we don't want it to be like, we try to share your post. We think that it failed because between, you know, Facebook got it and they posted it but they didn't tell us that they got it and posted it. And so then we post it again and we don't want duplicates and stuff. Yeah. So we try not to retry those. We just instead inform people this might have failed. Yeah. You should check. Yeah. You know? Then here's how you, like, redo it. And but this is actually one scenario where it's a 100% like Facebook for some reason cannot get the image for the post and Yeah. It's failing. I don't know why. Yeah. Like, when every time I go and check the image, it's there. It's there's nothing there doesn't seem to be something on our end. Mhmm. And so I was like, okay. That's actually like for these, you know, subcodes of digit a, you know, subcode this, subcode that, they're dumb. Error reporting, like, this is actually a thing that I should retry. Yeah. So, like, maybe retry it in, two minutes.

[00:24:01] Kris: Right.

[00:24:02] John: And because I've noticed a lot of them coupling to like the beginning of a minute and the beginning of an hour, which is probably when lots of people schedule things. Mhmm. Mhmm. And so it's almost like Facebook is a little not overwhelmed Right. But like maybe there's more, you know, noise that happens in that moment. So so my hot take is Codex is not all bad and evil and But actually, for like some things, I'm also finding it's not perfect for like like my daily driver, like to just use it for everything. I I was like, you know what? Yeah. I started using it for reviews and found it was really good Cause like, Claude makes the plan, writes the code Mhmm. Codex reviews it, you get the best of both worlds. Then I started using it for like, actually like writing all the code and I was like, it's got some problems. Like Mhmm. Things that it does that I'm just like, no. That's not that's not right. Like, you don't why do I literally, I was yelling at it, like, right before I came down here. I'm like, I'm typing in all caps, like, what are you doing? You know? Like Well, and what's what that makes me think of

[00:24:59] Kris: is another is, like, when all these people are not who are nontechnical and say, like, oh, I'm vibe coding this weekend. Yeah. And they have no idea, like, if it's even, like, doing it correctly. You know? There's like, whatever it does, if it does the thing, regardless of if it Yep. If it's doing it right, they're like, cool. Like, it works.

[00:25:22] John: Yeah.

[00:25:23] Kris: And that But I also feel like if it works,

[00:25:26] John: it works.

[00:25:26] Kris: Fair.

[00:25:28] John: Yep. So it's it's like, it just but it might just work for you.

[00:25:32] Kris: Right.

[00:25:32] John: Or it not work for a thousand people.

[00:25:34] Kris: Right. Right.

[00:25:35] John: Right. And things like that. And I I I honestly, I think that there's something kind of great about that, like, bespoke thing that solved my personal problem and not feel empowered like Yep. But it is yeah. At the same token, I'm like, yeah. You made this thing, but, like, that's not, like, that's not gonna actually fly.

[00:25:49] Kris: I guess in in my mind and now to get, like, an MVP of something, I think it's that's a great Yeah. Way to do it, especially if you're not technical. But at at the same time, go back to our other hot take, it's like someone coming to a studio or something, hey, I'm an artist.

[00:26:03] John: Here's here's my design. Yeah. They have a coder. Here's my code. Yeah. Make it better. Yeah. Yeah. So yeah. Well, that's where, again, like, g stack kinda comes in, I I think I brought up last time. It can I think it can take someone who's just a viber and Yeah? Kind of take them to the next level because it Oh, yeah. It does it through, you know, Socratic method of just asking you a ton of questions. Right. Right. Because the only thing I think the thing that I have gained this this one might be a clip. Can feel it. I can just feel it kind of the thing that I have gained programming over the last twenty years is what are the right things to think about all the way around the problem, not just the one core thing. And I think a noob is gonna come in and they're gonna say, oh, I can upload a file. I've solved my problem. They're not gonna think, oh, okay. Well, is it a file that's safe? Does it need to be scanned? Should we allow people to download it? What happens if, like, thousands of people start uploading these files? And what if like, are we even verifying that they have access to upload the file? And like, there's all these other questions around it. I think that's what, you know, if you vibe right, what you're doing is you're saying, what am I not thinking about? That's Yeah. Because like, most people are just like golden path, here's the thing, sweet, I solved my problem, I'm a black coder now. Yep. But I think what they should be instead is, okay, now I've got it working. Now, break it. Yep.

[00:27:16] Kris: And I That's what people don't do. And I and I love trying

[00:27:19] John: to

[00:27:19] Kris: break

[00:27:19] John: I think

[00:27:20] Kris: that's over just even the last decade of when I've worked with different teams, it's like I wanna I I don't go people often go to the path that they know works.

[00:27:28] John: Yeah. And I always wanna go to the path that doesn't work and

[00:27:30] Kris: say like, well, these things don't work because Yeah. Knowing that users aren't gonna necessarily just do it the way that you expected them to do it. They're gonna do it every other way Oh, than what you expected.

[00:27:40] John: They're gonna put in the title that's 300 characters Yep. And you only have 250 So if you just truncate it, you're you're they're gonna be like, where did it go? Yeah. You know? And now you've got a vibe issue.

[00:27:49] Kris: Yeah. Or even like on your box out, like, if you're like, oh, someone's gonna upload a JPEG for for images of headshots. Yeah. Well, someone might upload a PNG, someone might upload a PDF because like that's what their marketing person sent them. It could be oh, like it could be so many other file types that you would you would your assumption is they're gonna do like, I took a picture, it's a JPEG, but they upload a raw file. And like because, again, they just don't know or they assume they can upload an image so they'll just That's

[00:28:21] John: literally why we support HEIC

[00:28:23] Kris: or

[00:28:24] John: whatever the app format is because people just do it straight from Yeah. Their phone and, like, it comes in as this massive file. And so, again, thankfully, like, we get error reporting. We say like because I knew at the beginning when I wrote uploads, like, only allow these content types. Yeah. And if there's another content type, it throws an error, puts it in a system, I get notified. I see it, I say, do we wanna support this or not? Like, you know what? I don't wanna do a BMP. Yeah. I'm not gonna support bitmaps. We don't need to do that. That's ridiculous. But, you know, HEIC or, like, new things that come out, I can say yes. Or WebP, I can be like, yeah. Totally. I wanna support that. That should be easy. They should just drag and drop and not have to think about it. Right. So then I can build that in. And I don't, like, notify the world, like, hey, we support HEIC. I just make it work and then the next person doesn't hit that issue and the software is just smoother for them. So but yeah, agree. Like, that's the kind of stuff that like, I think most people who who haven't been building software for a while or who are vibing things, they're not thinking about all the things that can go wrong. Mhmm. They're not thinking about how do I even know when things go wrong? Because, like, they don't even know that error tracking is a real thing. You know? Like, are people experiencing frustration or not? They don't know because they just vibed a thing and Right. And moved on with life. So Yeah. Yeah. Oh, I like this new segment. Tell you what, there are sparks flying everywhere. Hopefully, nothing caught on fire. Nice. Okay. Fan of flames. Recommendation time. Kinda gave my recommendation already, I guess. My recommendation is to try more than one, you know, AI. Don't just use one. Try and kick the tires of a few. That's my my biggest one.

[00:29:52] Kris: I don't know. I don't know if I've been testing out anything lately. I've I've been on the same book as last episode.

[00:30:00] John: So What about your your busy light? You know? I mean, is that a recommendation? Or Yeah.

[00:30:07] Kris: So John recommended a very cool light so that people know when you're busy, you don't wanna be bothered because it's it's people right now just think if you're there, you're available to talk to. So he did a bunch of research.

[00:30:24] John: Did way too much. Claude did, let's

[00:30:27] Kris: be honest. And then sent me a link. He's like, cool. And I clicked

[00:30:29] John: on it. It was like $150. Yeah. It was expensive enough Yeah. That it's like

[00:30:33] Kris: It's just that I looked on Amazon. I think I actually found that same one on Amazon. It might be a little cheaper.

[00:30:38] John: Oh, interesting.

[00:30:38] Kris: Or it might be a knock off. It looks the same. Okay. So it might be a knock off. And then I was like, oh, there's this other thing that looks like a cool, like, salt lamp ish. Yeah. And then I based on how big the photo is, I was like, oh, like, cool. You can sit on the desk, maybe be like Yeah. Eight inches tall, and that can be like a red, you know, light if don't bother me or green light, you know, it's safe to come in. Yeah. And it shows up, and it's like

[00:31:03] John: three inches. Yeah. It was like three inches. It was very small. It's like you could probably see it if the lights were off. Like, would be there'd be like a little glow in there. Yeah. But probably not the effect you were looking for.

[00:31:13] Kris: Well, in in our our our friend Bethany had bought she had she had a party recently, and she was I I I think she we make the joke because T Mobile made with Alibaba, you know, same thing. And she bought these, like, lemon strings to like, vines that had lemons on them. And the lemons were

[00:31:33] John: like Not lemon sized?

[00:31:35] Kris: Yeah. They were not,

[00:31:36] John: like, what they looks like on the photo.

[00:31:37] Kris: Yeah. Mini lemons. And there were also, like, maybe, like, 10 of them for, like, four vines.

[00:31:42] John: Would you say that product was a lemon? I mean, I yeah. Sorry. Dad jokes. Yeah. So, I would

[00:31:47] Kris: say buyer beware. So we might be returning that one getting a new getting a new Switch, but I we'll see next time if it's a good buy to see. Do people respect the red light? They might not right away. Yeah. But I think once they you know, most people will see a red light and maybe get a little bit of a vibe. Yeah. I think put on the desk is not enough. Like, for the people that

[00:32:09] John: on

[00:32:09] Kris: the window. Yes. I think it's gotta be on the window

[00:32:12] John: or the outside. Yeah. And so something with, like, a three m stud But the

[00:32:15] Kris: outside I saw some that you could put, like, busy, you know, but you had to flip them from the outside. And I like, well, I

[00:32:22] John: don't No. You don't want that yet. That's Yeah. I was taking control. That's the one that I sent. It's just got a little flipper on your desk. You just flip it over and then it changes. Or to

[00:32:29] Kris: But but it would also be like if you could do that, but also if there was what I was hoping, it could be like a backlit lettering that said like busy or or Yes. You know, do not disturb.

[00:32:39] John: So there are some Etsy ones, but then you get into this whole like, I three d printed it. Yeah. Yes. Slap the light behind it. I was again, in the in the research, I found a bunch of Etsy ones. And I was like, these are cool, but like Yeah. They don't then have the

[00:32:51] Kris: brains

[00:32:52] John: that like some of the more productionized ones have. I feel like there should be something in the $50 range. It's it's like it's basically just a light and, you know or, like, I have the tidbit, which is probably a little more expensive. T I d Oh, that has, like, your Yeah. Right now, has Gale contributions and, like Yeah. Crypto prices and stuff like that. And so, like, what that that could actually work really well because they look classy. They have water on the outside, but it would be more expensive. But then you could actually have an app that's, like, busy, you know, free, and it would change the digital lights because there's like a 100 something.

[00:33:26] Kris: Just even wonder if we could make have like at one of the local, like, high schools, they they made some stuff for us here, like custom, and because it's kids, like, doing Yeah. And they have all the machinery.

[00:33:37] John: Yeah.

[00:33:37] Kris: And I wonder if they could create exactly what we want. Like Yeah. Because they've been doing that, like, we have some light covers they're working on that Yeah. Will specifically cover the light fixtures so that we can block out certain lights when we're doing presentations that that are hitting the screen. Yep. They made our coffee pod, like, twirly things. But if they did, a clear plexiglass that was etched that had like an LED light on maybe the top, so they would fill it,

[00:34:05] John: like, fill in.

[00:34:05] Kris: Yeah.

[00:34:06] John: I don't know. I'm just thinking

[00:34:07] Kris: I wonder that might be get get exactly what we want, but it might be take way more time of like r and d on it.

[00:34:14] John: Yeah. It would probably be cheaper in in the long run as well.

[00:34:17] Kris: We can sell them to everybody.

[00:34:19] John: Yeah. Exactly. A profit center. There you go. Increase the EBITDA. Nice. I like it. Yeah. So oh, did you so then you're You you did not technically give a recommendation. That was kind of like an anti recommendation. But if you have one, I mean, it's it's fine. But that was also That was great.

[00:34:36] Kris: Yeah. I don't know if I have one this time.

[00:34:40] John: I I Yeah. I I

[00:34:43] Kris: I've also the non recommendation is after an old fashioned, don't switch to gin over oh, no. Don't switch to tequila. That was Oh, well Yeah.

[00:34:54] John: I

[00:34:55] Kris: mean,

[00:34:55] John: the recommendation could have ended with don't switch to tequila ever. That would have been safe. Yeah. Yeah. That's funny. Yeah. So that that's my recommendation.

[00:35:03] Kris: Okay.

[00:35:03] John: I mean

[00:35:03] Kris: from an old fashioned to a margarita.

[00:35:06] John: That's a solid recommendation. It's not an anti event. Yeah. It's it's like, you know, it doesn't always have to it's it's learning. Yeah. We've learned things and we're sharing with the world. You're welcome. Cool. Let's go to Amber's small little updates. So what are, like, you know, just stuff that's been going on? We talked about the studio and things like that. What are some things that you're working on at Momentum or Access or that what's taken up the last week since we

[00:35:28] Kris: Yeah. At Momentum, we're, you know, working on I I would say, like, it's we're trying to create, like, a startup accelerator. And I would say it's more of a cohort because the for the first iteration of it, because we're we want to get six to tens young founders Yeah. Stay here, work on projects, and have a cadence of workshops, mentorship, help them keep going along their journey.

[00:35:56] John: I feel like we're getting close to be able to do that.

[00:35:59] Kris: Like, hope is we can launch an application in June and start in August. So I'm trying to wrangle, like, people and funding and things like that. And I feel like we're close and people like the outlines. And so I do think it's gonna it could be impactful. And that's kind of what we originally kinda built this for. Yes. It's for everyone else too, but we we need some of the we need help for those younger founders who maybe aren't at the scale where they can automatically afford office or a membership. So that's been going well. And with with this with the t I see. There's a lot of things, the teeter shop, we're figuring out how to scale better with, you know, potentially looking at contracts work. We are we're we're get send some of that the work out for contract so that we can take on more jobs, but not have to do all them in house. So that's that's interesting just because I don't know that we wanna get bigger and bigger from a floor plan standpoint, and this would help us get more sales without having to increase staffing. And then, we we did try out and I I I forget what if we mentioned it last time. We did try out some podcast art on the episode art Yeah. On YouTube. I got a lot of feedback from the team.

[00:37:22] John: Was cool though.

[00:37:22] Kris: Like, I I brought up

[00:37:24] John: Overcast and, like, it had the custom album Yeah. Art instead of

[00:37:28] Kris: Nice. And I was like, this is cool. Yeah. And and and so we wanna keep I think doing those things, it just makes it feel I know, like, I know you started watching YouTube, you mentioned it last time, and it's like, it's easier when I'm scrolling what I'm gonna click on because of the thumbnail to know which things I wanna click on. And so I think and you start in that mindset of would I click on this? Not because it needs to be click baity, but more as is it giving me enough context to what the thing is so that do I wanna learn more and watch or listen to it?

[00:38:00] John: You're you're being a good salesperson. A good salesperson is not just tricking you into buying something you don't need. It's helping you realize you need the thing that they're selling. Yeah. Yeah. So you're like, look, we've got some value here. Like Yeah. I wanna make sure that everything represents that the whole way Right.

[00:38:13] Kris: So those are the kind of the three

[00:38:15] John: different

[00:38:15] Kris: areas that I keep trying to push along. Nice. You?

[00:38:20] John: Yeah. I had a lot of stuff because of, you know, going back deep into like doing more things at once and stuff like that. So for me, I like a lot of box out stuff. We had Railway. So like we're using Railway. It's a hosting provider, kinda like Heroku. We're using them for our rendering stuff. And like we've slowly been moving a lot more things over. Yeah. So relatively, I mean, rendering for box out is the most critical thing that makes the graphics like Yep. If it's down, that's a big deal. It's been rock solid, super cheap. Love it. Great flow for the developer experience, all that kind of stuff. They had a five and a half hour outage, like, two nights ago. And it it seems like what it came down to was, like, Google Cloud shut down their account, and they were using that for, some control plane stuff. Is this, like,

[00:39:05] Kris: a

[00:39:05] John: small company or No. They just got, like, a 100 plus million in funding, like, in the last two months.

[00:39:11] Kris: It just seems, like, strange that someone else would shut down the account and they wouldn't yeah. It would be, like,

[00:39:15] John: a Yeah. So, like I I mean, I don't know. There's conspiracy theories, but, like, Yeah. They had kind of complained publicly about some Google Cloud stuff. And so I I don't know. I'm not saying that's why it happened. Yeah. And that could be totally just put the tinfoil on it on, you know? But, you know, they reached out. They got it, you know, but they're from something like that. Like, all the services were just running, but, like, nothing could get to them. Mhmm. Like, you know, the routing was basically dead. Mhmm. So that was a pretty big deal. So we were down. So we're I was like, for box out, was like, okay. I'm like, well, is this gonna be an hour outage, a two hour outage? Like, I don't know how long. Mhmm. So we're when we were like an hour in, I was like, okay. I'm gonna see how fast I can spin us up on another host. So I just, you know, went to town and it took about an hour, hour and a half, and I got us running on Fly, another service, and routed everything over there and we were back up. So Wow. In all, we're down a couple hours, but, like, it saved us the extra three, three and a half.

[00:40:16] Kris: Did you use, like, Claude or AI to help A 100%.

[00:40:20] John: Yeah. Claude was just like a ninja. It was just, I mean, with nunchucks and everything, it was just going to town. I had like three terminals open and it was all of them where I was just granting permission to Yeah. Just do more things on its own and move faster and because like the the problem with moving it, like you think, that's easy. You have the code. And this this service, it has no state. So there's no database. There's no anything. So it's just like, oh, just spin it up over here. But you know what it does have? I guess, counts as state is configuration. Mhmm. So you're saying like, there's like, here's how you access s three or r two or here's how you access this or here's how many workers of this to have. Here's how often to like restart things. Like, all these I just didn't even think about how much custom configuration that varies from the code. Mhmm. The code says, hey. Read the option. The option comes from the deployment. Mhmm. And so if you have a different like, trying to recreate that was that was the hardest part. So, thankfully, like, a lot of it I could pull from, like, the Box Out Rails app, which is hosted on Heroku. So I pulled a couple of those things. I pulled some stuff from a few other places and from local files I had, and I got it together enough that, like, it ran and it rendered graphics. And so we also lost image proxy, which we everything runs through image proxy for all the resizing and reshaping and all that. So between the two of those, it was just like yeah. I was like, okay. So again, managed to get it going. Super pumped about that. So that was, like, only a couple hours, but that was, you know, core memory type of thing. Right. Again. But I don't know. There's always, like, trade offs. Right. There's a side of it's like, well, I could run both going forward and I can I mean, I have Yep? You know, CloudFront in front of it, I can route wherever I wanna route. Yep. So I could route half here, half here, 10% here. Like, the hard thing is if you just have it cold, then you don't really know if it

[00:42:09] Kris: works

[00:42:10] John: until you make it hot. Yep. So you almost want to kinda run both. Yeah. But if you run both, you gotta make sure that like, well, if one's down, now some percentage are failing and I don't know. There's all these Yeah. So we went through this at GitHub a lot. We generally ended up with like, involve a human when something is down to make the decision to swap and stuff. Yep. Again, last week I talked about Brain and how Brain for Box Out is like making those decisions. Well, Brain was at what actually alerted me because it was like, I can't talk to Railway right now. I posted it in Slack. I always think you're saying Brian

[00:42:43] Kris: incorrectly when you say Brian. Brian?

[00:42:45] John: You know what? My dad

[00:42:46] Kris: received

[00:42:47] John: so much mail as Brain Dunamaker. It's not I'm not even joking. Like, so many people Yeah. Swap the a in the room. Yeah. Yeah. And he always said it was because he was so smart. But yeah. It's really just a typo from probably somebody in Yeah. Philippines or something. But so that was a big one. A whole lot of, like, performance updates for Box Out too because I was just, like, in that mode lately. There's just we we have proliferated a lot of change, a lot of new things in the last five months. Mhmm. And when that happens, you know, things get used in ways you didn't plan on. Right. And then you end up with, like, you know, performance issues. So I spent a bunch of time on that. That I think also that comes up because I'm working on Flare a lot, is our performance monitoring solution. And so, like, I look at that because I'm working on it, and then I see a problem, and then I have to go and fix it and pull the thread. And so so that was a lot of the stuff lately was fixing a lot of performance things and stuff like that. And then I've actually been working on click tracking for Box Out as well, which is kind of a neat solution. Every you know, you can create short links and the goal is to have this all done before NACDA, is like a big conference for us in in two or three weeks, I And so we have pages, which are like a way to like create a single standalone like, you know, landing page type thing for people. And so the next thing is to make links. And so then they can have QR codes and we'll track scans and clicks and all that good So I've been working on that a lot. That's that's been taking up a lot of time. And then the last thing I already mentioned about Flare, I've been doing a lot of work on Flare. So basically, the we have, like, the here's what's happening really cheaply from a high level. And what we didn't have was, like, the other side of the barbell, which is, like, on this individual request or this individual job, like, here's what's happening. And and you want that so that you can take the aggregate and give the, you know, the individual to Claude. Right. And then Claude can, you know, say, oh, well, this is happening 30 times. It shouldn't be happening 30 times or this is taking way longer than it should, so let's make it faster and here's how. Yeah. Yeah. So I've been working on that and it's it's turning out pretty cool. Like, I deployed it today just for myself, but it's behind a flipper flag for me, but it's it's basically working kind of the whole way through. So now I just need to like improve it, make it better. It's a very rough version. I would never show it to anyone else right now. Like, it's embarrassingly bad. But the technology of like from, you know, kind of the web interface to say like, start tracing this. I want some individual requests to then it gets to your app, you know, like Speaker Deck. Speaker Deck says, oh, I see a new trace. Okay. So then all a sudden it like kinda turns on a brain and says like, hey, if you see a request that looks like this one, then some percentage of time, send it over here. And so it kinda does that. Yeah. And so then you get those individual traces and it just sends them right to our object storage and stuff like that. Kinda designing this whole like, hey, start watching this thing. Okay. Cool. I'll start watching it. Okay. I watched it. I sent some things. Cool. I got those things. Okay. Now give them to Claude. Do that. Yeah. It's it's really neat to see the whole thing all the way through. Like, no, I don't I don't think there's bits of companies in this area that do similar things but not the exact same. Right. And what's cool is like because we have these high levels, like, I don't have to ask you like like how many traces do you want or, you know, for how long you wanna trace or, like, what percentage do you wanna sample the traces at? Like, because if you have a really high throughput, like, ton of stuff's coming through it, like, you can have, like, a really low sample rate and get, like, 10 of them in, like, a millisecond, you know? And if it gets like 10 requests an hour, then, you know, it might take an hour to get it. Right. Right. So you actually want a 100% sampling. What's cool is when you have those big aggregates, I know exactly what your traffic is. So I don't have to ask you any questions. I can just say, literally, it's just a button. You just say start tracing and then you can say, give me five more, give me five Okay. I got enough. I'm done. Like and the app just kind of does handles that all with your app. So it's it's gonna be pretty neat when it's done. I think when it's done, then you won't even hit the button. You'll just say, Claude, there's a problem. Claude will say it'll look at the aggregates. It'll say, okay. Well, this has taken up a lot of time. It wasn't before. So let's get some traces on it. Boom. It'll kick the tracing on. It'll get some traces. It'll say, okay. I got some traces now. I see the issue. I've fixed the issue in your repo right here. Like, do you want me to make a PR? You say yes, and your problem goes away. So I think it'll be pretty magical when But it's so those are the the the things for me. Just a lot of, you know, same old stuff Mhmm. Splitting time between things Yeah. Yeah. Trying to push them all forward. So Awesome. Yeah. That's it. Well If you have done

[00:47:20] Kris: so, you can like. Yeah. Subscribe.

[00:47:24] John: Leave a comment. Do it. I mean, would it be our first comment? Have we had a I'm just I'm realizing, like, have we had a comment? Like, think about think about where you're at, listener, right now. You could be our first comment. Leave a comment. Yeah. That would be that would be great. Yeah.

[00:47:40] Kris: I I actually and I do have some things that I'm working to work on that as we're talking. I was like, oh, for for next time. So I'll try to get those ready for our next episode.

[00:47:49] John: Love it. Great. We'll see you next time.