TRANSKRIP 🔗
Colleen: Good morning, Joe.
Joe Masilotti: Good morning, Colleen. How you doing this morning?
Colleen: Okay, okay. So yesterday, jumping right in, I am listening to our podcast and I am cleaning my bathroom. And you know, I have those gloves on you wear, because I wear gloves when I clean the bathroom, because like, I don't know, my kids are gross. And I had this idea and I want to tell you about it for your new service, which is getting apps into the App Store, right? We can do a whole crazy viral marketing campaign.
Joe Masilotti: Yeah, let's do it. Hehehehe... Hehehe... Mm-hmm. Mm-hmm.
Colleen: where we do something that's totally crazy, outside of typical B2B SaaS, like some kind of B2C iOS app, we could build it in Rails, use your service to show how quickly it is to get on the app store, and then run all of my crazy GTM automations on it, like let OpenClaw run meta ads and see if we could make money.
Joe Masilotti: Mm-hmm. Yeah, that sounds awesome. It feels like we have all of the pieces of the puzzle to grow something and distribute something. We just need a product. ⁓
Colleen: Something, right, so we need a product job. ⁓ my gosh. And so what is so funny to me about this is I've been thinking this. So I am building go-to-market automations for my clients. And I do wanna talk about that, because this was our second week of SEO results. So people are getting excited, was fun. ⁓ But I wanna run meta ads and I have one person I'm working with and we're gonna start meta ads probably this summer, because we're not quite ready.
Joe Masilotti: Mm-hmm.
Colleen: I just wanna try this stuff, right? Like I just don't have a good test bed. Like simple file upload is not a good test bed for running meta ads. And so I don't have a good test bed. So I'm like, what if we just did something totally off the wall that like we have zero expectations. It would be fun to build. We could use your service to get it in the iOS app store. So we don't have to stress about all of that. And then I could just build like open-claw meta stuff and we can see what happens.
Joe Masilotti: Mm-hmm. Yeah, sounds really fun. as, I mean, worst case it's a story. On like, yeah, yeah. And ⁓ it also gives me some more feedback on Ruby Native before, you know, the official launch, which is exciting. Cause I'll be able to like be in there in the code and fix anything that happens. ⁓ We could spin something up product wise over the weekend, I'm sure. That was, that's good enough.
Colleen: I something. Yeah, right, you lose couple hundred bucks running ads and your time.
Joe Masilotti: it would really come down to what are we building and who are we building for and you have an idea. I saw a bulb just go off.
Colleen: Yeah. I will know I'd there's of course there's a lot of like things I'm interested in that are outside of like work. ⁓ So I think it'd be fun to build for like personal trainers. Everyone at my gym is trying to start a side hustle where they do online personal training.
Load full transcript
So I know a bunch of people trying to do that. ⁓ I don't know like things like that. There's this all like there's I have a couple like really weird ideas ⁓ of things we could do.
Joe Masilotti: Mm-hmm. Mm-hmm, mm-hmm. Hmm. Yeah, I like the personal trainer angle. I was on a little bit of a tangent, but I have, I moved to Portland a years ago, found a barber like right away and have been working with them ever since. I go there, my wife goes there, my kids go there, like the whole family goes there. He went on a six week vacation to France, which is why my hair looks like it does right now. And texted me yesterday that he's not coming back.
Colleen: ⁓ no!
Joe Masilotti: Yeah, so he's gone, he's gone. I gotta find a new barber. And I'm like, you know, two months deep into no haircuts. Long story is that I had to find a new barber and the website that I got sent to to book that appointment was one of the worst built things I have ever used in my entire life. I had to enter, I had to verify my email address, my phone number, put in a credit card before I could even see schedules. It was...
Colleen: He's gone. Okay. Yes. I believe that. What? That's crazy.
Joe Masilotti: so bad and it was so clearly this like, you know, third party solution. It wasn't even on their domain, like the barber's domain, right? It was like some other thing. And then you go to the homepage because you want to like change your appointment. And the landing page is literally like book clients. And then it fades into something else and it's like get results. And it fades to something else and it's like fill seats. And you're just like, ⁓ this does everything. No wonder why the part that I used was awful. So
Colleen: Yeah, they kick you to another domain. Yeah. Gosh. was terrible.
Joe Masilotti: I mean, I'm sure there are a million and one of these out there, but I had a terrible experience with just like this side of it, and I have thoughts on how to make it better. For sure. For sure. Yeah. Mm-hmm. Like if you're doing the personal training with booking appointments and stuff.
Colleen: Excellent, okay, I love that. Yeah, that could be another idea. And what would be fun about this, yeah, like booking appointments. And what would be fun about it is it's not, like we don't have to evaluate this as a good business idea, right? Like it doesn't matter, because this is just an experiment to test out your stuff and to test out my stuff. So it'd just like a fun, fun experiment, like a way to make this more tangible.
Joe Masilotti: Mm-hmm. Yeah. Mm-hmm. So. Yeah. So if it's an experiment, it has to have a pass or a fail at the end of it. For Ruby Native, it's like, does it get into the App Store and is it approved and does it not crash? Pretty objective. As long as we have one person who downloads it, we can consider that fine. For running ads, what is the metric you would measure to see if they're working?
Colleen: Ahem. Yes, sir. Right. Yes. I would want to see that people are buying
the app. I would learn more about, I don't know much about B2C consumer apps. But I think the metric for me would be if I put $500 of meta ads behind this, kind of, I think too with B2C, it's more like how many users do you get? So it's not how many paid users do you get, especially upfront, it's how many active users do you get? If I put $500, I don't know what the click through rate is for any of these, but if I put $500 in meta ads,
Joe Masilotti: Mm-hmm. Hmm.
Colleen: behind something, like not how many eyeballs, but how many signs up and get active engaged users. That would be my metric.
Joe Masilotti: Mm-hmm. Yeah. So they would have to see the ad, click the ad, sign up on the website. So the offering for the product itself has to be not crap. If we're actually expecting them to like go through the onboarding and, you know, create an account or whatever it is versus like if you were just measuring, yeah, if you were just measuring like clicks, it's like, okay, we know that this is just based on the ad, you know?
Colleen: Sign up. Yeah, no, that's a good point. You actually have to have a product. Maybe we should just measure clicks. Maybe we should think of something even simpler, right? Like, so we don't get caught up in the details, in the minutiae. Maybe you're right. Yeah, and building a product that we're never gonna turn into a business. Maybe we should think of something simpler. If we go after small businesses, so the barbershop example, ⁓ click-throughs to a wait list. Or it could just be...
Joe Masilotti: Mm-hmm, mm-hmm. And building a product. Mm-hmm. Mm-hmm. Mm-hmm. Mm-hmm.
Colleen: I don't know, something like that?
Joe Masilotti: Yeah, I do love the idea, not to get too deep in the product, but doing a barber, let's just point out like with the barbershop thing, you have to get the right person who's in charge of the barbershop, like has decision purchasing power to actually buy that and switch over their reservation system to it, right? Pretty big ask, especially for something that with no testimonials, no like support, no onboarding.
Colleen: The barber shop, yeah, yeah. Mm-hmm. It is a big ask. I'm fake out.
Joe Masilotti: if it was something that was way more like single player where, know, you could get one of your friends who, who you work out with at the gym, who's like, wants to start their own side business. And it's like, ⁓ we like spin you up a website and a reservation system. It's like, they get the folks that are going from like zero to one might be an easier demographic because like they can just like sign up and do it themselves. They're having to like talk to anyone or change processes or anything. It's like, it's just a little bit easier for it's a little easier of an ask.
Colleen: Yeah. Okay. ⁓ and. Yeah. Yeah, okay, we should think on this. We should talk more about this. But I think it'd be a fun way to like do all the things.
Joe Masilotti: No. No. I like it, I like it a lot. I think my only reservation is I want to make sure we have a halfway decent product, because I don't want that to be the reason that it all fails. And we're going to be talking about this a lot, so it's got to be at least halfway presentable.
Colleen: Yeah. Yeah. Yes, I agree, I agree.
Joe Masilotti: but it can't sidetrack us, which is, know, kind of go odds at each other.
Colleen: Okay, well, think about it. I agree with everything you just said and we don't want it to be a whole thing, right? It's just a test bed. ⁓ But ponder on
Joe Masilotti: Mm-hmm. Mm-hmm. Yeah. Yeah, I already am. I'm distracted already.
Colleen: Okay, ponder. Okay, you're distracted. Well, I knew if I told you, maybe you'd get into one of those like 48 hour sprints and just do the whole thing. You're welcome. ⁓ I'm here. That's what I'm here. I'm just for inspiration, Joe. Just here for inspiration. Okay, so in other news, so I had this idea yesterday and I was like, this would be so much fun. We should totally do that. But I also have a lot of other cool things on my plate that I am excited about. ⁓ So, you know, I've kind of...
Joe Masilotti: ⁓ thanks. Thanks. Yeah, you're just the inspiration for me to burn some tokens. Got it. Mm-hmm. Yeah, tell me more.
Colleen: I've been really interested in ads for a while. And the reason I have been interested in ads is because I think SEO is just dying. I think that, and I know there's a lot of opinions about this and everyone is gonna have a thought, but I am just not convinced that people read content very much on the internet anymore. And so SEO as a channel is going to become harder and harder and people are turning to paid ads because distribution as a,
Joe Masilotti: Mm-hmm.
Colleen: in general is just getting so much harder. So I think paid ads are really going to have an upswing. And so I've been super interested. So as part of that, I've gotten a few people to let me manage their Google Ads.
Joe Masilotti: Ooh, exciting. Okay.
Colleen: Yeah, especially since I don't know how to manage Google Ads. But you know me, I think I can figure anything out. And it mostly works out for me. yeah, so I have a couple people with real budgets, like real businesses, and I am managing their Google Ads. So I'm going to set that up today and see what I can do. And again, I don't want to be, I'm trying not to overthink any of this, right?
Joe Masilotti: We'll figure it out. Yeah. Yeah, okay. Nice. Mm-hmm.
Colleen: Monday, Tuesday are kind of like my fun days where I can explore and I don't have to worry about billing hours. And so I'm just gonna start doing it and see what I learn.
Joe Masilotti: Mm-hmm. Do know what your first change will be to all of the, like to their ads or like what you'll look for first to optimize? Like what is your, what are you offering them? Like I'll make it better. I'll increase click through rate. I'll lower your click cost.
Colleen: So here's the. Yeah, but the trickiest thing is like, at least at the two accounts I've looked at so far,
and it's only three accounts total, the conversions, tracking conversions in Google Ads is kind of a nightmare. And you don't have a good way to like even know if they worked. Because for example, I'll just use simple file upload, because I don't want to talk about people I'm working with on here. For example, you come in and you sign up to simple file upload.
Joe Masilotti: Mm-hmm. Mm-hmm.
Colleen: I have a Google Tag Manager that's supposed to be going through a custom event in Google Analytics that is tied into Ad Manager. But if you come in and you sign up, I have an unlimited period of time where you can just try the service. So if 25 days later you upgrade to paid, it's hard to capture that conversion event on that user because Google uses these cookies that expire in a couple days or whatever. ⁓ So you have to...
Joe Masilotti: Mm-hmm.
Colleen: get the user and send off a custom conversion event. And so like just non-technical people setting this up, like it doesn't, I think it's just not set up right. So that's like step one. It's like, and there's not a really a good way to test it. And, sorry, I'm like on my Google ad soapbox. If Google sees that you are not getting conversions, it will automatically stop running your ads. So it's like, you, I don't know, it's just kind of, it's very confusing.
Joe Masilotti: Mm-hmm. ⁓ yeah. nice.
Colleen: especially for people who are less technical. And ⁓ that pipeline is just not very clear because, for example, if you were to go to the internet and be like, how do I set up Google Ads? You're going to get some people telling you to ⁓ use Ad Manager directly, some people telling you to use Tag Manager with analytics through that. ⁓ And there just doesn't seem to be... And Google changes their UI like every day. So it's kind of a mess. So step one is just getting everyone set up and knowing we're set up correctly.
Joe Masilotti: And that would be like making sure that if someone comes in through the ad, but, or no, someone signs up and then 25 days later clicks an ad and converts, they get attributed correctly. Yeah.
Colleen: Yeah, they get credit for that. And that, again, like it doesn't sound that hard because by the time they convert, you have their email address and you should be able to fire their email address. You have a custom event in your app, but it seems like it's a little complicated. So I'm trying to unravel all those pieces right now.
Joe Masilotti: Mm-hmm. Is it? Is it written in JavaScript to all of this? Like where do you actually make,
Colleen: Yeah, yeah. You just fire off a conversion event in JavaScript. it shouldn't be hard, but it's like oddly hard. I mean, there's literally like I have a who literally runs a business and this is basically why he started the business. Because he was like, no one can figure out how to track the user journey. Google doesn't do a good job of tracking the user journey. Like from first touch. I know, right?
Joe Masilotti: Yeah, conversion's a pain.
Yeah. Yeah? Yeah. You'd think with like all of their tracking pixels and all the stuff that exists, I'm curious, not because I want to try it myself, but out of just pure curiosity, if meta ads track better, like if a different service has a better conversion and or like a conversion tracking, because I do feel like every website that has ads is sharing a pixel in some way. And it's just like sharing cookies across, you know, cross sites, cross site tracking and everything.
Colleen: Yeah, I don't know. I'll let you know. Mm-hmm.
Joe Masilotti: So I'd be curious to see how all of that actually plays out. You'd think for all the data you give away, it'd be able to track you.
Colleen: Yeah, I think it's a, it'd be easy. Yeah, you're not gonna be able to track you. So I am looking forward to this, because I have wanted an excuse and a reason to dive into ads. So that is also here on the podcast, and part of this is accountability. Like I want to come back next week and be like, I set these up, or I reviewed them, and step one, two, three of what I've learned is whatever. ⁓ So that's one thing. And the other thing, I am having a heck of a time.
Joe Masilotti: Mm-hmm. Yeah. Mm-hmm. Mm-hmm. Mm-hmm.
Colleen: getting my website live. I hate everything. I hate all the copy. I'm like, this is terrible. So I, and you know what I'm trying to do? It's because I'm trying to optimize for something I don't know yet. Like I'm trying to be like, what is my unique positioning? And I'm like, well, I've only worked with three clients, so I don't know yet. Right? Like it's two weeks, we're two weeks in. I don't have, like the way to really make it strong is to say,
Joe Masilotti: Yeah, it's hard. Mm-hmm.
Colleen: I do go to market automations for small businesses. I do, you know, documentation intake for attorneys, whatever it is, but I just don't know what it is yet. So that requires a little bit of patience on my
Joe Masilotti: Mm-hmm. Right. Patience, true. And also, just, I don't know if empathy is the right word, but you might not have one thing, you know? And what you're trying to get is not ever going to happen because it's actually, do AI at small businesses. And that means something totally different to the client you're working with. And you have, you know, seven case studies and two white papers. And it's like, oh, this is what she does. But...
Colleen: Right. Right. Right. Yeah.
Joe Masilotti: to try and put it in a box with one sentence, it's like impossible.
Colleen: Yeah, so I'm having a heck of a time getting the website live, but I'm doing a webinar tomorrow, so I need to have it live by tomorrow. So I gotta do that today. Yeah, it's still not live. Maybe it'll be live by the time you publish this tomorrow. Just to make me do it. I mean, I've been trying to do it, but everything I come up with, I'm like, this sucks. But I'm just gonna publish it, because I just need something right now.
Joe Masilotti: Okay, that's the deadline, yeah? That's a Colleen.es. You We'll see. Maybe I'll publish today though to force you. That's hard. That's hard. Mm-hmm. Mm-hmm.
Colleen: Like I just need something. So I'm going to do that. So that's my second thing that I want to make sure I have said here that I will do. So when we get on the mic next week. Oh, and one more thing. I just, just 30 minutes ago, I deployed my very first Laravel application using Laravel cloud.
Joe Masilotti: Nice. Right? Yeah. Mm-hmm. Was it everything that Arales developers ever dreamed of? Yeah, of course it was. Of course it was.
Colleen: Yes, Joe, yes, it was. I immediately was like, it's time for my quarterly reminder to Joe that I want to build Larambeau Cloud for Rails with him. Actually, I want him to build it. I'll do everything else. ⁓ my gosh. was so, it's amazing. So good, Joe. Like, why don't we have this? Like, I don't understand. I mean, I know I'm Joe.
Joe Masilotti: Alright, alright, alright, yeah, You can manage the servers. I'll build it. I know! That's ridiculous. Yeah. Yeah. ⁓ We have Kamau.
Colleen: Yeah, right, right, because that's the same. No, not the same, not even close. I just want to come in, I think I tell you this every quarter, and say, we should do this because it's so good. I was in there thinking though, so two things happened this morning. I woke up at 6.30 in the morning and I checked my email as one does, and there had been a cert that expired about midnight for simple file upload.
Joe Masilotti: We have Kamal, yeah. Yeah. Yep. I know. I know. I know. Mm-hmm.
Colleen: So I already had four emails from people being like, the whole service is down. ⁓ This is so.
Joe Masilotti: ⁓ my God. Because they're going, they're routing their images through you.
Colleen: Well, no, they can still serve files. They just couldn't upload new files.
Joe Masilotti: Okay. Okay. So, not terrible, but not good.
Colleen: Right, it wasn't like they couldn't get their files. It was not good. So it was funny because my morning was fixing the cert issue and then, unrelated, deploying something to Laravel Cloud. So it was funny because as I was deploying to Laravel Cloud, I was like, oh yeah, but to own a service that has to have 100 % uptime would be incredibly stressful. Right? Like if you had a deployment hosting service and you basically have to have 100 % uptime, that would be a rough business.
Joe Masilotti: Mm-hmm. Yeah. Yeah. Yeah. It's also very clearly not a profitable or sustainable business because there are how many platforms as a service and none of them say Rails first anymore. They're not even like Rails only, none of them are even the one that you choose for Rails. It's just like, we'll do AI first deployment. And it's like, that's all you get. ⁓ I have a couple...
Colleen: Yeah, I know. Used for rails. Yeah, you don't even have one anymore. ⁓ huh. ⁓ huh.
Joe Masilotti: I have two apps on Heroku right now, which I'm trying to migrate over. I have five apps on Render and I have some static, a static
site hosted on Cloudflare. That's my, that's my like split. And I was thinking about bringing everything over to Render and the cheapest box that I could run on Render for even like SQLite, ⁓ but running Solid Q.
Colleen: Mm-hmm Mm, okay. Ahem.
Joe Masilotti: is like a one gig box, which comes out to like, I think 25 bucks a month. And you know, that's fine for anything that's bringing in any sort of money, but for something that's like literally the demo app for a server that I just need to not have not go down. It's not really viable. So I went down this long exploration of what I should migrate to and I might migrate everything over to fly. Fly might be...
Colleen: month. Yeah. Mm-hmm. Okay.
Joe Masilotti: But I've also just heard bad things about Fly itself, so I like don't want to, so...
Colleen: I'm, yeah, so I haven't tried render. Or I tried it for like a second, but it wasn't appreciably cheaper, so I didn't care. I have some stuff on fly. It's okay. Like, I mean, I don't know. Like, I don't, I would not recommend it. And a lot of it is personal preference with how you like to work, right? Yeah, but I wouldn't move everything to fly.
Joe Masilotti: Mm-hmm. Yeah. Yeah. Mm-hmm. Mm-hmm. Yeah. What I've liked about render is that I have that like render YAML or TOML file. And then that sets up my server with everything it needs. So it's like, you know, infrastructure as code, but proprietary. But what's nice about that now is that I can have Claude code just write that file for me and then do a git push. Like I don't have to actually, I'm not in the render UI, like clicking around and stuff, which is nice, but.
Colleen: Yeah. Yeah. Yeah. Okay. Yeah.
Joe Masilotti: I don't know, it like saves me a couple minutes when I first deploy and it doesn't really have any benefits at the end of the day. I'm just paying, like look at my bill, I'm paying like almost $250 a month on services between the three. And I don't think any of those actually bring in money. They're just like, you know, like, they're lead gen or they are demo apps for products that do bring in money, but directly they're not bringing any money. And I'm like, I gotta cut this, that plus Cloud Code Max. It's like...
Colleen: Yeah, that sucks. Yeah. Yeah.
Joe Masilotti: Starting to add up.
Colleen: Yeah, so add it up. ⁓ anyway, Laravel Cloud is awesome. That's my summary. That's my takeaway. We need that. Someday we're going to raise money and we're going to build that, Jo. You heard it here first.
Joe Masilotti: Set them up. Yeah. And yeah, I know. I know, we're going to. It's just, will we build it before?
Colleen: It's only been like five years of me complaining about it. ⁓ man.
Joe Masilotti: I would love to be, that's like the product, you know like when you're really excited about building something, you're really excited to write like that first line of code or that first prompt and it's like, I can't wait to build this service, I want to be built. Like I want it to be done. I
don't want to be involved in it, I don't want to manage it, I just want it to be like to exist so I can use it. And if I have, know, end up being the CXO on it and can...
Colleen: Yeah. I know, right? I someone else to do it. ⁓ Same.
Joe Masilotti: help direct the product, sure, but I just want it built.
Colleen: ⁓ this is what we need to do. Someone else needs to build it and then we'll buy it.
Joe Masilotti: Hmm. And then we get like, hire them to run it for us. Dear listener, if you want to build an entire platform as a service.
Colleen: I don't know, guess I guess there's details we'd have to work out there. On a tire for some service. Okay, anyway, moving on. Let's see. So we talked about how I am getting my website live today with the generic positioning of something like, right now it's something like AI slow or I don't know, it's stupid. Anyway, whatever, it's gonna be whatever, because I just need somewhere to point people for my webinar tomorrow. And I'm gonna get the Google ad set up. And of course, like all of my regular client work on top of that. But those are kind of the big fun things.
Joe Masilotti: Mm-hmm. Mm-hmm.
Colleen: that I'm going to do this week.
Joe Masilotti: Yeah, cool. I'm excited to hear about the Google ad stuff. Next week. You said you had something that was like two weeks in from your clients. Yeah.
Colleen: Yeah. ⁓ right. I did want to say this. So this is week two of these Google Search Console SEO agents and people are happy. ⁓ So it's cool because this is the first week it starts because it saves the Google Search Console report every week. So now you can see week over week. So like this keyword, you increased on this keyword, you decreased on this keyword. Here are your key opportunities.
Joe Masilotti: exciting. Hmm.
Colleen: ⁓ So this is like the first, the second week. So that's been kind of cool. That was cool to wake up to, right? That's why it's so great. It's because you just wake up to it. It's right there. You're like, ⁓ this is delightful. So that's cool. So yeah, there's something there. Again, it's not a product. To me, it's not a product. Like I do not want to get in that as a product. We've talked about that. But as a consulting kind of like, it's like I bring it with me. Like if you hire me to be your coach or you hire me to...
Joe Masilotti: Yeah. Yeah, that's nice. Nice. Mm-hmm. Mm-hmm. Right. Right.
Colleen: do AI in your business, this is something I set up for people.
Joe Masilotti: Yeah, I like that. It also lines up with what we talked about last week. You don't want to sell a service or set up a box for someone and charge them a subscription. You want to give it to them and hand it off. And this aligns with that really well.
Colleen: So, yeah. Right. Yes. Yes, yes, yes. Yeah. So, yeah.
Joe Masilotti: So this week for me ⁓ is I think 100 % client work for the first time, yeah, for the first time definitely this year. I have five projects that are starting today. Yeah, not starting today, but three new engagements are starting today to make a total of five, which is ⁓ they're not like...
Colleen: Hello! Amazing! That's a lot. Okay.
Joe Masilotti: big engagements, each one on their own is relatively small, medium sized. But this is the most, ⁓ the most number of people I've ever worked with. So I'll be really curious to, five is a lot, I'll be really curious to see how my, my context switching works and staying on top of everything. And three of the clients have like weekly check-ins that I'm responsible for. So making sure I don't like fall behind on those. So lots of notes this week, lots of to-do lists.
Colleen: Yeah, five is a lot.
Joe Masilotti: and lots of meetings to kick everything off and follow up with everything. But it feels good to have work to do, really, for the first time this year that isn't my ongoing retainers. So that's awesome. It does mean that Ruby Native is put on hold for a little bit, which I'm upset about.
Colleen: Yeah. Yeah.
Joe Masilotti: but also is not a bad time for that to happen. Like as of this weekend, I think that the iOS release is feature complete for where we are right now. Like I can always add, you know, in-app purchases and more native integrations and stuff, but like for a launch, it is feature complete for iOS. Like you can get this thing to the app store and get approved. And I think that that's a really good time to...
Colleen: Okay.
Joe Masilotti: let the beta tester sit with it for a little while, gather some feedback, maybe we run our own app, you know, and like get it out there. And then not next week, maybe not the following week, but like soon launch this thing for real and go at it confidently because I've had people use it and I can like point to apps in the app store and say like, check this out, this was built Ruby native.
Colleen: ⁓ I love that. ⁓ 100%. Like I feel like go-to-market is always start with warm customers. That's such a funny side note. That's such a funny thing that we're like talking about people like those warm people, those cold people. Anyway, start with people. Right? Start with early adopters who are excited and willing to sit through the bugs because of their enthusiasm. Get 10 of them, run it through with 10 of them. And yeah, that'll take a couple of weeks.
Joe Masilotti: Mm-hmm warm Yeah. You want warm blooded humans. Yeah. Mm-hmm. Mm-hmm.
Colleen: work out the kings, and then you literally have 10 apps you can point to and 10 testimonials, and that's when you start really pushing on the launch. And also I would like to say, I don't think you need, I know you know the market better than I do, but do we really need, do you really need Android?
Joe Masilotti: It's interesting because I've had five people ask me if Android is planned and only one person was like, this sounds good, but like, I'm not going to use it until you have Android. Everyone else was like, okay, cool. Like I'll use it when it's ready. And that's how I'm, you know, I'm
looking at this from the, how are 80, 10 years of clients that I've worked with. Almost everyone goes iOS first.
Colleen: Okay. Yeah, whatever.
Joe Masilotti: and then Android is an afterthought. So I'm following that same playbook because that's what everyone wants. And if I lose a customer because they don't have Android, like, okay, come back when Android's ready. I'll add you to a list.
Colleen: Right. Right. Yeah. I f- Yeah, I think that go do your client work, work out the kinks with this initial group of founding customers. I always like to say founding customers because they sound special. And it is special. ⁓ You're founding customers. Get it in the app store. Get the testimonials. And then in a couple of weeks, you can do a big push. ⁓ And then you get some revenue coming in. And as someone who has revenue coming in from a product, it's freaking awesome.
Joe Masilotti: Yeah. Yeah, I like that. Yeah. Mm-hmm. Yeah. Yeah, right.
Colleen: It will really motivate you to figure out what the next thing is and to double down on it.
Joe Masilotti: Exactly. Yeah, and come April, ⁓ one of the clients that I have will already roll off. Like, I'll be down to four. By the end of April, another one will roll off. Like, they're pretty short engagements. So, come May, like, I'll have more time for this. And that will be a good time to like, okay, this is either post-launch, we're doing Android, or this is big launch. So, it's like, it's pretty flexible, but I think that that's okay.
Colleen: Okay. Okay. Yeah, that sounds great. Yeah, I think so. I think that's an excellent strategy. Like this is the hard thing with this is like being a little bit patient. It's like you want to be impatient when you're doing it, but also you need to be a little bit patient to see it work.
Joe Masilotti: ⁓ yeah. I'm so impatient. I was going through... So I released the earliest of access to my paid sub-staff subscribers. And anyone who had a paid account on a certain day, was like, here is a coupon code for three months free. Like, go have at it. And I think 20 people responded and were interested in it and then tried it out. And I've been following up with them and almost every one of them is doing it on nights and weekends.
Colleen: Mm-hmm. Yeah, yeah, side project.
Joe Masilotti: Right? Because it's like their side project. So I talked to someone on Tuesday and they're like, I'm like, hey, I fixed this bug for you. Like, ⁓ now works with Google or whatever. And they're like, cool. I'll check it out Saturday. I'm like, no, check it out now. Check it out now. So having patience with that has been super hard because I don't have that type of patience. It's like, I'm in the code. I want to know. I want that feedback loop. ⁓ But I have set up a, which has been pretty cool.
Colleen: Yeah. ⁓ Of course.
Joe Masilotti: It's the first project where I'm leaning heavily into AI to build processes around my workflows. So what I mean by that is I a ⁓ parent directory, which has like the project name, and then four directories under it. One for the Ruby gem, one for the iOS app, one for like the platform that runs on Rails, and then one for, ⁓ something else, can't remember, like resources, like a... images and stuff. But I always launch Claude in the parent directory because it has access then to all of the subdirectories and it's not a mono repo because I have to make the gem public and everything else is going to be private. So it's like this weird amalgamation where I wish it was a mono repo but it's not. But I open Claude in the parent directory and then I have a bin directory in there that has a bunch of scripts that I run. So I have one that's just like bin release and it takes a version number, goes through all of the
Colleen: to everything. Yeah. Okay. Mm-hmm.
Joe Masilotti: subdirectories, make sure that the git status is clean, bumps the version, commits it, git tags, releases it on RubyDrems, and then pushes it. And it's like, yeah, sure, I could do that on a GitHub action, but it's because it's across multiple repositories, I can actually say like, go dry run all of these. And if they all come back successful, do all of them. And like now I keep everything in sync.
Colleen: Yeah. Okay, I like that.
Joe Masilotti: And then I have one that's like called bin status, which looks at the git status of each directory, tells me what branch it's on, if it has any dirty changes, if it's ahead or behind of main. It's just really good for like setting context and removing the fat fingering that happens so often with releasing stuff. Like so many times before this, I have pushed code but forgot to release the gem or release the gem but forgot to push the changelog. And it's like...
Colleen: Yeah. Yes. Yes.
Joe Masilotti: So if I want someone to use this, it's gotta be like, you know, tight. And this single script, which of course I haven't actually looked at any of it, works and is really nice. And it's using all of the command line tools. Like it's using GH, it's using Ruby gem, the gem, or gem, I should say. So it's not actually doing any like API stuff. It's using all command line tools, which are already off to me.
Colleen: Yeah. Yeah. Yeah.
Joe Masilotti: So I don't have to worry about like giving it its own token or running on a server or any of that stuff, which has been a breath of fresh air. Yeah, I like those.
Colleen: That's awesome. Nice. Yeah, I have a bit of that problem with simplify upload because if I want to update it, I have to update the server. Then I have to update the web component, different repo. Then I have to update the react component, different repo. So I like how you've put all those together in a shared context. That's solid.
Joe Masilotti: Mm-hmm. Mm-hmm. Yeah. Yeah, yeah, and then I have, ⁓ like the plans and the docs and the, and the Claude files all live at the parent directory. So when I open up a new Claude session, it has context on like where everything is, even though they're
separate, ⁓ Git repos.
Colleen: Interesting. Yeah, I've been thinking just about this concept of how we are going to build pipeline and how we are going to centralize context a lot, both in my consulting and in my personal work and everything I do. ⁓ I don't have a solution, but it does seem like we should have some kind of structure where you have docs, like you said, you open it in a parent repo that has context on everything. Maybe the parent repo has its own readme.
Joe Masilotti: Mm-hmm. Mm-hmm. Mm-hmm.
Colleen: You know, like it's own, that's like here's, yeah. So yeah, I've been thinking about that a lot with context and stuff because it drives me nuts when I'm switching from Claude code sessions and I'm in one, I don't wanna close the window, because I don't wanna lose the code and then I start another one, but I want this one to have context of this one and all that. Yeah, it's interesting, interesting.
Joe Masilotti: Yeah, it does, ⁓ Mm-hmm. Mm-hmm. Right. Well, is it? Yeah, yeah, yeah. I haven't gone as deep as like git work trees because I kind of try to stay focused on one task at a time.
Colleen: Yeah, I don't like it work trees. I mean, I only tried them a little bit, but like I just I didn't like it. Like, I don't know. That's just I guess that there's a lot of like your personal style comes into this. But like you lose context and then you have to merge everything together. And it opens a work tree like one directory above, which was like weird. I didn't like that because I have like too much. I don't know. There is.
Joe Masilotti: Yeah, has to.
Colleen: It wasn't good for... I prefer just to do it one at a time. It's so fast. ⁓ But anyway, yeah, different.
Joe Masilotti: Mm-hmm. Yeah, I have a long-running inertia support branch, and I kept switching back to it to do stuff, and then I'd realize a bug and want to switch back to main and fix it there. And that was a perfect example for a perfect way to try work trees, have inertia be its own work tree in this. And even then it just felt like I was losing context. My context window had filled up ⁓ every time because I'm keeping track of everything. ⁓
Colleen: Okay. right. Yes, your context. Yes, yes.
Joe Masilotti: Yeah, so I ended up just like, I'll just use a branch and if it's a little dirty to stash changes, it's fine. But one thing I have figured out is that I don't enjoy skills as much as I would prefer a script. I was thinking about this, like this commit process, I'm describing to it to Claude in text. Essentially, I gave it a markdown file and I said, create a script.
Colleen: Yeah, that's what I do. Really? Oh, interesting. Yeah. Yeah, a skill, yes.
Joe Masilotti: have been a skill on its own but I realized that once something like there are definitely things that are better for skills like what you know think this way do this thing versus that but when I want something that is deterministic I don't want it to be a skill so I've almost created a skill that generates a script you know and like that's the skill that's ⁓ which has been really helpful but the script
Colleen: Yes. Right. Yep.
Joe Masilotti: calls out to a skill to generate a changelog. So it's like this nested workflow where like the script is running and then because it's in a cloud session it will be like, ⁓ let me look at the dirty changes and update your changelog and then come back. So kind of like a mix of both. I don't know if I'm doing it extremely inefficiently or token inefficient but it works and it's been deterministic for me outside of the changelog entries which is what I want.
Colleen: Hmm. Interesting. Okay. Yeah, and then come back. Yeah, so my Google Search Console is a script. And people, there's tons of skills. There's tons of SEO Google Search Console skills out there. And for keywords everywhere and Google Search Console, I wrote a script. Because for that exact reason, I want the same thing. I want it to be deterministic every time. So I think that makes sense to me ⁓ in terms of the way you're doing it. I think most people are still just using skills for that kind of stuff. But I think scripts certainly have their place.
Joe Masilotti: Mm-hmm, mm-hmm. Yeah. Yeah. And I think that if you're on a Claude bot or OpenClaw and you now have to worry about token usage, going as much script as possible versus skills is going to save you tokens. Like if your API key is hooked in, you know, you don't want to just like waste it having Claude make API requests. You want a bot or a script to make API requests and parse it and then pass it to Claude. And yeah. All right, I have one more thing to talk about.
Colleen: Yeah. Yeah, for sure. 100%. 100 percent. Yes.
Joe Masilotti: RubyConf is in July. Austin, Texas? I think? No. Is that right?
Colleen: RubyConf. Okay, where is it? ⁓ nice. Excellent. I don't know. It doesn't... It's in Vegas! Okay, you weren't even close. Cool. Onward. Okay, RubyConf is coming up. I understand.
Joe Masilotti: No, it's in Vegas. It's in Vegas. Railsworld is in Austin this year. And I'm pretty sure the CFP is due Sunday.
Colleen: ⁓ dear.
Joe Masilotti: Yes, and I haven't started anything, which is pretty normal for CFPs. Everything happens in the last like two days. But I'm struggling to figure out what I should submit. I have three ideas and I don't know. Okay, so here are the ideas. Submit something very similar to what I've done at Rails World the past couple of years and Rails Conf, which is how to build mobile apps with Hotwire Native. And just like continue, it's like a talk I know.
Colleen: Mm-hmm. Okay. Okay. Mm-hmm.
Joe Masilotti: still defies me further as the expert. It wouldn't take as much time to prepare, but I think it's less relevant because it's usually pretty code focused. So it would be more about like maybe how to use AI to take advantage of hot work.
Colleen: Yeah.
Joe Masilotti: The second is more of what my new positioning is, which is like, what should a Rails developer think about when
they're building a mobile app? So that would be kind of like the mobile app framework for figuring out which framework to use for Rails developers, ⁓ or having it be like, I don't know if I'd get accepted, but I would try, like the internals of how Ruby Native works. And then it'd be more of a marketing pitch. Ruby Native, but I would be talking about how it actually works because the infrastructure is kind cool and it's less about the code and more about how it orchestrates all together.
Colleen: Okay, so let me recap. Number one is Hotwire is just kind of like keep going with Hotwire, which is what you have given two or three talks on already, right? ⁓ That's good, because it keeps you top of mind as the Hotwire native guy. That's bad, because it's traditionally very code heavy and no one cares about code anymore. ⁓
Joe Masilotti: Those are my three. Mm-hmm. Exactly.
Colleen: Two, what to think about when you're trying, when you are building an app. So this is good, because this feeds into the new consulting you're doing, which is the kind of like high level deep thoughts. ⁓ And then three is internals of Ruby Native, which sounds like, why would I be interested in that?
Joe Masilotti: Mm-hmm. Yeah, good question. It's more of the... It's less about like the code and more about the orchestration of like how a web view could work and how you could submit to the app store without opening Xcode. But it's all the stuff that the developer doesn't want to deal with. So I guess you're right. Like, why do you want to know that? I don't know.
Colleen: So our goal here with the conference talk is obviously to keep you top of mind both as a consultant and now as a product builder.
Joe Masilotti: Mm-hmm. but probably the former 90%.
Colleen: Okay.
Joe Masilotti: Yeah, so three doesn't make sense.
Colleen: Three doesn't make sense, I don't think. I think... So your Hotwire Native Talk, how could you tweak that to make it more relevant?
Joe Masilotti: It has to be how to leverage AI with it. It would have to. And that would be something like if I were to do a live demo, I would be doing a live demo with Claude. Like, I don't really know how that works. I don't know if it's me like showing off skills or it's more what to think about when building Hotwire native apps. So it's less code focused.
Colleen: Okay, so based on your current consulting, my vote is number two, which is like how to think about mobile apps in the time of AI and Rails, like kind of high level, how to think about it, how to pick a stack, what to commit to. ⁓ And I think that really tightly aligns with what you're doing, which is, does this seem really hard? Like I do consulting where I'm your thinking partner in making these decisions.
Joe Masilotti: Yeah. Mm-hmm. Yeah. Exactly. Pretty obvious when you say it like that, yeah. Thanks. All right. All right, for the RubyConf, you know, board listening, that's what I'll be submitting before the end of the week. You heard it here. Thank you. That was helpful.
Colleen: That's my vote. ⁓ I'm glad I could help. You're welcome. Good. I'm glad.
Joe Masilotti: Anything else for this week?
Colleen: I just want to give a little update on my open claw. I was so obsessed with it last week and it's just gone off the rails this week. It's so dumb all the sudden. Like a honeymoon phase is over open claw. It's been a week. ⁓ It just got really stupid and I don't know why. So there's a million articles about how to like ⁓ prune the memory and stuff. ⁓ But anyway, I was disappointed. Couple things, like couple like quick things is one,
Joe Masilotti: Yeah. ⁓ no! What happened? The honeymoon phase is over. Hmm, interesting. Mmm.
Colleen: I was like, this is great. I do not wanna give it access to my email. Like I'm too much of, I'm not a security person, but I'm also like security conscious. So it's not getting access to my email. I was like, I can forward all the kid emails. I gave it its own email using, think, Resend. I can forward all the kid emails to it and have it summarize the kid emails, cause I get so many kid emails. But, okay. The kid emails link to these external newsletters.
Joe Masilotti: Mm-hmm. PDFs.
Colleen: that are JavaScript based. So you can't just go curl and get the newsletter. And since I'm running on a Hertzner box, I don't have a browser. Like, so the whole thing is like, I'm like, can you, it's called S'more. I don't know if anyone else's school district uses. Basically they send us an email that says, go read the newsletter. So annoying. You have to click on the newsletter, but the newsletter is JavaScript rendered at runtime. So you can't just pull the HTML. So it can't just like pull the HTML. So it has to spin up a browser to read the freaking newsletter and it can't spin up a browser because I'm on a headster box. And I'm like, this is a total fail. Anyway, that's my quick update. I'll just give updates every week about how I feel about OpenGLOB. But that's where I'm at right now. Like this was worthless. Okay.
Joe Masilotti: That's really frustrating. ⁓ Browserless.com is a tiny API, very affordable, that will take screenshots of pages after they're rendered. So I think you could do that and have, like use the API to have Browserless take a screenshot and then you could ingest the image, which of course like wouldn't be, you'd have to OCR it then, which is of course very backwards. ⁓
Colleen: Okay, writing it down. That's what I need. Okay, so... Okay.
Joe Masilotti: But I think that BrowserList might be able to output HTML because it will render the JavaScript. I'd have to look at that. That was what I was, I used that extensively when I was building that OG image library tool. It was built on BrowserList to take screenshots. So.
Colleen: But that should work. Ooh, okay. ⁓ yeah, that's right. Yeah, yeah, yeah. Amazing. OK, I'll try that. I'll give it browserless, see if I can connect to that API so it can get the screenshots. That might work. anyway, it's just like gotten, like it couldn't, remember stuff from yesterday. I was like, what is happening? So I think I just don't know. I mean, I haven't spent that long tuning it, right? It's obviously just a play project. So I'll report back next week.
Joe Masilotti: Yeah, that might work. Mm-hmm. Yeah, please do. I'm interested to see where this goes off the rails and if it hasn't already.
Colleen: than I said already. was like, yesterday I was like, what do mean you don't remember my spring break trip? And it was like, I have no recollection of what you said yesterday. I like, I hate you. I know, right? Okay. All right. That's all I got for this week.
Joe Masilotti: ⁓ my god. Stop gaslighting me! Alrighty, if you want to hear more, check us out at nopermission.fm. Our email addresses are there if you want to leave some feedback or say hi, and we will catch you all next week. Bye, Colleen.
Colleen: Bye everyone. Bye.