A top iPhone app developer reacts to Apple's new Siri

This episode is brought to you by Accenture. When your advertising operations fall out of sync, everything else follows. Spotify and Accenture are working together to reinvent the rhythm of ad sales. Using automation, analytics, and smarter workflows to simplify campaign delivery and access better data across the business. The result? Less time spent on operations, more time connecting brands with the moments and fandoms that matter most. Learn more at Accenture.com slash Spotify. This episode is brought to you by Google Chrome. You think you know a browser, but Gemini and Chrome? That's new. It can help you with practically anything on the web, like restoring a vintage motorcycle from a 50-page restoration block, or finally break down that long article you've had open for weeks. Gemini and Chrome is here for it. Ready to make anything online make sense? There's no place like Chrome. Check responses, setup required, compatibility and availability varies 18 plus. You've heard Tim Cook say it a million times. They're not first, they're best. In this case, they're two or three years behind. They wish they were one year behind and that's historically where they are. They're pitching this local and private thing. I don't know if anyone would care or if it would change adoption versus Chad GPT or Claude or Gemini if it was like we sent it to the cloud and you just have to trust Apple. Like no one cares. Ryan Jones has been one of the top indie app developers in the app store for more than 13 years. He's watched new design aesthetics come and go. I don't like it. I'm going to say it right now. I don't like it. Played by Apple's ever-changing rules. You expect for Cook, his app store in particular, and that 30% cut it can take. Sold beloved businesses and even won Apple Design Awards. Now he's building Flighty, one of Alex and my favorite apps for flight tracking. So who better to discuss Apple's latest updates from WWDC, the new Siri, and where vibe coding is actually headed. Oh, is this the scoop for the show? This is the scoop. All right. Access, right? God bless. God bless. I'm Ellis Hamburger, founder of Meaning, a storytelling studio for tech companies. And I'm Alex Heath, founder of Sources.News, your home for scoops about the AI race. Welcome to Access. Ryan Jones, welcome to the show. How's it going? It's good. Been a whirlwind the last couple of days between meeting Apple people, Flighty fans, trying to figure out what all this means for Flighty and me as a user and the world. It was good to see you, Ryan, at your dinner you do after WWDC every year. I feel like that's become the quiet power center of indie app development. It's always fun to pop into there. We're not supposed to talk about that. So you're banned from now on. OK. OK. It's fine. It's fine. It's fine. You've been blacklisted. I mean, I'm sure the guest list is pretty popping, though, because you've been doing this for longer than pretty much anyone, I feel like, at this point. That is weird. I was thinking, walking around Dub Dub on Monday, that I wouldn't say we're the old guard, but it is weird because I still feel like we're brand new. But this is year seven of Flighty, and I had Weatherline for four years before that. So I guess I kind of am. It's crazy. So going back to your design aesthetic and the way that you've leveraged all the new iOS stuff over the years and won the Apple Design Award, what do you make of the current state of iOS design and the new Liquid Glass updates? I assume you're one of those people who's like, we want to have our own point of view with the design of this product. Do we really want sliders letting people adjust everything? I feel like some designers in the space are probably saying, oh, that's like admitting that Liquid Glass actually isn't a correct idea. What do you make of all that and the new changes? Yeah, iOS 27 just got it this week. Yeah, I think you nailed it, Ellis. This is a classic, users ask for a slider. Users say, I want to figure out how much transparency you want, and you should not give it to them. They're not asking for what they want. They want you to make it better because they're not going to know which is the right choice to make. And you just introduce all kinds of complexities when you put in a slider like that. This is. It's a whole design theory thing that would take a while to get into, but more or less the designers who are the experts who know how this decision of more transparent or less transparent plays out across all the different depths of the UI and the interfaces, how that actually matters. And when you give people a slider like that, you're kind of, in a way, in a very short phrase, you're showing that you didn't think it through entirely a year or two ago and figure out exactly how to do it. How it all played out. And you kind of got yourself in situations where you can't just pick one because it doesn't work across all the different depths of it. So you have to let people pick, and then they're going to run into scenarios where it's not great when you get into whatever two forms deep with a close button on top of another close button that's also glass, those kinds of things. Hardcore. Yeah. But you're exactly right that we wanted to have a design aesthetic that is our own, but feels at home on. Yeah. IOS. And we appreciate some of the refinements that they did. To start with the keynote a little bit, the first segment was 15 minutes of more or less like, hey, we need to clean up some of this liquid glass stuff. I actually thought the whole keynote, to give like an overview point of view, when you think about it, kind of the whole thing was, we might have made some mistakes. Let's do a little bit of cleanup. Like, okay. On the liquid glass stuff, we need to clean the UI a little bit. We maybe went a little fast, so we have some bugs, so we need to do some bug cleanup. Maybe parental controls wasn't working super well, so let's do a little cleanup there. Maybe Siri has, whatever it is now, 12 years of broken promises. We need to totally revamp that. So I thought it was a pretty. I think humble was the word Gruber used last night on his show, and it's spot on. Like, the whole thing was a pretty humble, hey, we're going to do some. We're going to do some cleanup across the board. And it was nice to see it on liquid glass because we've been running into problems kind of introducing it. They had this rule last year to give an example of, you weren't allowed to do glass on glass. So for example, the tabs on the bottom of an app, if you're going to make those in glass, then in Flighty, you have a sheet or card behind it with all the flight content and that they were like, no, that can't be glass. You can't have glass on glass. And this year they're like, whatever, just make it look good. So they've kind of loosened the reins on that. And then they've done a lot of work on that. And then they've done a lot of cleanup on Mac OS and then across kind of the edges of liquid glass where stuff gets blurred and you can't even really tell which layer the button is on versus the content behind it. So there's been a lot of great cleanup in that area. Literally the edges and corners they made an announcement of, which was pretty funny. And then some, of course, were saying that the corners actually weren't properly rounded. And every window on Mac OS now has the same tighter corner radius, ensuring greater consistency, across all of your apps, even if they haven't been updated. Ryan, we were we were both at Apple Park in Cupertino for the keynote. And I was noticing before you posted what you were going to look for and you said a cohesive and thoughtful A.I. strategy where Apple will get its 30 percent in the new Siri, how they're going to let apps monetize. How good is the LLM and then any Ternus impact? Did you feel the turn or not yet? Turn was a zero. It was zero. I mean, he was very much there. He was like taking photos. He was at the media reception the night before. But it was it was the final cook show, which, you know, I guess is to be expected. But it makes sense. Yeah. What surprised you relative to what you were expecting going in? And yeah, what did you make of the overall story? Yeah, this is kind of my channeling, my Alex reporter angle here. What am I looking for going into it? And it's helpful to kind of remind yourself because when you come out of it, you just think about what was done and you don't really think about like, wait a minute. They didn't say anything about vibe coded apps, that kind of stuff. I was surprised by that. I wasn't as much. I think that's a thing that's so new that they don't really have their hands around it yet. You know, they they think slow. They take six to 12 months to kind of think these things through. And it's such a Titanic of a ship. If you put out a rule about any of that stuff, you can't just later a year like you can fix liquid glass, you can fix bugs. You can't really change app store rule six months later. That's just not being a good platform anyway, not to go into that too far. But I did think they had a good cohesive AI strategy that plays to their advantages. You know, they're not a frontier model, they're not playing an agentic world. And that's kind of what I would hope, kind of what I hoped to see from them. I want them to recognize the advantage we have is one. One button allows Siri to index all the stuff on my phone instead of open chat and authenticate with Gmail, authenticate with G calendar, or those are the same one. But you know what I mean? Almost makes me want to go back to using normie apps like I did last year because Siri will have access to all of it. The thing that yeah, this is kind of jumping into Alex's kind of world, but I've been thinking about how this is kind of where's the end state of this and I I'm starting to think tell me. What you think of this, that they're real key that they have that other people don't have is I messages and photos. That's the thing that they can index that nobody else can index, at least in the States. Yes, that's true. But same theory applies like if you just put. put in WhatsApp for the example. Um, those, so you're going to have iMessage and photos if you use Siri AI, but these big companies like Google, Facebook, I know it's the same company, but WhatsApp, Instagram X, those guys aren't going to donate their content to Siri AI to be indexed. You think in a way you're going to have wondering about this. Cause in my Siri briefing, um, the other day they were explaining how it works and it's opt in to the index, right? And what we're describing for those who haven't seen it as the new Siri can appear on any screen and know the context of the screen. And it can do it in any app at a very basic level. It can understand texts, um, just like raw text and images. It doesn't understand UI or interactive elements. The Siri to ingest. Apple told me, uh, unless you know something different that this index is private. Personalized on device never goes up to Apple for training or any kind of like data mining. And if that's the case, I actually think developers might, even the big ones, they may opt into this unless they really just want to strategically stick it to Apple. You don't think so though? I mean, I, it's really hard for me to imagine Gmail saying, yeah, you can ask Siri to find your email. They want you to open Gemini or open Gmail and press the Gmail, press Gemini inside of Gmail. Yeah. Cause they're charging the subscription. They'll eventually have ads in there. And it ties into everything for them. It also just might be a big, like rebuild of how they even store data, right? I mean, I use poke all the time to add things to my notion pages and I'm not sure how those are stored, but I doubt it's as straightforward as Apple's local notes database on your phone or something like that. Yeah, that is. So one of the kind of, um, super nice things about this is that it's super easy. I mean, I've been using it for a long time. I've been using it for a long time. They announced this intense and entity model. So at a very simplistic level, Notion would donate is actually what the word they donate to Siri say, here's the text. No way. That's the word? That's the word. And it's actually makes sense because, um, it's not a two-way connection really. And Apple doesn't really have to use it. So you're there. Cause if you get like some spammy gross app that donating stuff, Apple's going to kind of ignore it. So it's kind of like here, use this if you want. Uh, it's, it's interesting. They're always super careful about the words they use with that stuff. What is the vibe with developers about this? I mean, do you think people will take a wait and see approach? Would you, are you going to let flighty, you know, hook into the index? I would love to, there is no flight entity. So because they have this entity, construct that Siri has to work with, that a developer has to offer up there. We're using that word and I'm like, w what the hell is an entity? I don't. Yeah. So imagine there's, imagine there's a layer, a horizontal layer at the bottom of all the apps. And then the next layer above that is Siri figuring out here's everything I know about you and it's going to tap into the apps below. And the way Apple does it, cause it's extreme control and all those things is they defined, let's say a hundred entities. Like, okay, if you have text, tell us what text you have and we'll index it. If you have photos, tell us what photos you have. And we'll have another agent index that if you have voice, tell us what that is. And we'll have someone index that they don't have a flights one. So we can't say here's Alex's flights. So then when you go up a level above that to Siri AI, if Alex says, what flights do I have coming up? It knows about the ones in Apple first party stuff, but no developers have been able to donate up into the context about you, your flights. And that kind of the, you asked about the developer thoughts. I think this is an example where Siri AI in a way is like GBT three GBT 3.5, because this whole concept of tell me exactly what you have and I'll then index it is not what these models don't need that just connected. They figure all that stuff out. So they're in a kind of, it's kind of a couple years old of you have to tell me what all the content is, and then we'll index it. Versus all developers are like, yeah, you could do a flighty MCP node and I could use Claude to control everything already. I mean, I don't think you do that currently, right? But you could, and it would be way more useful to the end user than what Apple's done. Yeah. And in an example for the audience, like APIs are very deterministic. You say, here's a flight number or I'm going to give you a flight number. Here's the flight number. MCP is not like that. It's like, here's a bunch of content. You figure out what it is. And the beauty there is the app that's donating it to use Alice's new favorite word can change the content over time and doesn't have the other person that's reading. It doesn't have to say, oh my God, we didn't expect a flight because this AI is so good. It just recognizes it. So it's way more fluid and faster evolving, but that's just not the way Siri is. It has these deterministic 100 things and flights aren't one of them. So we've filed our necessary complaints and raised our hand and said, we would love people to be able to ask flight questions. And. Fingers crossed for next year, I guess. Can you also help me understand, like, is this actually the holy grail for search? So I think, you know, a company like Apple, in my view, has always been historically awful at search, whether it's with mail, it's with messages, uh, notes, especially that's the reason I don't use the notes app anymore. It's because I can't get through anything in search and, you know, AI isn't very good at that historically. Like they use the retreat. We've logmented. What is it? Alex, the rag generation. I mean, is this actually like a new, uh, way to make your entire phone searchable in a better way? Like, is that new here or is it just kind of the same as we've had just giving it more and more stuff? I am with you a hundred percent. And then now Alex read my tweet of stuff I'm watching for, for the last 10 years, I've written on their search that actually freaking works and cellular. Fallback that actually works. And then the one year I didn't write it because I've given up. They, in theory, did both. Um, you're, you're totally right. I have no idea if they've actually executed it. It was in the section of the keynote that was frame it. Hey, stuff, here's all the stuff that never worked, that we're going to make it work. And search was one of them and they kept hammering it in a way you could tell in their tone that it was like, we actually make it at work. We promise. And I think I've had, I've downloaded. I was 27, probably you guys have too. And it's doing this indexing thing and I'm on day three of indexing and it is still indexing. So you're, I haven't done it yet. You're still indexing. My impression is that they're like, we're going to build a huge local index. Yeah. Here's mine still going a huge local index and actually make this stuff good. I just don't get this approach. I mean, I get it from Apple's POV because they're the on device privacy company. And they clearly think that this is going to be a strategic differentiation for them. Uh, where, where I is going, but it's just, it's very cumbersome. I mean, the fact that it's taking days to index your phone when I'm doing a currently right now, our plus long over 70 step multi-turn prompts with using five MCPs and Chrome use on fable. Like that's where the world is right now in Apple's taking three days to privately index your phone. I just, there feels, it feels like there's a big disconnect, you know? Yeah. That's where the nerd world is right now. I mean, I'm with you, but you've heard Tim Cook say it a million times. They're not first, their best, um, they're, they're never, they're always there. Um, there's a lot of interesting strategy angles we could take here. One is I totally agree with you that they're pitching this local and private thing. I don't know if anyone would care. Or if it would change adoption versus Chad GPT or Claude or Gemini, if it was like we sent it to the cloud and you just have to trust that like no one cares. I don't think anyone cares. And I brought this up to Apple execs I met with Monday and you know, I think it's a hunch they have that people will care that there'll, there will be a Cambridge Analytica moment for AI. Uh, I mean you're seeing, you're seeing some of the lawsuits about chat GPT already and I think they don't like that. There was a lot of subtle shade. I think. Yeah. I think that's a good point. value in my life. I fly way too much. And for those who don't know flighty, go download it and play with it as we're talking. Cause man. Yeah. Is that a, something your wife says, Alex, Alex, I think you've been a bit too flighty lately. This is the third time up to San Francisco. Ellis says the naming guy. I was worried about that, man. My initial spreadsheet of all the names flighty was like, that's a great name, but is it going to get attached too much to the other word? I don't know. It's great. Flighty is perfect. I mean, that's kind of the ideal in my mind, especially if you're in a category that isn't mega crowded, you have a bit more room to experiment. Even with headlines that I write every day, so often trying to take a familiar phrase and twist it a little bit or a familiar word and twist it. And what that does is it borrows the memories and feelings that people already have. Like the name already sounds familiar, but a little bit different versus, something that they have to hear, uh, completely brand new and build new associations with. So yeah, I think flighty is great. I mean, the, the bigger question with names like that, do you have any clones that are flightier or flights that are accidentally beating you in SEO? We had the first flightier and he literally just stuck an R at the end of flighty. It was just T Y R, which is like the easiest trademark claim of all time. So he got disappeared pretty quick. No one's tried to vibe. No one's tried to vibe code a flighty competitor yet. They're trying. Yeah, it started. I mean, as a product founder, it's so depressing because I really try to contribute something new. Whenever we like work on a certain feature, we look at what's out there and then bring our own thoughts to it and try to contribute something new. And it stinks to see people just straight rip it off and not add anything or try to like bring something new to it. Um, so I've kind of distanced myself from it. We got a good IP lawyer and I told him, don't go after people that are just, you know, having fun. But if someone's getting serious and they're copying us, just make it go away and don't even show me because it's depressing. I think it's nice. You have a user base of people who deeply care about quality and accuracy, who cannot miss their business flight to an elite dinner, which as we've discussed is Alex twice a week. So yeah, there's that as well. Well, and correct me if I'm wrong, but I mean, the flighty UI is obviously, you know, great. You're an Apple design award winner. The real edge you have is the data. Is that, is that fair to say? I think so. I think it is designed to though, like it's the first thing people say. Um, I guess that's credit to the data being great. If the data wasn't great, that would be the first thing they would say. They would say it's pretty, but it doesn't, it's wrong. Um, you know, when we were talking about or initially starting the company, um, I was pleasantly surprised at how many people said, there's a couple apps out there that do this, but they're just so hard to use. Cause I knew that we could bring a fresh perspective to that and like put it out there. And I think that's, I think it's great to bring a fresh perspective to that and like put in the work to make it really user-friendly. Um, and I didn't know if that was the thing people just said or would actually respect and understand when they used it. But I think so many people have experience with airline apps and that's what they kind of benchmark it against. Um, that when they use it and you know, you don't have to hunt around and guess where features are that it's kind of a breath of fresh air, but that along with the data has to be, has to be great. And yeah, that's, that's the core. And I think that's kind of harder part to copy, which, you know, I guess someone could do it over time, but it's going to take them a few years. And, um, we're starting to build real expertise and advantage in that a lot with the AI stuff we can do these days. And that that's the key thing that we see with vibe coded apps that pop up. People will say, bloody so expensive. Someone's going to vibe code this. And then it pops up and it's just so inaccurate, just bad. And we see constantly be like, we'll see a little app pop up someone. They'll tweet about how they're going to take on flighty. Then a week or two later, the person's like, wow, I had no idea. I am out. And you're like, yeah, this is, uh, this happens every six months for the last six years. You're right. I mean, air airline apps generally do suck. I will say I'm a, I'm a United. Delta the only good one. No, you're two years behind, man. Yeah. No Delta. Shit. Past Delta, man. Delta's Delta's very like mid, mid 2010s now. Okay. Don't mind me. United has really upped its game with its app. I'm a United loyalist. I do the Burbank SFO flight weekly. Um, and I will say like, I'm impressed with what they've done. It's the most, you know, full featured airline app, but you know, last night was a great example. I was coming home late from SF back to LA and it still didn't have like the terminal that I'm supposed to go. And, but SFO, it's a little confusing. I mean, you can get between them, you can walk between them, but if you have the Uber drop drop off, it's much easier to just go directly to the right terminal. And I'm looking in the app everywhere on the boarding pass, everything. And it has the gate hadn't been assigned yet and it didn't have the terminal. And I was, it was like an hour, maybe a little bit above an hour before the flight. Uh, and you know, I was just like, man, why am I even looking at this? I got to go to flighty. Of course, flighty habit. So you still have an edge even on United. Good to hear. Yeah. There, I never would have thought I would have lost a lot of money. If you told me seven or eight years ago that our biggest competition would be United, I would have to be like, absolutely not. They've got the live widgets, they've got activity stuff. I mean, they, they're really embracing all the iOS stuff in a way that's surprising. Yeah. A little, a little business insight. Scott Kirby said he's the United CEO. He said on, um, Stratechery that they spent around, this is just rough, um, like a hundred million dollars a year for the last, five or six years to get their stack right in order to make the app great. So this, that's kind of an insight into, they could just buy you for a hundred million dollars. We talked to them. They, yeah, they, they're smart. They, um, you know, he made comments about buying us, uh, publicly and, oh, he did it. Yeah. He said it on Stratechery. He said we should just buy flighty and just set those guys up in a office and just have them tell us what to do. Let them keep running flighty and just tell us what to do. But the magic of flighty is that your airline, airline, It's agnostic, right? So that would break that. Yeah. You'd be going in house. It's like open AI buying TBPN. Shout out. No comment. I like this guy. Good job on the analogy. Yes, I do. Um, and people, that's the first thought that they have with flighty. They're like, if the airlines try to buy you and if you think about it for a little bit, it doesn't make any sense. Um, it's not the right fit. Um, yeah, it's, it's a good person to work with maybe to partner with. Um, the good news is everybody in this kind of, you know, in this kind of, you know, flight tracking game is starting to get their data in better order and understand that it's not that great to just silo. Like, let's just start sharing data and see if we can get this stuff to be more accurate. So from a partnership, it makes a lot more sense for airlines. I guess that's part of it. Uh, a lot of them who haven't traditionally revealed their data, even to like Google flights, they're just trying to retain any semblance of differentiation they can. Right. And I guess you've seen firsthand kind of like who is and who isn't sharing their data over time. The best way I describe it is it's a spider web. People think you just subscribe to API. That's like, let's get the, all the flight information. And that is not how it works. There's multiple intermediaries. Some of them have this, some of them have that. Nobody has this four people have that. And the thing that it causes at a very high level is there's holes where nobody has, but then it's kind of almost a harder technical problem when multiple people have one thing. Like let's take Alex's flight yesterday. Someone was probably showing us one terminal. Another data source was showing us another terminal. One was showing no gate and one was showing another gate. And he knows the answer because he he's in the airport. He can go walk and find the answer. But as a server or as a cloud service, we had to figure out which one's the truth. And it's very hard to know which one of those multiple inputs is the correct one. So this is why your balance sheet has that strange, uh, 5,000 interns line item on it that are on the tarmac at every terminal around the world. I mean, it is, it's silly when you're at the airport and you could see the data in front of you, but not on your phone. Some of these old wires just don't seem to be fully connected yet. Yeah. And the weird thing that you're doing in your brain and you don't even notice it, or you don't even really know is people will go look on Google flights and then they'll go look in the United app and then they'll go look in flighty and then they'll ask chat GPT or whatever people do. And then you find out, you find the one that's accurate. And then you think like, why don't the others have this? It's like, well, that's the problem I just described. We don't know. That's the one that's accurate. Everyone's telling us different things and figuring out which one's the one that's actually accurate is the hard part. business. The result, less time spent on operations, more time connecting brands with the moments and fandoms that matter most. Make sense? Check responses, setup required compatibility and availability. Various 18 plus. Well, so is there anything in particular you're looking forward to using from the new iOS? I don't know if it's going to make it any easier to grab people's flights from the actual local mail app. I feel like you do that through Gmail right now, which is one of my favorite features. Anything else you're excited to use? To be honest, because that flight entity thing is missing, it's extremely limited. It's really tough. That is wild to me that you have to petition. All right, change.org, all our listeners, add the flight entity to the iOS 27. The fact that Ryan has to do this, that he has to petition Apple to create a flight, quote unquote, entity to let Siri look at it when MCP exists is just so wild to me. It's a little old school. I would not, Matt, I would not be surprised at all if iOS 28 is, we don't have to do that anymore. It's so easy. You just donate the contents of XY. A developer marks what's private and what can be donated. And you don't have to do the entity thing. We just read it and we made Apple intelligence so smart. And now it knows what a flight number is. It's coming. We have reinvented folders. 11. Well, there is one thing that I was looking through in the notes that I thought you might be able to use if you had any interest in it. Not that I'm recommending it at all. But I saw that developers. Have a new way to reduce subscription cancellations through retention messaging. Did you catch this? I wonder what that actually looks like. I did. And we use it already. We were one of the chosen few who got to use it over the last year. It was kind of introduced last year. And as like a opt in beta thing. And we've had it running for about a month now. I can't say that I've seen any meaningful impact or change. But what else is talking about for everyone is. When you press everyone knows when the manage subscription screen is now in your app store and when you press that button developers now have there's pretty limited as usual for that poll like you can have I think it's five hundred characters and one image and you can say are you sure you want to cancel this you're not going to get delay notifications from flighty anymore and calendar sync is no longer going to work and then you can offer like no I don't want to cancel anymore. Or. You should try the monthly plan because it's lower price that kind of stuff. Yeah. I guess the larger thing here that has been a kind of an undercurrent that may be interesting to you guys is. There's it seems like there's excitement and reenergizing at the app store level of what can we do to kind of embrace the fact that AI is happening a lot on the phone and is happening a lot in apps like two years ago I had this worry and all of a sudden I was like. I'm going to go back to the app store. about what the technology can do just build features and then add the technology into it i think that's a little problems first it's a little bit naive like you kind of have to understand you have this new tool set like i used to have a hammer now i have a jackhammer like i can do different things with that now and so you kind of got to have it in the back of your head so we've built that kind of skill set over the last two years and it is a completely different kind of way of thinking so i'll give the feature that's coming in a month um as an example just to kind of land it for people but oh is this a scoop for the show this is a scoop this is all right yeah we'll take it access right it's god bless so when you land let's say you're making an international connection and sometimes you're like i don't know am i getting my bag beforehand am i going through security am i doing passport control am i doing it at the final destination did i do it beforehand all of those steps am i riding a train can i do security here all of those steps during a connection which is a very high stress point for a lot of travelers previously if we wanted to make a feature that would help users get through those steps it would be i mean i don't even know how you do it you'd make a huge matrix of every single terminal at every single airport for every single nationality for every single final destination because those are the things that go into like ellis is a u.s citizen so he does different steps when he comes into the u.s but he does different steps if he goes to china as a u.s citizen so there's infinity variance to these things um and it would have never been we would have never even approached the problem of let's try to tackle a connection guide when someone lands like that but this is right up the alley of what ai can do so we've kind of built the harness that goes on a model of what terminal and gate are you coming into what terminal and gate are you going out of then what is your passport nationalities and you can input multiple if you want and then what's your final destination because that determines what kind of security you have to go through and then when people land in an airport we give them it's not we don't think people want turn-by-turn directions and you know you don't want to like turn-by-turn map you just want the big landmark stuff you've got to do so it'll say for example walk five minutes to the tram take five minutes to the airport and then you're going to take it going south two stops go through security then go through passport don't do the bag recheck and then go back to terminal three and it tells you the exact steps of those kind of what do i have to do in this connection world and then gives you the typical time that it's going to take a user to go through those steps alex is under your spell right now i can see his face ryan you are going to make this grease trip i have in july is this launching by mid-july for you sure okay please because i am terrified of what looks to be travel airport nightmare in greece right now eu has instituted a new kind of passport control system yeah yeah like they they have to do a biometric biometric capture phase and then you have to go back into the line the lines i mean you've you look on tiktok it's like hours and hours of lines it looks i'm like i'm borderline considering just paying for one of these vip services yeah yeah someone has scored us through all this uh so i may end up doing that but this sounds awesome and is going to have a lot of help me not feel stressed about my grease trip uh so thank you yes yeah it's got some cool you know it knows um again in the nationality context in the in the uh in europe they have these e-gates which is a faster way to process and only certain nationalities can do the e-gates so it knows that um what other cool stuff does it do if you put in your seat number we try to we have the model figure out how long it's gonna take you to get off the plane so alex obviously be in 1a and it'll take him a second to get off and of course me and ellis will be especially on united yeah you know yeah i'm down in steerage it's gonna be another 11 12 minutes for sure yeah alex is already sipping his latte and i have a child that is like a monkey on each of my feet so that also adds uh 21.6 percent uh increase in in transit time yeah we'll have to teach teach the model about your children that that's actually funny you could ask that non-boarding yeah this will make our predictions a lot more accurate just add 10 minutes to everything that's very cool uh this has made me realize that i want to know what our flighty passports are like for the year ryan uh how much have you been flying this year i want to know if i have you beat oh you know i have lsp i'll check mine as well so for 2026 i've got 39 flights 40 000 miles uh 14 airports which means i'm i'm dude i'm already about to surpass you just go to the same airport it's my word i only have 15 000 miles at 12 airports you just go you're just going back and forth between sf and i do a lot of sf uh but i've had a couple long hauls and i'm gonna have two long hauls in the next month this is the story of a founder all my passport is screwed up because a user will send in a like oh this is not working correctly with this flight and i add it and i look at it and i want to look at it myself so my passport and my flighty is just an absolute wreck the friend view you have now because i used to add my friends once to track them or like my wife if she was traveling without me and the friend view now is is amazing because you can see people's flights without adding it um but yeah i mean i i was also thinking is is that an opportunity for you to create more of a social experience around travel because you have flighty for moms you have paid a thousand dollars a month for this to know when their kid's flight is arriving yeah there's this social aspect is a thing that we have in internal debates about on the team. It's kind of, I guess, it's kind of like our, we'd be very, we're very careful about it, kind of like Apple is with safety and trust stuff. I'm kind of the one making us go a little slower. The other two guys are like, let's do more, let's do more. So we try to find a balance between, let's not do a leaderboard of like who can fly around the world the fastest, the most, versus it is interesting to know kind of what your friends are up to. So I think you'll see us go a little further in that direction, but we don't want to go the full, like you're ranked number 12 in the world kind of thing. It's just, it's just not, I don't think that's super long-term interesting. Maybe we could land it here. Land, ha ha. I didn't, I didn't even plan that. Not funny. Okay. Thanks, Ellis. Ryan, you, to me, exemplify this era of indie app development that felt like the golden era where you were, you, you had, you had another line before this, sold it. You've had flighty for seven plus years. You've got a team. It seems like the business is going well. You really think about the craft and the design and you focus on a platform, right? And there's, we were talking earlier about the other developers we all know in this community that have made great careers on this. And I've been thinking about like the future of development and app development with vibe coding, what we've been talking about with stuff like what Wabi represents, but also just lovable, all these tools, cloud code. And the idea that, you know, as like Patrick Carlson, I think said, like software is going to become like pizza, right? Like, like made on demand and infinitely personal personalized. So if you buy that vision, which I don't know if I do fully, I think craft will always matter and always differentiate, but there will be this explosion, this Cambrian explosion of people creating their own software. How are you thinking about this kind of amazing career you've had in the next 10 years? And if you had, you know, someone you were mentoring who was just coming up and trying to figure out, can they be the next flighty? Can they be the next weather line? What advice would you give them when you look out at what's happening with vibe coding? Oh man, there's a lot there. Those are good. Those are good questions that keep me awake at night. I'm not totally bought into that. And I try to be really careful that I'm not like just speaking my own book or whatever the correct analogy is for like in my own, oh, apps are the most important thing in the world. I'm not totally sold on this UI is going to be created on the fly thing. I think if anything, this is tapping into Ellis's world. If anything, it'll just get completely skipped over and everything will become voice or audio. I have a hard time believing that normal people are going to interact with AI through text. I think nerds like us are good with typing a lot and reading a lot. But if you just look at this, the numbers of, of TikTok and YouTube versus Twitter, people aren't text first, their voice, video, it's the native things that you can see in here. So if anything, I could see a world where we kind of skip over this. But I guess that also plays into the generate, you could generate UI on the fly thing. People instead of one of the problems I had yesterday with the keynote was, or two days ago, there was a lot of just text and text and text. These answers that Siri AI was giving were just mountains of text, there was no images, they weren't even doing bulleted lists, they weren't even doing tables, it was just paragraphs after paragraph. And I don't think that's what a normal person wants to interact with. So back to the theory here, I think there's a chance that there's a little bit of that generate generative UI that happens. But my guess is that's probably a stepping stone to whatever the next paradigm is. And in terms of the next generation of apps, like I said, a minute ago, I kind of thought two years ago, oh my gosh, all this AI is going to happen in the cloud because you need everything to be processing on servers. You need the servers to be crunching the information. So we need Flighty to be this MCP callable thing that's in the cloud so that someone's Gemini can tap into it and talk to it constantly. And at the moment, it's not really going that direction. Apps are kind of hotter than they've ever been. So I don't know if I have I guess I'm wise enough now to not make a prediction that is ill-informed. I don't know if anybody knows where this stuff goes. I think, Alex, I've heard you talk about that. We just don't know. In hindsight, we'll look back and we'll say, oh my gosh, how did we not figure this out? But nobody really knows. But in terms of new people coming up, I think what I, you know, we're in the fortunate position that at the dub dub stuff, people will see my Flighty shirt and younger developers will come up and say, oh, my gosh, and take selfies and stuff like that. And it's just super surreal, first of all, and flattering. And then they'll ask me similar type questions. And I think the advantage or the thing that I would tell them is nobody knows, but this is 100 percent a real wave and a real platform shift. It's not the crypto thing that was kind of a head fake. It's not whatever was before that that I can't remember that all the people told us was going to take over the world. Like, this is a real one. Get on the cutting edge, understand what's going on, play with tools. Don't just sit there and read about it. Go do stuff. You have to be creating in order to understand how this stuff actually impacts. And there's not that many people that are willing to put in that work and do it. There's a lot of people willing to sit on the couch and watch TikToks about it and and get informed on it and kind of think about it. There's not that many people that are willing to put in that work and there's not that many people that are willing to get nerdy about it and get their hands dirty and try to kind of push us forward. And, you know, not everybody. We don't all have to be whatever Boris. We don't all have to be like the smartest PhDs in the world at Anthropic and Claude. Like there's plenty of or Anthropic and OpenAI. There's plenty of niche, small applications. And I do think that that is a thing that I don't know long term what the format of it is, but I do think craft niche subscriptions, there's never going to be a scenario where these big models can do as much as flighty can do with flights. And there's always going to be a world where there's people that care a lot about flights. So if you understand what these tools can do and you understand the tool set and then you find your niche or your people that care more deeply about it than everyone else, you can put those two things together, no matter if the format is apps or MCP or voice, audio, whatever that is. That's where I'm at with it. That's almost what I'm more excited about is not necessarily one billion vibe coded apps that are all kind of messy, but a million, a thousand true fans apps from people who are only able to make something beautiful that they're passionate about for the first time. You know what I'm saying? And I mean, it's not like you need to generate AI for every single use case on earth when there really aren't that many use cases, you know, in the average person's life. Ryan, we really appreciate your time. Great to recap DubDub and hear about where flighty's headed. And thanks for coming on the show. It's great. Yeah. Thanks, Ryan. Good to see you guys. Thank you for having me. And that's it for this week's interview. Stay tuned after the break for Alex and my reactions. Spotify and Accenture are working together to reinvent the rhythm of ad sales using automation, analytics, and smarter workflows to simplify campaign delivery and access better data across the business. You think you know a browser, but Gemini and Chrome, that's new. Check responses, setup required compatibility and availability varies 18 plus. All right, Ellis, good to chat with Ryan about all things Apple. As always, I consider him one of the forebearers of iOS design developer culture. And yeah, it felt really great to have him this week of all weeks with WWDC. Yeah, I mean, even going back to when I was a reporter, I mean, seeing him on Twitter in those early days, but we never really made the connection, but he always seemed like one of those guys who was just very smart. And savvy and cared about the right things. And yeah, we didn't really you didn't get to ask your favorite question on the show, I guess, because it hasn't happened yet. But I will be working with Ryan and his team on some story stuff over the coming months. And I'm just really looking forward to it even more now. I mean, he really is a mensch and seems to be thinking about things in all the right ways. And, you know, I mean, there just aren't that many like business operators, I feel like, who are going to make a billion dollars to seem like they're impacting the world in a good way. And that also are like really holding that idea that design is how it works so close to their mind. So yeah, nice to see how he's thinking about using AI not just to add a chat bot to flighty, but to actually make sense of a billion lines of information and make connections. And as he said, it's right up AI's alley. And I think that's really what needs to happen next for AI to appropriately come. Yeah. And I think we landed on that you should not change the name. The name is good. We'll stick with flighty. Yeah, great name. Great name. But there's some good tagline opportunity, I think, for flighty. Well, it's interesting, too. I mean, it all depends what the tagline needs to do. Like, I think he has the luxury of having, I would assume, a, you know, somewhat wealthy, somewhat savvy, travel pro audience like yourself that really cares about, you know, when it's good, when it's bad. And I wonder, you know, how much he's trying to grow his addressable market and to whom. Because I do think the pain points he brought up, especially around international transfers, are just about as painful of a consumer pain point as you could ever find. That's solvable by the world of software. Yeah, it's one of those rare but incredibly high stress, high leverage moments where if you have an amazing product that can solve it, it's worth a lot, even though those things don't happen that often for people. And it's cool that he's thinking more about that. What did you think about the Apple stuff this week? We haven't actually really gotten to catch up on it. I mean, I was there, Ryan and I were talking about being there. But I mean, I agree with him. Like, you can kind of see how Apple's taking it slow on AI. They are fixing things. Seems very much like a snow leopard year for people who remember that. But yeah, I don't know. What do you make of it? So I think my instinct is very much in line with what he said and what our old friend MG Siegler wrote about, as MG Siegler does, about Apple waiting for the right time and then swooping in with the well-designed solution that is palatable to billions of people. I would usually be of that school of thought, and I think it might actually play out. However, when Apple introduced the iPhone, there weren't a billion people already using pocket PCs and palm trios, right? It was still such a nascent category. So even if the business and tech people were like Apple's late, for most people, they were still early. Meanwhile, today, ChatGPT has a billion weekly active users, and they're building a whole lot of association and muscle memory in people's mind that that is where you go to for answers. Meanwhile, Siri has never been the place for answers for 15 years because no one can find a way to really make it work reliably or get the answers that they want. And so, you know, coming back to the storytelling angle, I would imagine that there is going to be a massive marketing campaign around this. And I wonder if they will be willing to be a bit tongue-in-cheek about it, about how it hasn't been perfect thus far. Because I do think there are a lot of examples throughout history. Domino's Pizza is a big one. We're acknowledging that the product wasn't great, actually is a huge step forward to rebuild that trust. But yeah, it'll be interesting to see how they roll it out. Because to some sense, what they announced this week is what they advertised two weeks or two years ago with Bella Ramsey that they got sued over. So yeah, to some extent, they already made the ads. But yeah, you could look at it one way like, oh, they're finally shipping stuff that works, even though it's not anywhere near what Claude can do or chat can do. Or the other way is, well, they're finally shipping what they said they were going to do two years ago. and that in two years would think about what's happened with the rest of the ai industry you have people like boris the head of cloud code which ryan referenced talking about how you know not only does he not you know really write code anymore he's not even really prompting anymore he's writing loops these recursive loops that prompt claude to code for him so that's on one extreme of what's happening and then apple's like yeah we have like better models than xcode we don't really have a pov on vibe coding and you know siri can work a little bit like chat gpt but it has no memory and we have this weird thing called uh intense or you know whatever where ryan can't even ingest you know he can't even send flight info to us which just feels like antiquated but here's what they need to do mark my words uh setting aside ai and the existing expectations of ai around question and answer and writing help that you know that's what they need to do mark my words uh setting aside ai and the existing expectations of ai and the existing expectations of ai and the existing expectations of ai and the world is deeply familiar with what the world is not yet deeply familiar with but apple has the capability to nail here is easy automations and they showed off some of those during the keynote like every time i leave work text my partner my eta that leverages a handful of on-device sensors and data in order to do that what they should do is they should have a campaign called there's a right in your notes app with ai which is probably the same or even worse than what someone is already doing by pasting their emails into into chat gpt or cloud but showing off all these kind of ingenious little shortcuts that really aren't deserving of being entire apps but can really simplify your life or at the very least catch your attention i think that's something that they need to do um because they do seem really well positioned to actually execute on that i mean i'm i don't know if you are uh but even with poke you know it's like they go in and out uh all the time yeah i the mercury one i use a lot in claude um and it's it has like an off problem where it'll it'll need to be re-offed uh every now and then right but it works and what are you doing with that uh i mean i use it to help analyze all you say show me the money show me the money literally it tells you how much money you made literally i mean it's built growth projections for you it's super helpful and i just got access to the sub stack mcp that is not public yet they're doing it in private beta and while we had this convo with ryan early christmas for alex wow oh my god dude i had uh i've had fable the new anthropic model which was a source of scoop that that was coming um this week uh running a hour-long like 70 turn 3 mcp connector at least plus web search prompt to create me a massive uh sources growth report strategy document using my sub stack data and mercury data and all of the above and competitor data and can't wait to dig into that after we get off here well that all sounds cool but did you look at any of the insights so far well no we've been in the podcast i'll look at it you you've been doing it in the background i started it right before we started the podcast and it just finished oh wow i guess i'll let you go then alex um yeah no i'm excited uh anything else going on in your world everything good uh yeah what's going on um playing a fun new game in my free time called mina the hollower oh nice to get into a very cozy handcrafted little switch game um i'm trying to i mean it's an interesting thing with ai these days there's i don't know if you ever feel this just kind of the the decision or even the opportunity paralysis of like i can do anything now you know and you could theoretically spend all your time working on your business i i have a sense i mean obviously you don't have kids yet but i have a sense that you spend a lot more of your free time on the biz than me whereas i'm just trying to find like a few minutes to relax and really does feel good to just kind of hone in on something that isn't procedurally generated or user generated that took someone years to make and just kind of enjoy having that in your hands you know yeah every everything in moderation did you see uh ocarina of times being remastered for the switch i did i did very polarizing trailer i feel like uh a lot of people are saying that it just looks kind of ai in the in the way that they rolled it out um don't ai zelda that's the new that's the new kind of derogatory term i guess that's it for this week's show thanks to ryan for coming on if you like this show you don't forget to like and subscribe everywhere you get podcasts we are access.show online and you can find us in lovely high definition video at access pod on youtube and if you really like this episode we'd love a five-star review or even better share it with a friend preferably a human and you can find my newsletter at sources.news you can find ellis at hamburger on twitter and at meaning.company for your storytelling needs and access is part of the vox media podcast network the show is produced by hook creators see you next week bye booking.com is the easiest way from a day surrounded by noise to a state surrounded by nature that's nice go on book it it's easy booking.com booking.yeah saying how you feel out loud is terrifying but that's what hinge's new free audiobook no ordinary love is all about it's a collection of real love stories about five couples that met on hinge the stories are written and read by temby dinton hurst nicola dynan curtis garner raven smith and rufy thorpe i hunter harris wrote and read the forward the audiobook's out now give it a listen