Robin Heinze:

Hey listeners. Before we get started, I just wanted to make a quick announcement. In a few weeks, we're planning a very special ask us anything episode. You submit questions, Mazen and I answer them. Nothing, okay almost nothing, is off limits. Whether it's React Native, Expo, AI, or something completely different, we want to hear from you. To submit a question, head on over to our Twitter, Bluesky, or LinkedIn pages, which are of course linked in the show notes and look for the pinned posts. We're so excited to see what you send in. All right, onto the episode.

Jed Bartausky:

Welcome back to another episode of the React Native Radio Podcast. Episode 363, building React Native apps in an AI era.

Mazen Chami:

So Robin, do you have any vacations coming up?

Robin Heinze:

Actually, funny you ask. In a couple days, I'm leaving to go to Legoland with my family.

Mazen Chami:

Nice.

Robin Heinze:

Which I'm super excited about. I've had a lot of people ask me what's even in Legoland?

Mazen Chami:

Legos?

Robin Heinze:

Which is a good question. It's not super obvious, but it's basically, it's just an amusement park. They have roller coasters and other fun rides and then a bunch of Lego experiences like, "Hey, build with just giant bins of Lego and you just go and you build." And it's geared towards a lot younger kids than Disney. The park is way smaller. The rides are less intense. Some of the rides even have height maximums. The grownups can't go on them. So I'm really excited. I'm hoping it'll be a little more relaxing than some of our Disney trips.

Mazen Chami:

In December, well, kind of going back, not future, but in December we went to Universal with my wife's family and that family for my mother-in-law, she has eight grandkids, but they're anywhere from the oldest I believe is either 19 or 20 and the youngest is my daughter who's now nine months. So imagine going to Universal and trying to find a ride where everyone can ride. Well, minus my- It's

Robin Heinze:

Not going to happen. It's not going to happen.

Mazen Chami:

The carousel, right? The carousel. We all rode the carousel. So we did that. There you

Robin Heinze:

Go. Yeah, that's usually our go-to.

Mazen Chami:

We did that.

Robin Heinze:

And Disney, that's our first ride is always the carousel because everyone could go on it.

Mazen Chami:

And then my son is obsessed with roller coasters. I can't do it. My wife-

Robin Heinze:

Yeah. How do you feel about roller

Mazen Chami:

Coasters,

Robin Heinze:

Mazen?

Mazen Chami:

I hate them. There's not a lot of things I hate in life and I can't know. It's not that I hate them. I am always interested in them and always like, oh, I could do this. But as I get closer and I'm standing in line, I'm like, "Nope, this is not happening." Yeah. So I'm the type of person that you hear about coming off a roller coaster and barfing and stuff like that.

When I first met my wife, I was down in Florida coaching a soccer tournament at Disney. So ESPN Worldwide Center of Sports, which is part of Disney. They give all the participating teams and players and coaches and they give every player two extra passes to take their parents to a park for a day because you always have one day off in that tournament. So I took my then now wife, then girlfriend at the time to ... Where did we even go? I forget. I think we went to Epcot. We went to Epcot because I was like, those rides aren't too crazy from what I've heard. And this is what we did for me to survive it. I'd go buy a Coke, those massive Cokes and go and hide it at the exit and then go and take the ride. And then on my way out, just sprint to that and just chug it to get my blood sugar back up because I just couldn't not go.

Oh my God. And still dating at the time, so I still want to do all the fun stuff and all the rides.

Robin Heinze:

Were you trying to put on a good show for your girlfriend? Yeah.

Mazen Chami:

And the worst part was the-

Robin Heinze:

Pretend that you

Mazen Chami:

Were- The first ride she chose was horrible. And I just had to be like, "Listen, I think that's the last ride. You can go." But no, but yeah, literally my son, we went to King's Dominion in Florida, not Florida, sorry, in Virginia about a month ago for spring break. And it's not that wild. It's a Six Flags. If you've heard of Six Flags, it's a Six Flags park, but it's not Six Flags level. It's right below it. Now they do have some crazy rides. Sure, you can find them, but all the rides that my son could go on, my son is four, mind you and he's as tall. In his class, he has three, four, and five year olds. He's as tall as some of the fifth graders, if not taller than some of them. So he's able to go on a lot of rides and some of the rides he chose, well, he's also still young though, where a parent is required to sit next to them.

So for one ride- Which

Robin Heinze:

They do by age, not by height,

Mazen Chami:

Right? They do by age. So it's weird. They do it by height, but then they look at the kid like, "Okay, you're still a kid. You need a parent with you." Just to let you know, my wife took my son on this ride. As she was going, I could hear her screaming, nobody else on the ride. And I'm like, "This is it. We're done." So I literally come out, I meet them at the exit because I'm with my daughter. She's sleeping in the stroller and my son is screaming more, more and more. And my wife looks at me, she's like, "For him to go more, we have to switch." So they do a parent swap where if you have a younger kid, they let you hang out by the handicap exit entrance and they swap you out.

Robin Heinze:

Yep. And then you swap

Mazen Chami:

In. Yeah.

Robin Heinze:

We use that a lot at

Mazen Chami:

Disneyland. So I did that one ride. I had my Coke afterwards and I took my daughter and I just left my wife and son. I was like, "You guys fend for yourselves. I'm not going on any more rides."

Robin Heinze:

My favorite roller coaster story from you is when you took ... So I'm going to link this in the show notes. I tweeted about it a year ago, but there's this picture of Mazen with his son in a little teeny kiddie roller coaster. Mazen, for anyone who doesn't know, Mazen is really tall and it's just smushed into this teeny, tiny roller coaster.

Mazen Chami:

For anyone that's been to King's Dominion and can spot this ride, it's a pumpkin patch ride where it's like two hills and it's meant for kids, like young kids.

Robin Heinze:

But he didn't want to go alone, right?

Mazen Chami:

But he can't go alone. Oh, he can't go alone. He can't go alone because he's not yet at the specific threshold. They have a weird thing where it's like, if you're between this height, you need an adult with you. So it's weird. Some rides, they're like, "Nope, you definitely need an adult." Some rides they're like, "Okay, if you're here, you could use an adult or not." But he's like, "I want someone to come with me." So I had to do it and I couldn't even do that simple ride and he did it five times in a row. But yeah, it's a funny picture. I'm going to put it in there.

Robin Heinze:

But yeah, so that's our upcoming vacation stories. Anyone who's going to try anything and try to find me at Legoland, this episode will come out after I've been home for a while. So sorry about that.

Mazen Chami:

Yeah, exactly.

Robin Heinze:

But yeah, we should probably get to our topic. I'm guessing people are like, "Where's the Reas media?" Which is

Mazen Chami:

Like, I want to talk about cruises next, but we'll do that for another episode.

Robin Heinze:

Oh, I have so much to say about cruises.

Mazen Chami:

My TL;DR for cruises is I don't know if I will like them, but I also think I will like them. So take it for what it's worth.

Robin Heinze:

Right. Same.

Mazen Chami:

Well, yeah. I'm Mazen. I'm your host, staff software engineer here at Infinite Red. And with me is Robin, my co-host, director of engineering here at Infinite Red.

Robin Heinze:

Hello.

Mazen Chami:

Before we get into our topic, let's hear from our new sponsor.

Jed Bartausky:

Infinite Red. Infinite Red is a premier mobile app consultancy focused on Expo and React Native. We're a team of 30 that's located fully remote in the US with highly experienced mobile app developers. And we've been doing this for over a decade. We're also one of the first development teams to adopt agentic coding in a way that keeps high quality standards while also being able to dive in and do things the old school way if we need to. If you are looking for mobile app, Expo or React Native expertise for your next project, hit us up at infinite.red/contact. Thanks. And now back to the episode.

Robin Heinze:

Also, if you haven't already, go check out chainreactconf.com. We're hosting Chain React, the React Native conference on July 30th and 31st in Portland. We would love to see you there. Go to chainreactconf.com for tickets. We will be there. A bunch of people from Infinite Red will be there. Jamon will be there. Gant will be there. Todd will be there. So come and chat with us and see some awesome talks. We'd love to have you.

Mazen Chami:

Yeah, I'll say it now because I feel like by the time this comes out, I would've probably announced it on Twitter, but I will also be giving a talk at Chain React.

Robin Heinze:

Oh, snap. Yep.

Mazen Chami:

We'll see.

Robin Heinze:

Come see Mazen on stage. Yeah.

Mazen Chami:

Come see us. Come say hi. We're partnering with Expo for this one, so there's going to be a lot of people from Expo. I'm sure Amazon might be there too. So definitely a great event in a great city. Be there. Pack your bags.

Robin Heinze:

All right. What are we talking about today?

Mazen Chami:

Building React Native apps in the AI era.

Robin Heinze:

So we've had a couple of episodes now about AI. We had one with Todd talking about just the evolution of software since the '80s and '90s to today and how AI is kind of fitting into the evolution of things. And then we have one with Jamon where he shared a lot about how he's built his new game, Gunship Origins, go wishlist it on Steam, and how he used AI to help him build that. But we haven't really talked yet about how AI and React Native are fitting together. And we are a React Native podcast. I'm pretty sure it's in the name. So we figured we would just chat about how we're using AI as a React Native consultancy and how we're seeing people in the community use it and kind of what it's been like building React Native apps in this new world.

Mazen Chami:

Yeah. And I think just to give a high level, if you're out there and you are 100% leveraging AI and Claude is doing everything from writing your tickets to creating pull requests for you for a feature or bug fix, we have clients that are doing that. We also have clients on the other spectrum of it, which is actually my client, where if there's any code that is user facing, we're not allowed to use AI on it. And if you think about front end mobile app, that's like 90% of the code. Take tests out of the picture. That's like 99% out of that code there. So we have both spectrums. So don't feel like whether you feel like you're behind or you're missing on something, there's just all over the map. Everyone's trying to figure out their footing in the industry, right?

Robin Heinze:

Yeah, that's the big thing. It's a huge spectrum right now of how people are using it and to what degree. I don't think there's a right answer yet. I think every sort of risk profile is valid in one way or another. And so yeah, don't feel like you're doing anything wrong if it doesn't match up with exactly what we're doing.

Mazen Chami:

So I think when it comes to tooling from our clients specifically, I think what I'm hearing about is mainly and almost not exclusively, but mainly Claude. Using Claude- A

Robin Heinze:

Lot of Claude. I saw a lot of Cursor too.

Mazen Chami:

A lot of Cursor. Yes.

Robin Heinze:

Cursor was bigger at first and then Claude Code kind of caught up with it and overtook maybe a little bit. But honestly, I see the Cursor Claude divide I see is more like how people like to use it. People have said people who really want to keep their handle on the file structure and how things are changing seem to prefer Cursor because it feels like VSCode and they have their file tree and they can keep their bearings on their code base a little bit more and feel like they're not losing control.

Mazen Chami:

Because you're almost in the IDE. It feels more like, even though you have your terminal for Claude and it's like the terminal is in Cursor and in your IDE, it feels less in the IDE, if that makes sense, while Cursor is almost like part of the IDE in a sense.

Robin Heinze:

Right. Well, and there's so many different ways to use these tools now, but there's Claude Code by itself. You just pop open terminal and you're running it and you're giving it commands and it's editing your code, but you're pretty blind to it. Vibe

Mazen Chami:

Coding in a sense.

Robin Heinze:

You're letting it do stuff and you're kind of trusting or you're reviewing the code, but you're doing it separately in another IDE, or it's not part of Claude Code. Then there's Cursor where it's a full application and you've got your agents, but you're also seeing all your files. But then Cursor has a terminal. So you can also run Claude inside Cursor, which is trippy. But I actually really like doing that because you can have Claude working on stuff, you can see how it's changing. And then you can have Cursor agents code reviewing the stuff that Claude has made. There's so many different ways to set this up. I swear every day I'm seeing something, some new innovative thing coming out on Twitter or wherever.

Mazen Chami:

I even saw one, I forget its name. I know we're kind of deviating away from React Native specifically, but I saw some IDE or some trying to be an IDE, but it's like agent IDE where it's like the file tree, but you can't really open your files. You can preview them, but you can't open them and edit them. And then it's like on the left-hand side, it's like all the different agents. You can pop in Codex, Claude, Copilot, whatever you want, and then opens up a chat box and you can spin up your agent. So it's like agent first IDE type thing. So yeah, let's kind of bring it back to how we as Infinite Red and even just the React Native industry is using AI. I think like I mentioned, our clients are pretty much trying to figure it out for themselves. Some of the larger clients are waiting to hear back from lawyer people to tell them, "Sure, this AI can do the work for you."

Robin Heinze:

Right. There's definitely the bigger the company, the more risk conscious they tend to be and have very valid concerns about leaks and data security and stuff when using agents. And I think it's understandable.

Mazen Chami:

Makes sense. Yeah. And for us, we respect that. And when it comes to using the AI on their project, we're obviously focused on quality first. So whether it's letting the agent do the work and then reviewing the PR and making sure that it's quality before merging it or even making fixes from that. But we're also doing our own experiments. I'll talk about my experiment later on. I know within here people are using AI and leaning into AI fully to figure it out, figure out where is it good, what's its limitations? Where are some issues there? I feel like almost every day there's a new skill that's supposed to fix something. I know Expo has their own skills, which are pretty good at keeping this happening.

Robin Heinze:

Well, and Expo has gone beyond skills too, and they released Expo Agent.

Mazen Chami:

I've used Expo Agent or I am using it.

Robin Heinze:

Yeah. What was your experience?

Mazen Chami:

It was great. It was great. I mean, to me what it felt like was basically, think of it this way and I'll put it in terms of our podcast episodes. It's basically like Claude and Radon in one interface where you're talking with it and the app just keeps refreshing and you can do deep links and stuff like that to it. Oh,

Robin Heinze:

That's

Mazen Chami:

Really cool. Yeah. So I think it kind of goes back to what I said before, the AI chatbot IDE type tool where this is essentially the same thing, like Radon and Claude in one interface. And I think that's potentially the future that we're going towards. And side note, it built an app that worked just fine. It accomplished what I needed. I added features, I adjusted features along the way and it did just great, just like any other Claude agent would.

Robin Heinze:

Well, I should go bug them to let me off the wait list then.

Mazen Chami:

You should. You should. Yeah.

Robin Heinze:

So one of the biggest questions that I've seen posed in relation to React Native, I saw this, someone tweeted this and then I kind of ran with it, was the suggestion that now that we have agentic coding, is React Native fundamentally irrelevant, which is like a big assertion. The idea being that if we had agentic coding back in, what was it, 2015 when React Native was invented, the problem that React Native was trying to solve would've just been solved with AI, namely how much work it is to build two concurrent native apps, Android and iOS, and keep them in sync with the same features and how much work and human power that took. And so React Native was created to be able to write two apps with the same code base essentially and could we solve that problem with agentic coding and sort of take React Native out of the question?

So of course that's an intriguing question and one that we are curious to have the answer to as a React Native consultancy. So we did some experiments. We put a team on it for a few weeks just to see how agents do. If you ask them, if you give them a single spec, like, here's my app, here's what I want it to do, here's what I want it to look like, build it for me in Android and iOS, build me two apps with this spec. And the results were mixed interesting. It did really well for the first 80% and then really struggled with the next 20%. It had a lot of trouble with visual feedback loops like, "Hey, this card looks wrong or doesn't have enough padding or whatever." Giving it visual feedback and having it correct the problem was difficult. It's really good at doing the boilerplate, the scaffolding really, really fast, but then you're still spending a ton of time doing the hard technical problems like timezone bugs and deduplication and getting the stupid tab icon to center correctly and that ends up still being human craft.

In terms of whether it's a React Native killer, they said not really. They did the same experiment side by side with asking the agent to build the same app in React Native and doing it in React Native was still faster and more consistent. Whether a human is doing it or AI is doing it, you still end up with weird platform divergences that you have to solve for and things like that, you end up having to have the AI fix the Android app and then fix the iOS app. You still have more effort. It's just overall everything is faster and the same with React Native. Building the React Native app is faster and then building the two native apps is faster, but RN still beats double native, if

Mazen Chami:

That makes sense. Coming back to the root, what is it, the root tagline of React Native, learn once, write anywhere kind of thing. It's the same with AI, right?

Robin Heinze:

Right. The AI still only has to learn once.

Mazen Chami:

And write everywhere, anywhere.

Robin Heinze:

And write anywhere. There we go. Yeah. So I think React Native will still be relevant, plus it has other benefits unrelated to how fast you can produce two apps. Being able to do over the air updates is uniquely React Native. Just like the concept of having the JavaScript bundle and being able to do updates that quickly is still a huge pro of React Native. Being able to use JavaScript developers is still a huge pro. Yeah, there's still going to be plenty of reasons to use React Native. I don't think it's going anywhere.

Mazen Chami:

And I think this, I forget if I've mentioned it on this podcast or on others that I've been on. The big phrase is this is evolutionary rather than revolutionary when it comes to AI. We need to evolve with it. We need to grow with it. We need to figure out these edge cases that we did in this project and figure out what works best and whatnot. And I feel like I've been saying it from the beginning, I also don't think I've said this out loud, but I'm still potentially a huge skeptic on full lean into AI. I'm still finding some stuff that's hard to get over and I'm just like, just you can do that. I'm hearing a lot of people being like, "Just push through the AI and it'll eventually get it." And then you have the code. I'm always like, "Just eject from the AI, you know how to do it, do it, and then come back into it and then ask it, review the code."

Robin Heinze:

What kinds of things do you find it's bad at? I

Mazen Chami:

Think it comes down to visual with it, but then also when it comes to these native packages, it might get its hairs crossed when it comes to I'm in React Native, but I'm in native, I'm in React Native, I'm in native. The bridge, it probably just gets stuck on the bridge kind of thing. And that's where I say it's evolutionary rather than revolutionary. Let's just take our time to get to the end. And I know when we get there, we'll be better off for it because we took our time. I'm jumping in the pool one toe at a time. Let me just put it that way.

Robin Heinze:

Now I'm trying to imagine what that looks like.

Mazen Chami:

My first foot's not all the way in. But no, to me, I think there's a concept where if I'm going to lean in, whether it's ... Jamon was on here and mentioned his night shift idea. To me, where I think night shift would be like 100%, sign me up. I'm going to do this today. I want, let's just say Claude, for example. I need Claude to basically give it a document and go to sleep. When I wake up, what it did was it wrote either a Maestro test or some sort of end-to-end test that failed. It recorded all that. It spun up simulators, it did all that. I'm going to tag a ... There's a repo I found, honestly, right before we came to record this episode and I think this is kind of leaning towards what I'm going at.

It's called Mobile PR Reviewer and you'll see where I'm going with this. It'll spin up these devices, it'll click through it, it'll show the fail happen. It's all recorded. It's all uploaded to the pull request eventually. It has the failing Maestro test. It'll then go to fix it. And I'm probably talking about a bug here specifically, or even just a feature. You write a Maestro test for a feature that doesn't exist, it's going to fail immediately sort of thing. You start fixing it and as you're fixing it, you're always running the Maestro test and having it fail, fail, fail till eventually it passes. And then what you're doing is you're then going through and doing this spinning up of the platform of the devices up and going through the Maestro end-to-end test and making sure it passes and then buttoning all up. So then revisiting, clearing our context and starting all over again, you confirm that the bug's been fixed, but then also everything adjacent to it has been handled within there.

Does that make sense?

Robin Heinze:

Yeah, it does.

Mazen Chami:

But that still doesn't get rid of the visual part that we're talking about, more padding, less padding. I don't know if it's going to be able to do that. That's probably a combination of snapshot testing and I don't know, giving Claude access to your Figma file to be able to get it in dev mode and click through it and see the padding and confirm the paddings, right?

Robin Heinze:

Yeah. I think we're going to see that it's going to be really important to have tight integration between your designs and your agents, which will probably look like Figma plus Claude or whatever.

Mazen Chami:

Well, it's funny we're talking about Figma. I don't know if you saw this, but did you see Claude Code design or is it Claude Design? It's basically Figma. And it's funny, I saw a tweet and I was like, "I need to confirm this." I went to the Stocks app on my phone and checked Figma stocks. Literally one minute after Claude Design was announced, their stocks plummeted like 20 or so. I feel

Robin Heinze:

Like I said that on, I saw that on Reddit. Yeah. Claude Design just launched and Figma dropped 4.2% in a

Mazen Chami:

Single day. Yeah. And that goes to show that yes, all of these tools that we use as developers from Figma to the code to the devices, end-to-end tests and spinning them up and clicking through them, they all need to kind of almost come together into one for us to get that evolutionary platform. And I do think, Robin, when it comes down to it, for React Native and even mobile development potentially to be fully successful and fully AI driven, we need an IDE that is like Expo Agent, just the agent and the devices. We always want to test in both devices and the devices all in one contained where we're working in that one and in the future, this is also probably done on your mobile device built in React Native, wouldn't that be cool, and it spins up the app on your phone and tests it and does all that or even get to test it.

It's that IDE, but also the ability for us to inject where I can drop my agent and pull up my text editor and start changing a specific file as I want to. So that's where I'm kind of at. And I think going into the last segment of this, I wanted to kind of talk about I'm building an app and I'm building this app 100% AI. That's my goal at least.

Robin Heinze:

And he's not going to tell us what the app is because it's a secret.

Mazen Chami:

It's going to be in the store.

Robin Heinze:

And he doesn't want all y'all copying his idea because this is real.

Mazen Chami:

Yeah. And my goal is to actually have this in the store. I want RevenueCat to be involved. I want it to be paid and stuff like that. This app was an idea my wife came up with during the pandemic where one night pandemic, no kids, just stuck at home. We have a whiteboard in our office and she just started writing stuff on the board and she comes up and she's like, "Okay, I know I sound like those people who you just told you're a mobile app developer and they have a crazy app idea." She's like, "But just hear me out." So I did hear her out and it was actually a cool idea then and I started building it, but then life happened. Now with AI, my goal was to revisit it. I put together my initial spec and I said, "Hey, installing Expo skills." And I gave it, I actually think installed all the skills.

I was like, start building this app, but how we're going to do this is I'm going to hold your hand through it. I did not give it ... I feel like most people that I know that are using Claude give Claude full access to all the ... When it wants to do pnpm and grep and all this I'm actually not allowing all, it'll ask me, "Hey, can I run this command?" I'm like, "Yes." Which

Robin Heinze:

Is the opposite of what seems like a lot of developers just are so annoyed that Claude keeps asking them for permission. They're like, "How do I just let it do whatever it wants?" And Mazen's like, "Nope, you will ask and I will approve every single

Mazen Chami:

Thing." But let me also step back. When I first started this, I was like, "You know what? I'm also going to do this while watching videos that people have been posting on how to build an app with no technical experience using AI." Beto did one of those videos, I forget, Frank sent me another video that I watched and I want to see how others are using it while also just see how Claude is doing it kind of thing. And let me just say this, the app is up until the paid feature stuff, the app is halfway there and I think I've probably sunk I don't even think I could say a week's worth of actual work into it. This is just something that at nighttime after the family goes down to bed, I'll have it build a new feature while I'm watching TV sort of thing.

And it's almost there. Now, here's the part that's crazy too. I'm still having it write tests. I'm still having it write end-to-end tests, Maestro tests. I'm still having it keep the docs updated and stuff. And as I'm manually testing it, it's working. But the part that's getting me, and like I said, I'm not doing anything to it yet. There's going to come a point where I'm going to get in the code and just button it up before I release it. But what's bugging me is after it does a new feature, it'll always give me this output. And the outputs, I don't think I have it here in front of me, so I can't really read it, but it's something like there's actually a TypeScript error, but it's like, this TypeScript error was there before so this passed and it'll move on sort of thing.

Robin Heinze:

Which is hilariously similar to what a human would sometimes do.

Mazen Chami:

Correct. Yeah.

Robin Heinze:

I know lots of developers that would be like, "Well, this error is happening on main and on my branch. So it wasn't something I introduced. I'm not going to fix

Mazen Chami:

It." It's the whole thing of it works on my machine type thing, right? Not my problem. And basically, here's actually, I'm going to read it. It's like, npx tsc --noEmit. My four touched/new files are clean. 16 preexisting errors remain in unrelated files. Some-

Robin Heinze:

Not my

Mazen Chami:

Job. Same ones as before, 27 to 16 because filename.tsx was deleted.

Robin Heinze:

You need to teach it to have a continuous improvement mindset.

Mazen Chami:

So yes.

Robin Heinze:

Leave the code base better than you found it.

Mazen Chami:

Oh, I like that. I like that. There's also this other skill which Frank also led me onto that I installed, but I haven't built a feature yet with it. It's called the GrillMe. Have you heard of that skill, the GrillMe one?

Robin Heinze:

It's Matt Pocock's.

Mazen Chami:

Skill, right? Matt Pocock's skill.

Yeah. And I think for me personally, one thing I'm learning as I'm going through all this journey on this, building this app is I think AI needs to always grill me. Sometimes when it just asks me the two or three questions and then it's done, I feel like it did a good job, but I also wonder what would've happened if it asked me more questions. Would it have gotten a better ... I'm also thorough with it. You might have some people as I read, you know how ask Claude to build you something and asks you questions, it'll give you option A and option B. Sometimes I'll be like, "Yeah, option A, even though you recommend it and I agree with it, I'm actually going to chat with you and I'm going to give you option A plus more sort of thing." You're

Robin Heinze:

Basically saying, "You've given me two choices, but it's more complicated than that and we need to answer more questions and get more specific."

Mazen Chami:

Yeah. And long story short, it's doing a good job. I basically gave it a screenshot of an app that doesn't actually relate to the app that I saw on dribbble.com that I liked its aesthetics and I was like, "Take this for inspiration and go." And it put together a beautiful app and it's doing such a good job and it's actually giving me Supabase SQL queries to run whenever it needs to change stuff and add stuff. And it warned me, it was like, "Oh, Supabase is probably going to say this is destructive because we're actually deleting a table because the way you're now describing it should no longer be members. It's now attendees." Little hint. She's like, "Oh, so we need to delete the table. So it's going to tell you it's destructive. Just accept it and let it move forward with it." Yeah, I'm excited, but it's evolutionary not revolutionary to me.

Robin Heinze:

Well, what you're describing exactly is what we're going to see so much more of is people like your wife who just has an idea and never had a way to get it done in an economical realistic way. Now they can. And we're going to see so much innovation that we never saw before.

Mazen Chami:

Totally. Well, cool. If anyone has any questions or feels like there's something that you want to talk about, please feel free to reach out to us, let us know. We'll be happy to chat with you all on social media or even I believe we have an email somewhere. If you can find it, email us. I don't even know it. I think there's one. I don't know. But also one thing I really want to say is if you're building something cool, please reach out to us and let us know. We'd be happy to share it on here. I would love to bring back RNR RLR, Real Life React Native.

Robin Heinze:

Real Life React

Mazen Chami:

Native. I would really love to bring that show back. And the idea was that we had a set of questions that we weren't going to veer from and just ask those questions. We will tweak those to include some

Robin Heinze:

AI questions

Mazen Chami:

Within

Robin Heinze:

That. Yeah, but I would love to do AI. I would love to just hear from people who built a React Native app from scratch with agents and just hear how that process went. So if you are one of those people, yeah, reach out to us on Twitter. That's probably the best, most direct way.

Mazen Chami:

And check out Frank's Twitter. He's also built an app fully ... Well, he built it and then had AI rebuild it and it did a good job for it too. So check out his Twitter for that. All right. We're at time, Robin. And as always, before we go, do you have a mom joke for us?

Robin Heinze:

Yes, I

Mazen Chami:

Do. With hesitation. This is going to be fun.

Robin Heinze:

I was choosing. I was

Mazen Chami:

Picking.

Robin Heinze:

Which one?

Mazen Chami:

I thought you were cringing as you were going.

Robin Heinze:

No. We have a very active jokes channel in our internal Slack and that's usually where I get my jokes from. And so every time we have an episode, I go choose one of the ones that's come up lately. Okay, but this is a good one. I have some really good advice. Never scream into a colander because you'll strain your voice.

Mazen Chami:

Okay. That's a good one.

Robin Heinze:

That was from Darren. Good, good, good one.

Mazen Chami:

Thank you, Darren.

Robin Heinze:

Yeah.

Mazen Chami:

Well, thank you all. And like I said, please reach out to us and we'd be happy to hear from you and actually even probably have you on the show. And that's it for the episode. Yeah. See you all next time.

Robin Heinze:

All right. Bye.

Mazen Chami:

Bye.

Jed Bartausky:

As always, thanks to our editors, Todd Werth, Tyler Williams, and Jed Bartausky, our marketing and episode release coordinator, Justin Huskey, and our guest coordinator, Mazen Chami. Our producers and hosts are Jamon Holmgren, Robin Heinze, and Mazen Chami. Thanks to our sponsor, Infinite Red. Check us out at infinite.red/radio. A special thanks to all of you listening today. Make sure to subscribe to React Native Radio on all the major podcasting platforms.