535: My Computer Supports Math

52:38I wish I knew and I was playing coy, but I'm not. But I don't know. I just feel like based on no facts and just a gut feeling, I feel like there's going to be a lot this year, even if it isn't the headset. I still feel like it's going to be a busy keynote. Obviously, the drop of Final Cut Pro and Logic Pro just last week, that could, you know, lead, lend credence to the idea that it's going to be busy. But I don't know. I feel like there's going to be a lot here. If there's no headset, how can it be busy? Explain that to me. I don't know. I'm not sure. I really don't know. But, I mean, as a couple of examples off the top of my head, it sounds like we're getting an all-new watch OS, or for some loose definition of all-new, which we may or may not talk about this episode.

53:26We should be getting widget support on iPad OS. I'm trying to think of other things. Presumably, there'll be a lot of improvements to SwiftUI. I mean, this is a developer conference, strictly speaking. So hopefully we'll see a lot of SwiftUI improvements. I'm not sure. I feel like they could fill an hour to an hour and a half with just normal software stuff. I mean, hell, as Chris Fanazzo says in the chat, no new features in macOS. That would need an hour and a half just for applause, right? That would be a crowd pleaser, but I don't know if it would count as a big keynote. I kind of feel like headset is the obvious one that would make this big.

54:01And honestly, if they have the headset, nothing else matters, essentially. like not that they're not of course they're going to have new versions of all their os's and the new versions will have nice features but if they have the headset that alone makes it this an extremely like these come rarely when does apple introduce new platforms not that often watch os tv os like they you know sometimes they're big sometimes they're small but this is so hyped so long room it's as if they were coming out with the car right all you need is that if that exists in the keynote it is a big keynote if that doesn't absolutely it isn't released in this keynote i think to make it a big keynote you need a lot of things that currently seem unlikely i think you would need the 15 inch macbook air to be announced you need all new versions of the os's with more than with some significant features that people care about and i think you'd have to probably throw in the mac pro that's what it would take to match the headset and of course it always comes back to the mac pro i don't honestly no one cares about the mac pro but but me but if you're trying to make a big keynote you need the things people do care about 15 inch macbook air all new versions of all operating assessments that people love swift with the new features new version of xcode and then you throw in the mac pro as a you know fun kicker if you remember the mac pro was announced at wwc we were all there i believe um the 2019 mac pro that is and i think that was like the star of the show in terms of they like leave that to the end like despite the fact that nobody buys them or whatever it is a it is a glamour product for apple to roll out because it's big it's powerful it's shiny it's expensive it's good looking if you like holes um and so i even i know people think like oh nobody cares about the mac pro that doesn't make a big keynote it does it would it would add to the bigness of it but you need all that stuff that i described and on the other side of the scale just the headset by itself yeah that's fair but i don't know i i've been talking a long time but suffice to say i get the feeling that whatever it is it's gonna be a big year which is why i really hope that the three of us get to be there.

55:59So Apple, call us. We'd love to hear from you. But no, I think it's going to be big. I think that would be excellent because I don't feel like the last couple of years have been bad by any means, but I think just a really firing on all cylinders kind of year would be pretty, pretty awesome. I'm excited. I'm a little bit nervous as a developer. I'm about to have my summer ruined in the best way. The same way it's like the Jonathan Colton song about you ruined everything in the nicest way. When you have a kid, it's like, oh, you ruin everything in the nicest way because you disrupt my life in a huge way, but you're my kid and I love you.

56:40Forgive the horrendous summarizing of this art form. I feel like this might be like that for developers in the sense that whatever plans we had for this year are probably about to be derailed in a pretty big way by the headset. You know, it seems, by all accounts, it seems like the headset is being announced at this event. And there's going to be a developer story to it because obviously why else would they announce it if that'd be quite a move at a developer conference to be like, here's a brand new platform. A sweet solution. You can make web apps for it. Yeah, right, yeah. So anyway, yeah, you can program it only with Siri.

57:19Oh gosh. I would quit the business. Anyway, so I think our plans are going to be disrupted. If our plans weren't, you know, wait for the headset, it's going to be disruptive. And I think it's going to be in a very exciting way. I don't know what the headset will be. I don't know what the headset's chances are on the market. But, you know, when I use the existing VR headset hardware that exists, you know, on the market today, the limited amount I have used with it with the exception of the Oculus whatever 2 that we have upstairs I'm super uninterested in it and it kind of gives me motion sickness a little bit it doesn't work optically well with my eyes for whatever reasons there's a lot of minor paper cuts or major paper cuts with it and I don't find the games that interesting but Apple is full of really smart nerds and many of them have apparently decided that what they have is something special and and worth shipping for for apple in a pretty high profile way so I'm optimistic, I'm tendinly optimistic that whatever they have is probably pretty good.

58:35because I trust them to make that call and they call it right most of the time. So whatever they have, we can look around the market and we can be super unexcited about all the other VR headsets out there and also trust Apple to say, you know what, if they're this excited about something like this and they're releasing it now, there's probably something there. It's probably really good in at least some ways for at least some uses and possibly more. So, you know, Because otherwise they wouldn't be releasing it. Apple holds back on a lot of stuff. So we can be fairly sure that what does get out there, especially a new platform in a high-profile release like this is bound to be, they've done their homework.

59:16They've probably built something really cool. And it's probably going to be really exciting when we see it. So that's why I'm very interested not in the VR or AR or XR space in general. I'm very interested in what the heck did Apple do here? because I trust them to, if they've gotten to this point, it's probably something really exciting and really good. Now, how many other people will think so and will be willing to spend$3 ,000 and deal with the whole butt-mounted battery situation? Who knows? We'll see when we get there. But again, I trust them to have gotten this right. If they do that, I don't think that'll be a thing to deal with, by the way.

59:51I think that'll be a feature that everyone will say, oh, of course they did it. I mean, we've all discussed all the reasons. It's just that once the product comes out, all the reasons we've discussed suddenly become like concrete and people say i always thought it was a good idea because the headset weighs less and you can swap batteries and so on and so forth but right now everyone thinks it's dumb right and then the whole rest of the industry will change to butt batteries and they'll be like well of course this is the obvious design we didn't copy apple exactly right yeah it was obvious we didn't copy apple at all i mean apple's not going to be the first one to use a battery pack external battery pack a bunch of other people not anyway yeah i do feel like for this for the headset again i get the vibe that it's coming this year if it's not i think a lot of people will be disappointed i think it's you know whatever like it would quote unquote justifiable it's like well apple never announced anything that was all on you was just rumors but anyway um the good thing that i hope the headset will do uh is provide cover for every other product and i know this is not the way it works inside apple but like it's plausible that there could be enough internal communication to let the other products that are announced every wwc all the new versions of the new os the new version of swift updates and frameworks the new version of xcode gives them some cover to not feel like they have to press press press to get their exciting feature out the door asap even if it's buggy because the headset is just going to block out the sun at wwc right yeah and so a lot of times at wwc i feel like i feel some groups are pressing they're like we really need to wow them in the keynote and they show something and we are wowed But then we realized like, oh, this isn't even in the beta build at WWDC.

1:01:29It'll be coming in a later build. And when it does come, it's buggy and unfinished. And it just ends up being frustrating and disappointing. And it's like, you don't need to try that hard. Ship what you're actually ready to ship. Like, I don't want to see anything in any of the other sort of platform stuff that isn't in the WWDC beta. because I feel like the headset provides them cover to be more relaxed with their, you know, with their choice of what to include and what to save for a future release. In the ideal world, I would have liked, you know, Apple to have the forethought to say, hey, you know, back in 2022, hey, WWDC 2023, we're going to ship the headset.

1:02:09So everybody else, when you're planning your new releases for the new versions of all the different OSs, everybody take a Snow Leopard, right? That didn't happen. like that's you know that they don't have that kind of coordination they can't tell the future that far in advance it's foolhardy to try to make plans like that i do think it would be wise for them to give each platform a snow leopard once in a while we've talked a lot about mac os not being annual if they could get the bugs out of it but it's just too big of a company to be that coordinated and making plans that far in advance about software releases and releases and headset most of you know the part of the reason you can't make that plan is that all the other groups aren't even probably disclosed on the headset if they don't have to be uh or subsets of them are because only some people who work on tv os don't have to know about the headset not all the people and certainly the people who work on tv os hardware probably don't need to know about anyway it's very compartmentalized to the detriment of apple internally i feel like sometimes and this is kind of one of those cases that we just assume that despite the fact that if the headset does launch wwc it will provide cover for lots of other os's the only thing that helps them with is what i just described which is okay now maybe they don't have to press maybe the higher levels of the org chart can cannot press to have the xcode team you know push out some feature of xcode that is buggy and going to cause the the app to crash all the time they'll say we'll just save that to next year it's fine that you don't have it and they won't know that's because the headset is coming but only like the people seven levels off the org chart who are disclosed in the headset do know that and they let those things travel downhill this i know this sounds byzantine and complicated but honestly this is the way it works in big companies especially ones like apple that have everything compartmentalized this is whole that's why reading the tea leaves is a thing that's whole kind of game of telephone of like what gets prioritized what gets given money and why uh by people who at various points in the chain don't know have all the information available they just know what the people above them tell them and what the people below them send back and feedback but they don't nobody has the whole picture until you get pretty high the org chart and it makes it very difficult to act in a coherent manner it's apple has an amazing job of it to be clear as a company they speak with one voice their their marketing their pr makes it seem like they are a coherent mass but internally it's lots of little kingdoms moving in all sorts of different directions pushed and pulled by forces they can't see and don't understand for the most part yeah apple can put that slogan in their in their recruiting material we are brought to you this week by collide collide has some big news if you're an Okta user, they can get your entire fleet to 100 % compliance.

1:04:44How? Well, if a device isn't compliant, the user can't log into your cloud apps until they fix the problem. It's that simple. Collide patches one of the major holes in zero trust architecture, device compliance. Without Collide, IT can struggle to solve basic problems like keeping everyone's OSs and browsers up to date. Insecure devices can be logging into your company's apps because there's nothing there to stop them. Collide is the only device trust solution that enforces compliance as part of authentication, and it's built to work seamlessly with Okta. The moment Collide's agent detects a problem, it alerts the user and gives them instructions on how to fix it.

1:05:20If they don't fix the problem within the time that you set, they're blocked. Collide's method means fewer support tickets, less frustration, and most importantly, 100 % fleet compliance. Visit collide.com slash ATP to learn more or book a demo that's collide spelled spelled k-o-l-i-d-e so collide with a k collide k-o-l-i-d-e dot com slash ATP collide.com slash ATP thank you so much to collide for sponsoring our show so in terms of like the overall like what i expect from the other platforms um you know it's hard when in a new platform year like what we have most likely here with the headset launching this is a year as john mentioned like this doesn't happen often this is a year of a new platform what tends to happen with apple when a new platform is being developed is a bunch of talent is pulled from other teams to work on that new thing and obviously that doesn't it doesn't all just happen in one year but what has probably happened here is probably a bunch of people from from other os's and other other projects in the company have probably been spending a lot of time recently working on the headset and so i would expect based on that for not only this year but also last year to have been somewhat slow and i would say that's kind of the case you know that being said apple's a bigger company now than they used to be and they are able to maintain things on multiple fronts better than they used to be they're still not great at it but they're way better than they used to be but anyway so i would expect this to be a somewhat slow year for the other platforms but that being said in terms of like what is actionable for developers to do with the headset right now i think you know we're going to see a we're going to see version 1.0 of this os we're going to see version 1.0 of the sdk and version 1.0 of all the frameworks and the ui and everything else on some level it's very exciting and very disruptive to our summers, as I was saying earlier.

1:07:19But also, there might just not be that much for us to do yet on it. You know, first of all, I don't know when they plan to actually ship this thing to us. But that's going to itself be a pretty big limitation. Like there's going to be a lot of things where we want to develop something for this, but we kind of can't until we have one. Or we can only do preliminary work maybe in some kind of simulator or something until we get one. Or maybe, you know, maybe Apple will use the new developer center that they built. They went up last year to, like, you know, invite developers out by, you know, by invitation only.

1:07:53Come out to the developer center in, you know, July and August. And you can sign up for, like, you know, a one-day slot to use the hardware and test your app on it. You know, they did that with the watch. So, you know, and they had, like, the whole dev kit situation with the transition to Apple Silicon. Like, they've done stuff kind of in this ballpark before. so I'm guessing that kind of thing will happen whether it will be open to many developers is a different question I'm sure some developers will get invited I don't know if it will be like anybody can sign up kind of thing I think it will be more likely don't call us, we'll call you that's much more likely just for scale but anyway, so I don't expect there to be a ton for developers to do on the headset yet this summer because presumably we won't have them yet We won't have access to them yet.

1:08:42And we probably will have a very early SDK. And so it's going to depend a lot on what you do. Like if you make an app or choose to make an app that is really a great thing in AR, VR, then you're going to have maybe a busier time. If you make, say, a podcast player, it might not be that important. And it might just be, all right, find a way to show a UI, but otherwise everything else is the same. And if the stories are accurate about them being able to run iPad apps in little Windows, it might not be that much work at all for developers to get their apps to at least run on the thing. And then once we have physical access to them, then we can maybe worry about making it good on the thing.

1:09:26Secondly, if we don't have physical access to them, neither do our customers. and so again it's like there's only going to be so much we can really do until this thing is out in the market or at least widely available to developers so that's going to hold back some of that so i would i would also hope that there is some nice to haves on the other platforms this is what we're gonna get to in a second and i think based on the rumors and everything so far that seems fairly plausible and also just based on you know what apple has done even when their teams have been busy doing other stuff you know on new platforms whatever they do tend to have decent ios releases you know mac os oftentimes you know is last priority and doesn't get much and and frankly that's often a good thing because when when they when they touch mac os without a ton of effort behind it they tend to break things and make them worse so you know that's it's good for mac os to not have not have a lot of updates when they're busy um so but ios i think is is likely to have some decent quality of life stuff there um that you know even if there isn't some kind of massive you know sdk overhaul some kind of massive new language or framework or anything else like even in the absence of those really big updates there tends to be a bunch of nice small stuff with new ios upgrades both as customer facing features and on the on the developer SDK end.

1:10:52A lot of little quality life improvements, little new APIs, single little modifiers for SwiftUI, or hey, here's a new view container that makes things easier for you. Stuff like that. That tends to get there in every release. So I'm expecting that for iOS. And then I think what we're about to get to is I'm actually very excited and interested to hear that there's pretty strong rumors that watchOS is getting substantial updates. Do we want to talk about that now? So yeah, this is reported in Bloomberg. what was this last month, I believe in late April, I believe it was a Mark Gurman piece that talked about how Apple is to upgrade its watch operating system with a new focus on widgets.

1:11:31And so Mark writes as part of watch West 10, Apple is planning to bring back widgets and make them a central part of the interface. The new widget system on the Apple watch will be a combination of the old watch West glances system and the style of widgets that were introduced in iOS 14 on the iPhone. The plan is to let users scroll through different, a series of different widgets for activity tracking, weather, stock, tickers calendar appointments and more rather than having them launch apps the new interface will be reminiscent of the siri watch face which was introduced in watch os 4 but it will be available as an overlay for any watch face it's also similar to widget stacks a feature in ios and ipad os that lets users pile many widgets into one and scroll through them as part of the overhaul apple is testing the idea of changing the functions of some of the watch's buttons currently a press of the digital crown launches the home screen for the next version of watch os Apple may have that open up widgets instead.

1:12:18And it is worth noting that apparently the Watch App Store has fewer than a million monthly users in whatever region they were going through some court proceedings. Apple disclosed versus 101 million on the iPhone. So literally 100 times more people using the iOS App Store than the Watch West App Store. I can't say I'm surprised. But yeah, I mean, I think the idea of widgets on the Watch is appealing. I don't know if I'm quite as enthusiastic as you are, Marco, but I will say that even as someone who has only dabbled in watch development, one of the things that makes watch development very tough, particularly for things like complications, is that you have extraordinarily little time as a developer to run code that will update information on a complication and other things like that.

1:13:09And so if widgets were given more flexibility and more time to update themselves such that they're getting updated more than what feels like three times a day, I know that's not literally the case, but that's what it feels like sometimes, that I would be more enthusiastic about. Like when I look down at my carrot weather widget and it's not, or excuse me, complication, and it's not carrot's fault. Oftentimes it is clearly behind, you know, especially on, you know, spring or fall days where, you know, the temperature can vary pretty dramatically from morning to afternoon. A lot of times I'll look in, in my watch will say, oh, it's, you know, like 40 degrees outside, but I can tell it's easily, you know, 20 degrees more than that.

1:13:50And so I think what I'm most potentially interested in is if there is either a better API, which presumably there would be, and I'm going to give you a chance here, Marco, in a second, but a better API to make this work a little better or be easier to do on the developer side, but I guess more importantly, just to have a little bit more ability to make the watch feel alive rather than something that occasionally wakes up and then goes immediately back to sleep again. I don't know, Marco, does any of that make sense? Yes, and I think you've identified a couple of the key parts of both why I'm excited for this and also why some of the issues I've had with WatchOS in the past and present.

1:14:34So Apple has a long history going back to, I believe the first major version of this that we saw in modern day was dashboard for macOS. a long history of these views that you can get where you have to like go to the view or you look at a view of something and at that point when you're looking at it the app is asked hey update this data and in the meantime maybe it shows a placeholder or stale data and then the new data pops in like a second or two later and from a user point of view that sucks it's terrible when you look at something on your watch or in the old dashboard interface on the Mac or different places like that or the old Today screen or the Today widgets on iOS.

1:15:19You would seek out this information, you would go to it, or you would look at it, and then at that point the system would ask the app, oh, hey, update this. And the reason why it was built that way was that these systems, for various reasons over time, there weren't the system resources or they were chosen not to be spent this way to have those apps constantly running in the background, being able to update their data whenever they wanted. In the case on the watch, the watch has always been extremely power constrained, and that dictates everything about it. It's a very small battery and a very small device, and they really can't make the battery any bigger.

1:15:57It's already most of the device. So it's like you've got to deal with what you've got here, and that dictates everything. That dictates the slow processor the watch has relative to the phone. Everything is about space and power and minimizing drain on that battery. speaking of that there are rumors that they're going to actually change the soc not just change the number but finally change and i don't i don't see the rumors around three nanometer are really confusing because i know tsmc is ramping up on three nanometer but the lead times are such that i can't figure out uh when apple will ship a product fab done tsmc's three nanometer process but i also can't imagine the watch soc being substantially better without going to a new process the problem i don't know what the current process is and is it actually the same process sizes that the like the good five nanometer that the phones are or is it it might still be seven i mean because the the watch soc like the the cpu in the watch i don't know about the whole soc the cpu cores in the watch have not changed in three years i think the whole soc hasn't i think you're right the die shots that like not much has changed it for a long time yeah so it's it's using a three-year-old processor and so that's that's not fun yeah i think the series six seven and eight they're all the same chip um so so the rumor is this might be a new chip and i think if it uses a lower a smaller process size than the previous one that could give apple a little bit of headroom to change some of the things you just described yes because that's the thing like with with watch os in like since its beginning it was all about like conserving extremely aggressively conserving the resources on the watch because there's not much to go around so that impacts the way apps have been built from the beginning in various ways but the gist of it is what case you said that your app you you you don't always run like if you if you're not being actively used on screen with very few exceptions you're not running and so if you have if you rely on a third-party app that has say a complication that you're looking at say you know carrot weather you know a good weather app you know you're looking at you know whatever the complication is on the screen like that app is only being woken up in the background every so often to you know fetch new data from the internet update the data in the app and you know update the with the uh the complication you know data and it tells the watch hey over these next you know six hours here's the hourly forecast to display from this time to this time show this from this time to this time show this etc it gives this timeline of updates and during that entire time the watch may not and probably won't ever ask the app hey refresh give me give me give me new data here so what you're looking at when you when you look at watch complications you are looking at a series a timeline of pre-rendered data that is that the app has provided during some like three second long window it was allowed to run in the background a few hours ago possibly and and there's a certain limit on how many times per day can do that the app servers can send push notifications to request an update but those aren't always obeyed it's more it's like it's a request not a guarantee like please give me some time i have some new stuff i i would like to show it and the and the watch can say okay or the watch can say i don't know what you're talking about and just walk away and and so there's is all these limitations and the widget system on the iphone was built the same way if if you were making watch complications you know a couple of years ago then when widget kit came out it was like oh this is clock kit just you know slightly different and with swifty y which made it a lot better but that that that idea of when you swipe over when you view a widget on the iphone that app is not running and this is why again like there's there's also some rumors with ios 17 of possible interactive widgets um and i and this kind of flies in the face of this because it's kind of questioning like well what is what would that mean and i think it would be very uh limited if that if that is a thing but anyway because again widgets on ios work that same way and then they actually they later then brought that widget system back to the watch and replaced clock kit which was based on widget kit was based on the ideas of clock kit but did it better with swift ui then they redid the new complication system using widget kit because it was better anyway so they brought it back full circle anyway so on ios you know you have widgets those apps aren't running it's the same thing where they are periodically asked for updates and the app gives the gives ios a timeline to say all right show this from this time to this time etc and again same thing applies this the that app servers can send a push notification requesting an update it may or may not be obeyed there are there's there's limits there's throttling etc on ios there's you know one set of limits on the watch it's way more aggressively throttled and there's way less opportunity for apps to update their complications on the watch so anything they do first of all it makes perfect sense to bring that system more to the watch because it does what the watch needs it gives a timeline of updates without having to have all these apps running in the background draining the very very tiny battery in the apple watch so it makes sense to have this be more built into the system.

1:21:15That being said, if there's no other major advances like more frequent background updates, a more robust system to request updates, higher limits on how much CPU time or memory those processes or extensions can use when they're updating, without other changes like that, it's going to be a fairly limited system. And it's not going to enable a lot of what people would expect or want. so i hope when they bring this over assuming they have to i mean i'm sure it's too late now whatever they've decided to do it's that's it but i hope they have when they have presumably brought this over i hope it has come with some increases in those limits or some you know less aggressive throttling some more frequent updates being permitted or something uh and and all of that i think will be a great thing once they establish it because when you look at the way the watch works now apps on the watch still suck they just suck less than they used to way less i mean they used to be horrendous uh with with you know watch kit one that was that was rough but you know they're still very limited they're still way more limited than ios apps and it makes it very hard as a developer to make a great watch app experience that will match what our users expect as a result it's kind of unfair you know like like i i don't use a third party weather widget on my as a complication on my watch because third party apps even the best written ones like carrot carrot's a great app it is even the even the best written weather apps on the apple watch cannot update their complications as often or as reliably as apple lets theirs update so i use apple weather complications on my watch even though they're less good because they they work better and more consistently and there's and that's not to say that's not saying anything against the developer of other apps they can't do anything about it it's out of their hands uh so and i know this like because i i have my own custom little complications don't do anything useful but i just kind of use them for cosmetic you know things on my own face and i know that occasionally they just stop getting asked to update and i can't do anything about it like it's So the whole system is pretty limiting for third-party developers.

1:23:32So what I'm hoping, if they're moving this to the system from a developer point of view, I hope those limits get raised in a big way. And maybe that'll need the new processor. Maybe not. I hope not because I would like them to be available on all the existing install base. But hey, we'll see what happens. and then as a user of watch os the honeycomb screen has always been a huge failure like i anytime you have to go to that screen you've lost and it's going to that screen it's like trying to plug in a micro usb cable it's like i know i'm going to be i'm going to spend way too long staring at these tiny icons trying to find the app that i want and whatever i tap on whatever i guess is this alarm or timer whatever i guess is going to be wrong the first time and maybe the second and so it that screen is terrible you can change it to a list if you want by a long pressing but the list is terrible in different ways it's first of all it's just too long there's all these different built-in features many of which you can't turn off that you know you need some way to access so that whole system could use a rethink it's a hard problem i i get you know they're trying to cram 40 apps onto a two-inch screen like yeah it's that's a that's a tough problem but it there is room for improvement in the way they've done it now so if the rumor is correct that they are actually re-implement or rethinking large parts of how users interact with apps on the watch at all that's a very good thing too and then finally i i would say the the rate of adoption of third-party apps among non-nerd people on the watch is really low it and and i mean this is anecdotal but you know look around people in your life who have apple watches we've talked about on the show before how many of them even know that you can change the complications ask around look around see how many people in your life use an apple watch face that even has complications and then see how many of those are on the defaults and it's a lot it's most of them so the watch has a lot of functionality that is just that people just don't know about and you can't blame them because the way to access this functionality is often really um hidden or or convoluted or on the phone yeah or on the phone which i mean fortunately they have been moving a lot of that off um like i recently i tried to edit like what metrics were shown in different workout modes and that used to be phone only and now it's watch only like now you have to like hit the little like dot menu on the workout and go in there and edit stuff there it took me a while to figure out like first i was looking all over the phone app for it couldn't find it anywhere there i'm like well crap did they just remove customization of workout faces and then eventually i found it in the in the watch app but it took me a while and i and i and that was like i know this feature is probably still here somewhere and i know roughly where to look at it and i'm a computer professional and it took me a long time to figure it out so there's a lot of stuff like that on the watch there's a lot of functionality on the watch much of which people would really enjoy if they knew it was there or if they knew how to do it and it's it has a massive discoverability problem and a lot of stuff is super clunky about it once you do figure it out so there's a lot of room for improvement there i'm really excited that they're that they're tackling this because in most for most of the last three to five years it has seemed like the watch is not well staffed like it has seemed like when wbc comes comes and goes every year for the watch it has seemed a lot like man is do they have like more than four people working on this entire os like because they you know they do good work but it's just really bits and pieces come out every year like it's not it not a lot comes out from the watch from watch os every year it's a very it's kind of like tv os like you know you get a couple of things here and there but it's it's a very very small set of new features every year and so it has seemed under invested in in the last you know x years this suggests that they have actually put a lot of effort into it this year so that's exciting to me as well i'm really looking forward to hopefully the watch having like major major updates in usability or in app capabilities or app structure or whatever it is i hope they actually deliver on that i hope it's really good and i would also hope still maybe the more you put a widget like architecture into the watch i think the the easier that makes it for them to offer third-party watch faces as an option yeah now whether i know i know happen eventually whether yeah whether they are doing it whether they will choose to do it whether they will permit it these are all very different questions but from a technical point of view using widget kit or a widget kit like system with some pre-built parts makes a lot of sense for enabling for the structure to enable third party watch faces because then you don't have to have the third party process running constantly on the on the watch you can again just wake it up periodically um you can you have you can define certain pieces that you have to use like all right well you have to use this digital face of this size and it has to be in this location or or you can choose to use these specific analog watch hands and they have to be in this location you know whatever it is if they wanted to dictate certain parts of it and certain requirements say all watch faces must be able to do xyz or must use pieces xyz in this way they could do that really well with with what with widget kit technically speaking and practically speaking so it seems like the the most apparent and obvious way to enable third-party watch faces based on the tech stack we know of today is with widget kit so bringing widget kit to watch os in a bigger way yeah that makes a lot of sense for maybe maybe maybe enabling this either now or in the future i hope it's now if it isn't now and i know it probably isn't i hope it's in the near future because i still want third-party watch faces and this is just such a such a clear like you know i wouldn't necessarily call anything easy but this is actually a straightforward and and dare i say obvious way to offer them you know using this kind of architecture.

1:29:38So anyway, I'm looking forward to this, and I hope this is an exciting year for watchOS. So imagine this. It's WWDC. We're watching the video. Since it's a fantasy, we're all sitting next to each other for the first time since 2019. We're watching the video together, and Kevin Lynch comes on the video and says, I only have one thing to tell you about the watch. Communication between the watch and the phone is reliable now, and then walks away. That's not going to happen. Is that enough? It'll never happen. But Marco, is that enough to make this a classic, just favorite WWDC of all time? Oh, no.

1:30:15Because first of all, I don't think it's possible. Second of all, I wouldn't believe them. I think the way to fix phone to watch communication is to get rid of it. The way to fix it is to give me the same controls that I have on URL background download sessions and URL communication. you have most of it already give me the same background download abilities and APIs on the watch that I have on the phone let it download, let background download start immediately and let them run over Wi-Fi and not over Bluetooth to the phone there's a number of points in that communication stack that are either unreliable or ungodly slow and again, this is all in the name of saving power you can see why they did it but I think a lot of those decisions they might have made sense seven years ago maybe they make less sense now uh or maybe they're not as necessary now as they used to be and those decisions really hold back apps and and so again i hope they like just let my app initiate a backgroundable download that will run over wi-fi and start immediately that's it i can do it on ios i can do it on all other apple os's well I don't think macOS supports background downloads anyway I can do it on most other Apple OSes I can't do it on the watch and that really holds stuff back to the point where I had to make a breakout game in my app so that you would keep the screen on like the whole reason I had to make a game in my watch app in my podcast app while you download a file is so that you have something to do that will keep the screen on so I don't have to rely on the background downloads when you want to download something right now before you go out for a run or something and you don't want it to be canceled for no reason or to never start for no reason so please apple make make that stuff better but anyway um i will complain a lot less about that if i have custom third-party watch faces no that's fair so john pulling on this thread a moment what is your i don't care about anything else that happened this was a great wwdc i assume new mac pro uh i mean yeah well i'm not really expecting that because of the rumors although Well, I do have a brief aside about something that Marco talked about, how people don't know, don't customize their complications on their watch because they don't know they can do it.

1:32:36And the interface of doing it is clunky. And it's lots of things on the watch are non-discoverable. And when you were describing that, I thought this is another place where it seems like Apple is, I mean, we don't know because we don't know what they're doing, but it seems like they're behind where everyone else is zooming forward with their large language models. One of the things that current large language models that are out there right now are really good at doing is exactly this task, which is a voice interface to describing what you want your the complications in your phone screen to be like.

1:33:07You can go to one of these large language models right now and say, make me a web page in the lower left corner, put a picture of a sun in the lower right corner, put a counter that counts up every second. And like you could actually describe the functionality of the widgets and it will do it. Like how easy would it be for a well-trained large language model to say, um, I want the weather in the upper left, the date in the lower left, uh, and then to continue a conversation and said, no, I want the date to be, uh, the day of the week plus the month in the year, or I don't want the year. Like those types of conversations, you can have them with large language models right now.

1:33:38And the canvas that they're painting on is like anything you can ask for more or less. It'll try to do it with varying degrees. the problem space of picking which widgets you want to picking out from a fixed set of widgets to appear in a fixed set of location with a fixed set of options large language models would eat that for breakfast the problems are one apple seems not to be on that train with everyone else as far as we know because they haven't announced anything in that area two their existing thing that you can talk to siri sucks balls and we know that um and then you know three i don't think this is the type of thing that apple can uh you know say oh well they're popular we're rolling out this year you know like it's again it's hard to know because we don't know what's going on but like it seems like apple is behind everyone else on this so you can't expect them to surprise watch os 10 is out and it's got a large language model that you can talk to oh and i guess this ties into what marco was saying to us even if you could like you don't run the large language model on the watch right so now it has to do network communication oh no the sky is falling batteries being destroyed and i get that like like it's true of what marco was asking for i want my download to start now and on wi-fi it'll kill people's watch batteries but it's like but they want that to happen when they're down they you know it's the balance between which would you rather have your podcast literally never downloads and it frustrates you to no end or we burn part of your battery downloading your podcast like you have to pick one you can't like there's no way to get the podcast on your watch without burning battery so either you want it and you're willing to sacrifice the battery for it and you don't and the same deal i think with talking to your watch and setting up with a large language model and setting up your complications.

1:35:09Maybe as a one-time tutorial that you guide people through when they're setting up their watch and they have a conversation with it, and during the course of the conversation, the watch face forms in front of their eyes, is a great way to A, show people that that's possible, and B, let them know that they can do it in a way that doesn't involve them poking their fingers at the screen trying to find the interface. Because I have the same frustration. I'm not familiar with watchOS because I don't wear an Apple Watch, and anytime I want to do anything in watchOS, I don't know where it is, I don't know how it works, I can't figure out what things I'm supposed to be touching or what even what the vocabulary of gestures is to do stuff.

1:35:40Large language models would destroy this test. They're so good at it. It's such a narrow problem space and you can talk in a freeform manner. And Apple's speech to text actually works pretty well. So once you get over that hurdle and you can throw that text to a large language model running somewhere on an Apple server that it can communicate with, that is a i think a really good use of current technology for you know talking to things and the problem space of dealing with any kind of ui on a two-inch screen that's on your wrist yeah and on on the point of large language models by the way um this is obviously you know a huge question is like we're in the the this like you know boom time of this new generation of ai capabilities what's apple going to do at wbc with them and and i think unfortunately the answer is probably nothing there is nothing uh because like that's not what apple does like apple does stuff you know more slowly than that in a lot of areas and i mean they could if they were on the ball on this and were were sort of ahead of the curve or even you know even with everyone else they could have stuff to announce the wbc but that whole part of the org seems dysfunctional now again because we don't know what's going on for all we know they've been working on large line go on stuff for seven years and they're about to roll it out and wow us but it certainly seems like that's not the case no i i think what's more likely is look apple apple knows that again these are smart people they know that the eyes of the world are going to be upon them and expecting ai announcements and so i think what they're going to most likely do is brand some things as ai that are not actually using this new generation of of technique and it's just using you know the the previous generation you know like you know what we used to call machine learning um you know these terms and i know these terms are you know oftentimes vague and shifting around over time but they i think they might just you know show off stuff that is using that kind of technique and say this is ai powered you know just because those are the buzzwords that people expect the day the rumors surrounding the like siri and the headset i seem to recall reading something somewhere that was like supposedly inside dirt on like the the headset team didn't really want to use siri but like that's what's available to them they're kind of like wanted to go their own direction because siri sucks so bad they just like well we you know i don't know if they want to use large language models maybe that predates that but like they know they have a problem here because they the headset is similar like the interface to it is more limited than it is on a phone ipad or mac so you have to figure out how do i manipulate things i don't know the whole camera's looking at your fingertips and so on and so forth but it's a harder problem right it's harder than just using a pointer or cursor you know or touching the screen with your fingers you can't do that with a headset so let's use the voice assistant that our company already has oh but that voice assistant is bad so you don't want the headset to be bad people work on the headset is that we want the headset to be good and we do need something that's going to listen to us and so i kind of agree that they're like as they promote siri and the headset maybe they'll have a new suffix siri ai sir i mean i don't know how they'll go that far but like maybe they'll just say hey siri's in your headset too and everybody loves siri and look how amazing it is you can talk to it and make little vr apps as the rumor was or whatever but like yeah this is another one of the benefits of the headset that the fact that it blots out the sun i think it can blot out the fact that apple doesn't really have anything to show uh for it uh regarding large language models wwc which is fine they got a whole headset and a new platform isn't that exciting i think they can get away with it this year next year when we come to wwc next year regardless of how the headset turns out maybe next year it's finally just shipping to people or something next year apple needs to start having some kind of answers with large language models because i think they've proven they're useful in enough context and like the reason i gave the you know setting up the complications on your watch uh you know watch screen uh problem space that is a limited domain where large language models can excel in in ways that siri absolutely has not because you have to know how to talk to siri in the special way that Syria understands.

1:39:41And even when you know, sometimes it does ridiculous things that don't make any sense. Whereas large design models don't work that way. Again, if you've never tried one, go to one right now and ask it to do something with a limited problem domain or even with a not particularly limited one, like I said, asking it to make a webpage for you or something. It's really good about having a conversation with you and understanding what it is you want as long as you're talking to it about something that was able to train itself on that's out there in the world. and Apple can absolutely find a large language model and train it on enough data to understand watchOS complications.

1:40:16Yeah. One thing, by the way, in terms of WWDC hopes and dreams here, one thing I would love to see, I don't know if it'll be this year, it'll more likely be next year, but I would love to see Apple make available to developers through APIs on the phone some large language model capabilities that they have optimized like crazy to run on their hardware. So for instance, the stable diffusion algorithm for image generation, Apple actually, I believe it was that one, Apple actually contributed to the open source project or whatever to optimize it for running on Apple Silicon Max or Apple Silicon in general.

1:40:57And then that kind of enabled people to make iPhone apps with it. I would love them to do more of that, that kind of thing, and build it into the OS so that we don't have to download each app. A good example of this right now is if you want to transcribe sound into speech, into text. Apple has actually shipped an API to do this for a number of years now. It's called something like SF Speech Synthesizer or Speech Recognizer, whatever. There were a number of problems with it. Number one, because I've been trying to do features like this in Overcast for years. One of the biggest, clearest examples of where I could use this is like when you do clip sharing in Overcast to be able to show as part of the video clip the text.

1:41:41You know, many apps like TikTok, you know, they do this, they have this built in now. This is a common thing and I need to do it at some point. And in the past, this was limited by a number of factors. Number one was in the past, I don't know if this is still the case, that API, to use Apple's built-in speech to text API, your app had to request and get access to the microphone. even if you were feeding it audio that was not from the microphone. So I could not, as a podcast app, feed the pre-recorded audio that I downloaded to my podcast app through their speech-to-speech system until I prompted the user for microphone access.

1:42:20And that alone was enough for me to say, I'm not using this feature because I'm not going to ask my users who are all educated nerds for permission to access their microphone in a podcast app because there's no reason I would need that except to be creepy and spy on them with ads and stuff. So I don't even want the appearance that I might be doing that. I don't even want the ability to access their microphone. I don't even want that to be possible in my app. So that alone ruled it out. But also, it was using older style models and things to do the speech-to-text conversion. And it wasn't very good in my testing.

1:42:58I made a couple of test prototypes, granting myself microphone access just to see like you know how good is this and it was bad and it was it was too bad to ship so i didn't ship it well whatever however many months ago open ai released their whisper model and whisper is really massively better it's not a small difference it is way more accurate and there there is a project called whisper.cpp where some guy has has done amazing work in like transforming this this thing that you would otherwise need a whole bunch of python to run into a very simple c++ wrapper that you could make ios and other apps with and it runs fine on apple stuff and i tried that and i built a prototype with that and it was super easy to use and uh and and the downside of using whisper there's two main downs number one if you use the large models that have like larger vocabularies and more accuracy uh it's just too slow to run on an iphone um number two these models are huge to download like the the base model that provides like okay accuracy decent speed is something i think it's something like 70 megs and then if you want something more accurate it goes into like multiple gigs very quickly um and this is just it's untenable to run that in the way i would want to run that would be you know stuff like transcription of what you're listening to uh you know there's lots of things i could do with that if it was fast and and you know inexpensive intensive computationally to run that but if i had to download like a two gig model you know as part of the app to do those features or whatever that's that that's not going to fly and if it's going to be super slow it's not going to fly well they could build in some medium-sized model into the os if they wanted to and they could use it themselves for their own speech attack stuff and then make that available to developers they could also for instance optimize that model you know the reason why the whisper.cpp doesn't run very quickly on iphones one of the reasons is because apple has not done their work to optimize that one the way they did with the stable diffusion also they uh the the way these models usually work you if you're running in the background you don't have access to the gpu or the neural engine as far as i know but you definitely can't run you can't run ml models on the gpu if your app is in the background you can only it it makes your app only use cpu based computation and that makes them way slower to run in many cases And this is the case for whisper and whisper at CPP only ever uses the CPU.

1:45:25It doesn't have any access to the hardware acceleration besides, you know, CPU raw stuff. And so this makes it slower than it, than it could be. So if Apple actually wanted to offer a really great dictation to speech API or to speak to text API, they could massively improve the, the API they've already offered for years by using this new kind of technique, a new kind of model like this, optimize it, let it run on the acceleration hardware, ship a base model with the OS, and all of a sudden, every app could have really easy access to speech-to-text. That's the kind of thing that I think is more likely for Apple to offer with these new AI techniques sooner than having them offer some kind of full-blown customer-facing, quote, AI-powered feature.

1:46:14I think it's too soon for that, and that's kind of less Apple-y to do that kind of thing. the customizing watch face is a great example of how apple would do it it's not like you can have arbitrary conversations with a thing like they would be very narrowly constrained and i know the large language models out there now are supposedly constrained but they're it's more like everything except for these five or six things they try to keep them away from whereas this would be you can discuss nothing except for watch faces and i feel like that is easier especially if they train the model themselves right um which again costs money uh and you know regarding the the speech text thing and the size of them and everything there's a reason large language models are called large language models they use tremendous amounts of storage and tremendous amounts of memory and tremendous amounts of compute not just for the training which is itself huge is obviously you're grinding through tons and tons of data but even just for the part of it where it runs the the inference part of it where it does its thing even that takes a substantial amount of uh of power and resources which is why i was saying the watch would have to be communicating with the server obviously and then even just communicating with the server would be taxing for the watch uh and drain the battery and so on but not to mention the servers yeah well i mean the servers yeah but but again to be that incredibly constrained to just be doing watch faces it's what we talked about a while ago and the people have actually implemented a large language model sitting in front of essentially the functionality exposed by siri or like you know limit functionality that's already on the watch the watch can already pick complications and configure them and put them in places on the faces it's just a question of translating from speech to text and from text to what is this text telling me to do and then oh actually i i the watch have apis for doing all the things you're asking for it's a very constrained problem space which is miles from like what microsoft did was like here's bing you can talk to it about stuff good luck yeah but like ultimately i what i want to see from apple in the AI age really is make a lot of this cool tech available on device for for on device use by apps that are not communicating with their own servers that can just run everything locally and Apple has you know with the exception of of having what seems like organizational or cultural trouble with AI machine learning techniques and people in general that's that's a large issue they need to figure out if they haven't yet.

1:48:34And so far it seems like they haven't. But what they have resource-wise is they have all these supercomputers in everybody's pockets that have amazing hardware and software capabilities. And they care a lot about privacy and things running on device whenever possible and everything. And that's a level of sophistication and an amount of resources that their competitors don't have. So I I hope Apple leans into that. And I think this is how they're going to ultimately tackle, you know, quote AI, leaning into on-device stuff as much as possible. And if they do that themselves, they will most likely do it in a way where there is an API for developers to hook into or use those same capabilities in our apps.

1:49:18That's what I'm hoping to see from Apple in this new age of these new styles of AI. is cool on-device capabilities that run quickly and locally that don't need us to, as a developer of one app, we don't need to go train our own models. We don't need to go run a huge server farm or have our own open AI tokens that we have to use and figure out how the heck to pay for. I want to see what Apple can deliver us on device that they will maintain, they will train, They will update it over time. They will make it faster over time and let us build on top of it. And they have a very strong history of doing that in other areas.

1:50:00So I think this is what they will ultimately do whenever they get their act together on AI. And we don't know when that will be, and it probably won't be all at once. I'm hoping to see little hints of it this year. Even that, I think, is optimistic, but we'll see. And then maybe next year or the year after, maybe we'll see big stuff. And that'll be exciting. they're already doing all that it's just a question of what letters they put on it they used to put the letters m and l and now and now we're saying the letters a and i what we mean is exactly what apple has been doing which is they they do provide you a bunch of models that you can use like that text-to-speech one that you were saying you know and they do improve it over years they do make it run fast in the harbor they've done all of that we're like yeah but but no but the stuff that you have now there's better stuff apple that uses not the m and the l but it uses the a and the i and it's all this you know right we obviously there is there are different techniques to doing this and the current technique that lots of other companies are experimenting with is better what apple is doing so what we want is apple do what you've been doing with all the things you used to call ml keep doing that we love frameworks we love you making it fast on your hardware it's just that there's a new thing that does everything you've been trying to do but better not everything but certain things it does better so please give us access to that in a way that makes us not have to worry about it and i do think that they will do that and i just got done saying the large-gigging models are too big you know they're just way too big to even run at all and even like phones or even big mac sometimes that's why you have to communicate over the network but that's today like those things change like part of what makes that tractable is just the march of progress but part of it is also companies like apple deciding that it is important for their socs to be good at doing this and that can really knock down these things it's the reason the socs we have today are so good at doing the things they do so good at image processing so good at you know doing the kinds of computations that has to be done on a phone on a mac or whatever those special process processes for video encoding and decoding and you know the neural engine all that stuff was developed because apple said these these types of functions are useful to do on a phone or on a mac or whatever and we want to do them in low power i think that will come for large language models as well i don't think we're going to you know models that currently run that require hundreds of gigs of ram to do inference are going to run on your phone anytime soon but i think there will be a meeting in the middle of like you were saying marco cut down versions of these things limited problem space and and to be clear with the limited problem space you're like oh there's no way they can ever limit a large language model enough it'll always be scary i'm thinking of the scenario when you were talking about configuring the complications it never answers back the only thing it does in response to you talking to it is show you what it thinks you're asking for as you're screening your complications and basically says you know is this okay you know press okay or say okay or like it never talks back to you it just does what you and there's no way for you to accidentally have your watch insult you or something or there's very few ways perhaps watch insult you by arranging complications i guess each of each complication was a letter it could spell out a naughty word or something and some would figure out how to do it but like it's really limited problem space we just but basically what we're just asking for is you know the promise of siri well large language models are currently delivering on a lot of that promise i can have a conversation in text with a large language model and get it to do useful things that reflect what i'm asking you no it doesn't know anything about facts no it has no understanding of anything no i can't ask it questions and trust its answers but i can ask it to arrange a bunch of complications on a watch screen and the good thing about that is i'm the one who gets to say you did it correctly or you didn't and if you didn't i could say no the upper left should be carrot weather and not apple's weather widget uh and maybe and maybe it still fails to understand me or whatever but the result is okay well fine i'll do that one manually or whatever but the point is like trying to do that through siri now is laughable like you can't you know people people would never try to do that they're they're lucky when they can talk to their watch to reply to a text or something and that's only like nerds who have practiced that whole thing whereas large language models seem so much more flexible about understanding the foibles of our language and and eventually doing what we want.

1:53:58And so I really do hope that, again, not this year probably, but that eventually becomes one of the things they start rolling out, if not as a way to interface with tiny screens or headsets, then at least as something that we can expect our powerful Macs to be able to do. Because I would love to be able to say things to my Mac to do drudgery without having to open up a web browser, go to one of these large language models, type a bunch of stuff in and all that. Thanks to our sponsors this week. collide squarespace and trade coffee thanks to our members who support us directly you can join us at atp.fm slash join you get that awesome new pizza special and everything we've ever done before that plus all the membership perks you'll see atp.fm slash join thank you everybody we will talk to you next week now the show is over they didn't even mean to begin because it was accidental Oh, it was accidental John didn't do any research Marco and Casey wouldn't let him Cause it was accidental Oh, it was accidental And you can find the show notes at atp.fm And if you're into Twitter You can follow them At C-A-S-E-Y-L-I-S-S That's Casey Liss, M-A-R-C-O-A-R-M-A-N-T-M-A-R-C-O-R-M-N-S-I-R-A-C-U-S-A-C-R-A-Q-U-S-A.

1:55:32It's accidental. Accidental. They didn't mean to. Accidental. Accidental. Tech podcast so long. All right. So I wanted to talk to the two of you. I wanted to kind of invoke the brain trust, perhaps call the diamond dogs together and see what you two thought about pricing for CallSheet. We talked about this. I don't know when it was. I think it was a little bit last week. I think it was a little bit a few weeks before that. But I'm getting, knock on wood, getting really close to sending this to the App Store, which means I really need to lock down my pricing schedule, for lack of a better word.

1:56:12Well, you need to lock down your initial pricing. Sure. You can always change it. That is true. And so I think what I've landed on as a general scheme, and then we'll talk about the different options therein. I think what I'm going to do is you get to use the main screen, which shows you like popular stuff in your country or really your language, popular stuff in your language, new stuff in your language. So that would generally be new stuff in English for me and popular things in English. and sort of kind of American-based when it's English, but not always. It's complicated. It doesn't matter.

1:56:50But anyways, you can do all that. That you can do. You can tap through that whenever you want, and there's no real limit there. So hypothetically, if you can six degrees of Kevin Bacon yourself to the information you want, that'll take a while, but you can do that for free always. but to search, I think I'm going to have a daily limit of something to the order of five, maybe 10 searches at most. That's so many. What do you do? That's too many. Is it? Okay. Have one or two. Okay. Well, hold on. Let's shelve that for a second. Put it in the parking lot. We'll come back to that. Because no, that is something I want to discuss.

1:57:28I didn't realize that that was too many. So I want to discuss that, but hold on a second. So some number of searches per day. And then after that, it's tough noogies. Subscribe or get out of here, wait until tomorrow. And I feel like in general, that is the right approach. And that's what I think I want to go with. But if you disagree, either of you, we can talk about that in a second. Then the second question becomes, assuming we have this some amount of free searches per day, and then you have to subscribe. All right, well, what does subscribing mean? And I think when we first talked about this, we had concluded kind of as a threesome that somewhere to the order of$8 a year seems like a reasonable price for this sort of thing.

1:58:09And I think I feel pretty good about that. But there's two other options on the table that I think would be additive, not a replacement. One of them is, do I offer a monthly option? And if so, how expensive is that? And currently I'm leaning toward offering a monthly option for either$1 or$1.50. I'm not really sure what feels right there. Now, that would be kind of considerably more than the yearly option. So I want to know your thoughts when we get there, when I stop talking in just a second. But I think for sure, I want a yearly option. The question is, do I do a monthly option? And if so, how much?

1:58:45And then the other question is, do I do a... something like a lifetime unlock. I think using the word lifetime is a dire mistake, and I wouldn't phrase it that way, but maybe a one-time unlock if I go this route at all. And I'm currently, I think I'm against it because I just think that it sets me up for failure. And even if I'm leaving a little money on the table, it makes me nervous promising anything that far out. That just makes me feel a little uncomfortable. So that's kind of where things sit right now. And I am curious what your thoughts are. Now, Marco, you jumped all over me justifiably.

1:59:25You know, I'm not upset. You jumped all over me a second ago, well, both of you really, about five free searches per day. So talk me down from five. What should the number be, Marco? I'd say one or two. I mean, that's tricky, though. It is. Look, again, it's tricky. And there's a question of, like, what defines a search? That's a big question. If you search and you mistype it and get no results, does that count as a search? Yeah. So the way it's currently implemented, It's not in the test plate yet, but the way it's currently implemented in a branch that I have going right now is that you can, strictly speaking, type however many searches you want during the day.

2:00:00But the moment you click, the moment you tap on a result, then that counts as a search. And now you've lost one of your searches. Yes, that sounds good. I would also, you know, the branding of this, I would say, you know, you've like, you call it something like free searches. So you have one free search left today, and then when you hit the end, you've used all your free searches today. You can buy the subscription and get unlimited searches. That's how I would phrase this. You have free searches, and then you have this paywall that gets you more searches. Speaking of free, you're doing a free trial, right?

2:00:34Well, that was my initial thought, but now that I have this whole free search thing, I feel like it's kind of redundant. Yeah, I don't know that you necessarily need a free... If you have this... Here's what you need. You need to have... What you need to have, the goal, is someone needs to use the app enough to like it. And that is the tricky part. So using enough to like it, like just a single use instance, like I was watching one TV show and we had a question on the couch about what was this person in. That one instance of use that has, I feel like that one instance of use has to be successful and enjoyable without any pay crap in their face.

2:01:11And you're like, okay, but how many searches is that? and that's why i think that the picking the searches is tricky 10 is too many right because no one's going to burn through 10 a day 10 is enough for a single couch instance and maybe you'd use the app once every few days um but one is too few because for a single couch instance like no one will ever get the point to like the app because they'll be like i used it for two seconds and then it threw a paywall on my face and i hate it that's why the free trial is good because the free trial period is unlimited and that allows the free trial period to be this is the time when you get to like the app because it's unlimited it's a free trial everything's open full access to everything but that door closes and that allows it takes the pressure off your number of you know searches per day to allow people to get to like the app they're supposed to get to like it during the free trial and then the paywall sole purpose is to gently remind people that they can pay to get that experience that they previously had if you don't do a free trial you have to somehow thread the needle of like how much will let people use it enough so they'll like it but like because because i think some people will never go above one search per couch incident or whatever you know like they'll just they'll just use the app forever for free to say i you know i don't do this often when i do do it i look it up and again what you just described i can do as many searches as i want until i find the result that i want and then i tap on it and i get my answer i put the phone down never look at it again they might not even know you have a paywall with your limit of one and other people in a single couch incident to go through 75 searches because they're like what was this person what was that a search for this person i think the person's name was this and they're tapping through results and they're doing all sorts of things and they will be incredibly frustrated by a limit that's like one or two but if you have the free trial you could be like look you had the you had your salad days you got to use the app you got to see all the features you got to you know and then it's just a question of how long is the free trial it would It would kind of be nice if the free trial only started like when they, you know, when they had their first big usage scenario.

2:03:06But yeah, I think a free trial would hypothetically start upon subscribing. So you subscribe, you get your one week or whatever, and then then you're charged. I don't know. I hear what you're saying. I don't disagree with it. But I feel like, especially if I were to crank the limit from five to like two or three, then maybe I'm wrong. But I feel like at that point you have gotten at least a vague enough notion to know, okay, no, no, no, this is all right. And if you want to try again tomorrow, try it again tomorrow. I don't love the idea of both a free trial and the search limit, but that is a weak opinion held loosely.

2:03:43See, this is getting very complicated. like i i almost think like john's right like you know you you want to give people a chance to like the app of course so that's making me now revisit like what if instead of just instead of having anything per unit of time per day per month what if it was a higher limit that was a single view so for instance you get 10 free searches after that you got to pay that's it like because then you could make the number higher and you can give people more chance to get to know the app you could display it more easily in the ui it's it's much clearer to people what the limit is some people take a while to come around to pay for stuff though like some people need to be convinced that there is actual utility they may like the app and they may say oh i found that app useful but they haven't come around to the idea that i like it enough to pay for it and that's where the doling out a small amount periodically is better than the big buck even if it's a big bucket of 100 once they burn through that 100 and they like the app they may not be ready to pay still but if they'll but if there's one doled out each day after that that they'd be like you know what i do keep coming back to this is the same reason so many people including me eventually paid for the new york times because it's like they have a bunch of free ones and i don't think i want to pay for the whole thing but you know what over the course of months weeks years you realize i'm more sick of of uh running out of my free articles of the new yorker or whatever than i am sick of paying for the new yorker and then eventually you do it and obviously the numbers here are way smaller than the new york or the new york times subscription or anything like that but it does take people's time to come around if you slam that door closed you use your 150 free searches and that's it forever and ever and ever they're you're giving out the opportunities for them to eventually come around because they're just going to delete the app yeah i guess yeah because it's just it's tricky because like you you don't you don't want to make this too complicated you want to keep it so that people know that there is a limit they know when they're going to hit it and then when they hit it, they don't say screw you and delete the app.

2:05:40So I get what you're saying, John. I mean, I just think once we get into you have X per day, they recharge per day, and then you can start a free trial that lasts X amount of time, then you're paying me eventually after that. And by the way, you have an annual or a monthly. That's a lot of complexity for this pricing model for a very simple app. I mean, a lot of the apps I see on the App Store are exactly like that. Like I know I've been subjected to prompts and screens describing this exact scenario. And as a customer, I tend not to get hung up with understanding what it is that saying the customer.

2:06:15All I know is, can I use the app when I want to? And if I can't, how much money would I have to pay to make it happen? And then the final thing is, will I be able to use the app in the future? Should I just delete it now? I don't actually understand or know or try to parse the text about what their scheme is and what the limits are and what the whatever. All I know is I can't do what I want now. Why is that? And how can I make it go away? And is this a permanent situation? And I feel like that's how most people interact with this. Like they don't have to actually understand your business model.

2:06:43Well, but that's even more reason to me to say, look, it's either you pay or you don't. And this is what Marco was saying a moment ago. Like if you don't want to pay, I'll give you a little sip. You can just make it a paid up front app like an old person. Well, yeah. I mean, but if you don't pay, I'll give you a little sip of it each day. And if you pay, then that's that. and then we're good here. As long as you keep paying, you get as much as you want. And like Marco was saying, I don't disagree with you in principle, but I couldn't put my finger on what felt gross about it. And I think it's what Marco said, is that just feels too complicated.

2:07:14Like either you're paying me or you're not. There's no in the middle. You know, there's no free trial. It's either you're paying me or you're not. You really want to use this app? Free trials are insanely popular for a reason though. That's true. And you can't overlook that. They're everywhere. They're everywhere. I think they do really work. and I know it does complicate things. It does, especially for an app. Your app is different in that it's not, like it is very situational. You're probably not going to use this app while you're commuting on the subway, right? It's when you're on the couch, it's when you're watching the show or when you're discussing a show.

2:07:47It's not a general purpose app. So you don't have that many opportunities for engagement. And when you do have the engagement, the value of your app is it's there when people need it. And if it's like, oh, they go to look it up and your app gives them the Heisman, they're not going to come back. Like I said, it needs to prove value. It needs to say, I wanted an answer, I got an answer, and the app was pleasant to use. And this is a thing I find myself doing frequently. So we'll see. Now, here's the other thing that I'm sure people listening are probably thinking, and we've discussed as well. The audience for your app, at least initially, is not the same as the general public.

2:08:22It's going to be able to listen to your podcast and know about the app. No one knows about this app except for people who read your blog and listen to your podcast or are related to them. because it's not out yet uh but that is a ready-made customer base and that customer base has totally different values has a totally different idea of the of the exchange going on here they might be buying it because they're a fan of yours because they've heard us talk about it so much the general public is coming from a totally different perspective so you know i feel like and i think it's reasonable to like this app has nothing to do with atp it is a general purpose app that anybody who watches television and movies can use the question is do you want to design this app to be optimal for the people who listen to atp or to be optimal for the people who have no idea what atp is or who you are and those are two very different designs unfortunately what makes you say that like what what's different because the people who are listening to atp they have different desires out of the app their their impetus to buy their their motivation to buy is so incredibly different for people who have no idea who you are you know what i mean like it's just they're totally different but they're very they're limited they're small in number and so like you could just say i'm just going to set them aside and i'm going to build this app for the general public uh because you know practically speaking the people who just want to be the app buy the app because they're curious about the app that casey made or because they want to have more context for the next episode of atp they listen to that's not a market that's going to sustain the development is that you know what i mean everyone else is and the everyone else you know the design for them often involves the gross stuff that you find unappealing and complicated and i base that on all the apps that i download in the app store that have those types of models that they're you know complicated not appealing and gross feeling yeah i mean i i i'm not sure i see a free trial as this make or break thing that you seem to because and maybe it's because i'm too myopic in the way i operate but like i think a free trial is useful if it's the sort of thing where that's the only way i can try the app but this kind of homegrown alternate approach, I think gives me kind of what a free trial does.

2:10:26Like I view them as equivalent, even though they're not literally equivalent. I view them as equivalent, spiritually equivalent, where I either need a free trial to prove its worth, or I need to distribute little sips of the app to prove its worth. And it seems redundant to have both to me. But I don't know, maybe I've got this all right. I mean, if you give them the sip, like I think the main thing when you give them the sip is to make sure it's clear to them this is not the end that you should not delete this app because even though you i'm not letting you do this search now at some point in the future that i'm going to tell you about tomorrow next week in an hour whatever you will be able to use the app again and by the way if you don't want to see this nag screen here's a subscription or whatever because that's what you're trying to stop is i've hit a paywall delete this app that is you know because you never get that person back they're not going to redownload it or whatever they have no you know i mean so that's that's what you're trying to avoid and free trial let's the idea of the free trial is they get to try it with full functionality and hopefully fall in love with it.

2:11:24And then when that door closes, eventually they, you know, they say, Oh, I really did like this app, or I did find myself using it a lot. I mean, I, again, I hear you, but I feel like that's the whole point of this. I'm going to say five searches a day. Maybe that number's wrong, but I feel like that's the whole point. And so I don't know. I, it, it feels currently like it's, like I said, like it's redundant and I don't love it. And, and, you know, for what it's worth, when you tap on, when you tap on like a search, I'm going to send you to, I'm just going to put this in Slack. I don't think we're going to put an image in the show notes.

2:11:59But when you, when you tap on the like search area, what, what you're seeing that I'm putting in Slack right now, that's not in the context of search. It's just, you know, the, the, the one view in and of itself, but this would be where the search results and stuff would be. It's like a kind of mini paywall that says, hey, you've got to keep going or wait another few hours in order to get whatever more searches. And that would show. And then what I haven't mentioned to you guys is the way I currently have it is if you think about the area on the, or if you think about the main screen and how there's a little magnifying glass that slides in from off screen on the right as you're drilling into things.

2:12:44When you're in free mode, to the left of that magnifying glass, I'm trying to generate an image real quick, so give me a second, I'm stalling for time. But to the left of that magnifying glass, there's a really obnoxious red banner at the bottom that says you have one search left today. I think Marco made a good point earlier. It should say one free search left today. But anyways, it's written right now. You have one search left today. Subscribe now to remove limits or whatever. I can wordsmith that however I want. But what I'm driving at is there's this big-ass banner right at the bottom saying, hey, this is your situation.

2:13:18You've got, you know, whatever, however many searches left. And then you're going to have to subscribe in order to, you know, get more. Yeah, I think you could – I mean this all makes sense. And I think the UI, you are conveying all the information you need. You might – you could actually do a fun kind of – I don't know if this is an underscore-esque thing, but it makes me think of him. kind of like where you know you have uh you know whatever you have uh one free search remaining and then you use it and then someone goes to try to do another search and you're going to show up the paywall on the paywall screen have a button that says can i just have one more search and you press it and then your app says okay and it lets you do it i don't know again this is a lot of complexity just the one just a one-time extra that again i i come down with marco i i don't want it this complicated that's not complicated that's fun that's surprise and delight what i'm basically saying is the thing that i do don't tell anybody but the thing i do with the the atp sales that we have when we sell merchandise um i always tell the the uh the listeners that the end date is one day before it actually is well because then that helps you know resolve things like what time zone does it end in like that's that's the real reason exactly i don't avoid all those issues the reason people don't say oh i thought the sale was supposed to end today but it's not i don't have to deal with time zones the end date is always one day later i and that the having the limit be one you know having one extra in reserve that you give people one time it's not like you can always ask for one more forever and ever but just having the one time extra that's another enticement to eventually buy because all this app was nice to me i was thinking about buy can i just do this one last search oh yeah you can then the paywall is really down and no you don't get to ask for one more forever and ever like a like a two-year-old right that's the type of thing that you can do to soften this to make it more to make it less likely they delete the app immediately all right here's i've been thinking about it i played with the app i saw your paywall screen in slack i think this is where i'm landing you get five or ten free searches single time use and then one per day after that well that's interesting then when they hit that one you show your your screen i will workshop the hell out of the screen for you don't ship it the way i know you i know as soon as i said unlimited searches every day what is the everyday phrase there for anyway yeah it works out the screen but it's yeah this is by no means the final screen this is very yeah you get five or ten five or ten searches up front for free then one a day free searches then you hit the screen subscribe to continue after that one a day and i like how you said you know or wait x hours to get you know your next free search like that's great and then you have your two options you're kind of killing me even offering two options monthly and yearly we'll get there you offer your subscription option parenthesis S and that's it and there's no free trial you get your one per day after you've used your five single use ones that's it, you hit this and you go forward if you want that gives people enough time to get to know the app with their five or ten stock free ones that aren't timed and then it gives them that little out to keep them in the door um on the way if they want to keep going after keep going after that um but it but it you know i i think any more than one per day and you will have the the limit that like you know which i was saying earlier with like how many incidents per couch event are we are we going to even be using this every day so so i i think that's probably your best balance i think that if you went with that model i think the initial bucket should be more generous just because that's the falling i think a little bit more than 10 would be because you really you want to give people more time to fall in love with the app and some people hit it after 10 some people it will take them three months to go through 10 some people go through 10 and one counters in it so i you know and it's because it is finite and one time it's not like you have to worry about it recurring no one's going to live forever off of your free bucket of 20 instead of 10 it's just giving people more time to use the app enough that they are convinced that it's the type of thing they want to pay for or at least not delete so for the sake of discussion i'm not sure this is the right answer but you get 20 searches for free then once that bucket is up be that in an hour or in a year after that you get one freebie a day and if you if you use that one freebie tough nuts either wait another day or pay me yeah i actually i kind of like that and i still think you should have a free i still think you a free trial but whatever i i think i'm team marco on that one i i disagree john also thinks we need to offer our t-shirts in 17 different options every time and i'm the one person like can we just do like two or three people love it people love it and it's even it even works with cotton bureau's new model now we're not even penalized for doing it all right all right well i'll think about the free trial although current my current thinking is marco's right all right so i i actually really i didn't think about this height well it's not really hybrid but i'm gonna think this makes a lot of sense.

2:18:13And I think it's reasonably easy to digest. That's the thing I'm worried about. And that's what I keep coming back to with the free trial is I think it's adding a layer of complexity that just isn't helpful. And so I like this because it seems pretty straightforward. You got your bucket. Once you use it, you get one a day after that tough nuts. All right. So I think it's just as complicated as all the schemes you were describing as complicated. But again, I say that the people who are using this app do not need a mental model of this system. They just need to know how do they feel about what the app is putting in front of their face and doesn't make them want to pay money like that's the level people are operating at when they're poking at their phone they are not trying to suss out your monetization model and and how you've structured things so i don't think that's a concern at all but if that was a concern the thing you just described is just as complicated as everything else we've described i mean maybe well and also like some people will suss it out some people will complain some people will be you know will be you know bounced off and they won't go through with it because they'll think it's too ridiculous those are only going to be the listeners of our podcast though the regular people just do not know who they would complain to assume casey doesn't exist and is just a giant application mill somewhere like they think apple made all the apps like that's not the people who complain are so much more like i was gonna say sophisticated but so much more in tune with how the world actually works than the the most of the people who are just poking at their phone screen yeah fair enough but all this is to say like you know you're not going to please everyone no matter what you do don't sell yourself short like don't rip yourself off trying to please everyone like if some if you hear from people who are like well 20 free searches is really not enough for my needs like you know fine like two then then they should they should pay you if they say that's what i'm saying if they say i use that up in one couch incident a you know they listen to the show because they said couch incident and b it's like okay but like fine you are you are outside the bell curve 20 for one couch incident is way higher than you know so i'm sorry but i'm not you can't configure the app for the outliers that's what you're just trying to run no and and if anybody uses it that heavily they should be paying for it that that is exactly like that is the kind of user that like okay then that makes sense you have you have heavier needs for this you should be paying for it simple as that you know because this is not a free app this is a this is an app that has limited free functionality that is really a subscription priced app and there's nothing wrong with that and people will there will be people who will try to make you feel bad about that but don't because that's life this is what modern software is you have you know you have to keep it up it's yeah modern software is ongoing revenue in some form because you because people expect ongoing updates i'll just use imdb one star all right good luck with that all right so so tell me if i'm if i'm bananas to say that monthly should be an option i i like it i think it's good i think it's a lower number on the screen i think marco was saying should yearly be an option.

2:21:02No, no, no. Well, no. I mean, I think if there's only going to be one, it should be yearly. But I think I'm okay with two. But I think what you want... So, I mean, look, cynically speaking, not everyone who signs up for a year will use it for a year. So you kind of want to push more people towards the yearly if you want to maximize money. So there's that to consider. I would suggest if you're going to offer both monthly and yearly, which i think i'm okay with price them in such a way that people who do math will go for the yearly yeah and that that argues for uh not making it 150 because that makes the math harder for people right like like you know so if you made it you know a dollar a month eight dollars a year that's pretty good like in terms of that will drive people people can do that math in their head real fast right or or make it two dollars a month and make it ten dollars a year you know something like that whatever whatever it is like make it so that people can clearly tell oh i should really go for the year and because that's really what you want if i care about that but i think the smaller number for monthly is is a smoother on ramp for people who say i don't know if i want this eight dollars worth correct and after three months of paying monthly they'll be like oh this is dumb i should just pay for annual maybe yeah but ultimately like you you want this screen you want most people who choose to subscribe to go for the longer term one for lots of reasons number one you'll make more money from people who abandon it early number two those people will be reminded only once a year instead of 12 times a year that they are subscribing to your app.

2:22:33And every time someone's reminded they're subscribing to your app by one of those emails from Apple saying these are about to renew, that's a chance for them to cancel it. So you don't want to have to bother people with that every single month if they're willing to pay for a year at whatever price that is. So ideally, you want yearly customers more than monthly customers. So that, you know, because you'll both make more from them, you know, in terms of the abandonment and then the ones that don't abandon it, you'll keep them longer probably. So a yearly is probably better for you. So yeah, I would price it.

2:23:02I think if your yearly price is going to be eight bucks, I think a dollar a month is a good price for the monthly. I would maybe push yearly to be 10 bucks, but it's not that big of a difference. I know that you're going to have trouble asking for 10 bucks because it's you and you keep talking yourself down and giving yourself pay cuts. well it's it's it's just i mean i honestly if it were no it is up to me but if it were up to me and i felt like you know i could get away with whatever i wanted then i think i i think ten dollars a year is reasonable in like ten dollars a year and like one or maybe two dollars a month but i think that a regular person would look at this and be like no that's way too much money and that's what gives me pause those people are going to say that no matter what these prices are it It could be 10 cents a month, they would say that.

2:23:54Yeah, maybe. I don't know. I feel like keeping it under 10, it feels like it can be an impulse buy at under$10, even on a yearly basis, which is bananas. Because if you go out to eat and get a soda, it's like$2.50. And that literally is pissed away in the span of six hours. But I do this. On the consumer side, I do the same thing. And I'll look at a$10-plus dollar a year app, and I'll be like, eh, do I really need this? So I feel like the most I can get away with per year is$9. When I see those$10 apps, you know what I do? I usually go for the free trial and see if I like it. Oh, Jesus. Of course.

2:24:35No, okay. So look, if you're going to go with$8 or$9, then yeah, I would say in order to maximize the best ratio of monthly to yearly, I'd say$1 a month,$8 a year. And you don't think that$1.50 is – because, I mean, the only reason to do more than$1 and less than$2 is just to eke out a little bit more from the monthly people, which maybe that's dumb anyway because I'd be – strictly speaking, if they stay for a year, which is a big if, then I'm making more with$1 a month than I am$8 a year. But I don't know. It just – does it make sense to make it$1.50 or$2 or something? I mean, you could try. Look, and this is the kind – you can play with this over time.

2:25:16you can actually you can change prices all the time like people don't need to have a comprehensive history of the pricing of your app in their head well i thought for subscriptions though got a little dodgy because then you have to like get the permission or something like that you know better than me mark i don't know if you can change you can actually increase the the subscription of existence but i'm just thinking of like for going forward for new customers for setting aside what existing customers do you just like retire one and start another subscription and like and just keep it going but not have it be purchasable in the app so like like there's there's ways to do it like yeah you just have multiple subscriptions at different price points like you can change like recently you know the whole the disney plus mechanic they added you can increase the price of an existing subscription and bring those users along but i believe they have to you have to give them some amount of like the system gives them some kind of notice and i think they i don't think they have to opt in but they have a chance to opt out so you know that's that kind of runs a risk of loss there so you know ideally you don't need to change these prices but you can change these prices yeah so but yeah i mean look i mean a dollar 50 a month is not bad i just think a dollar a month is an easier it's a it's a more obvious sell it lets the screen look nicer like you have more design options um in terms of your biggest markets like the u.s where it's going to be a nice even number you can you cannot even show the decimal points like you know like so you have options there uh actually no you can't because it's gonna it's gonna say do the new tiers have even like one dollar exactly or is it still 99 cents i thought so i thought they do i that's why i gotta double check i think you're right yeah yeah because i that's why i have it currently is literally 8.00 and either 1.50 or 1.00 you know um it's because i i was thinking about it and and i was listening to um uh shoot uh thoroughly considered earlier today and they were talking a lot about you know the 99 cent thing and i feel like you know having the round number i feel like that just feels nicer yeah and and that's that's partly why i don't want to go all the way to 10 i could be convinced to go nine but i don't think i want to go all the way to 10 because that just even 9.99 i mean that's effectively the same damn thing nine nine feels like 10 like i think that's why i keep going back to eight as instead of like if you're going to be below 10 eight's a really good number.

2:27:30I think people perceive that well. That's kind of where I am too. So, all right. So I think we all then, I'm asking, not telling, we all agree that about$8 a year sounds about right. And Marco, you were pretty perturbed about the monthly idea, but have you come around on that or are we still not loving it? I think I, I mean, look, I obviously I don't have a lot of data here. I have my app that is yearly only, but that's a little bit different in the sense that my subscription is not getting access to core features of the app so it's a very different mechanic um you know in your case this is like you need to convert as many people as possible because anybody you don't convert with this paywall as john said is likely to abandon usage of the app completely and you and then you make nothing from them you know in my in my case if they don't like my my you know premium subscription they can just keep using the app for free and yeah i make money from the ads from them so you know it's a very very different scenario in your scenario i think it does probably make sense to have monthly and yearly especially because the monthly is a cheaper way in the door i don't think you need free trials uh but i could be wrong i have i haven't used them in the app store so like as a as a developer so i'm not sure you know how the conversion rate would be different it does introduce a level of conceptual complexity for the customer uh it introduces complexity to the screen.

2:28:46That being said, I don't think it introduces meaningful complexity to your implementation. That's a flag you set on the in-app purchase and whatever validation you're doing server-side or whatever, it makes that a little more complicated to account for, but for your purposes, it almost doesn't matter if somebody's in a trial or not. You're just trying to make money from them long-term. You don't really care if this search came from a trial user versus this one came from a paying user. I already have support for it in my kind of facade in front of store kit too it's it's generally fairly straightforward unless you're trying to figure out how much time is left in the trial and or how you know when the trial expired then it becomes a little bit interesting but i think i could support the trial if i wanted to i just don't think i want to um and because again i just feel like it's more complicated but yeah it is um but john where do you come down on both monthly and yearly i think you should definitely on both.

2:29:41The other thing that's interesting about monthly, which I thought I've, I've been thinking about if I should message this in app at all, but certainly I'll talk about it on the show is that if you're a Casey list super fan, you could do monthly. And if you stick with it, then you're giving me an extra, you know, what, three,$4 a year. And now I don't think that that's necessarily going to be what anyone would choose to do, except like five of you who I, whom I love dearly, but, um, but it is, it is another way to kind of have an implicit tip jar without actually putting in a tip jar, which I kind of like.

2:30:17Yeah. I mean, there will be some amount of that for sure. I don't know if it's, if it'll be enough to make that have to dictate any choices you make about the pricing of the monthly. Um, it's, that's probably not a massive factor in determining the price. Yep. All right. So it sounds like the brain trust has concluded. a batch of non-renewable tokens up front, or I shouldn't say tokens, of searches up front, somewhere in the order of 20-ish, I think. I'd still go 10, frankly, but go ahead. I'd still go with free trial. I mean, there's no consensus here. This is not decision-making by committee.

2:30:54No, there's some consensus. I have opinions. Marco has opinions. But Casey, you get to decide. You can do whatever you want. Yeah, yeah. No, I hear you. If you're trying to get me and Marco to come to the consensus, it's not going to happen. No, no, no. Not necessarily. I'm just trying to for as much my own benefit as the listeners I'm trying to get kind of the brass tacks where did we all individually land and if there is something that vaguely smells like a consensus what does that look like and it sounds to me like again something to the order of 10 to 20 free searches and then one a day after that I think it sounds like we all agree that monthly and yearly is not such a terrible idea a little bit of disagreement as to whether it should be$1,$1.50, or$2.

2:31:36Do you think I could go all the way to$2 on monthly? You could try it. I mean, I'm not sure. I mean, having something begin with$1 a month sounds like nothing to a lot of people. That is a really good sounding and looking thing. $1.50 is a little more complicated. As John said, it makes the math a little bit less likely to end up with the yearly purchase. But ultimately, that still looks really cheap. two dollars is really cheap but doesn't look as cheap as a dollar fifty so you know i i think there is a lot of benefit like psychologically visually in getting that that one one point something a month that that is i think very attractive and then yearly eight maybe nine dollars i don't know i feel like the the delta between 12 monthly installments and one eight installment, I feel like that's quite a lot of savings for the yearly subscription.

2:32:35That's what you want. You want people to be going to the yearly. Yeah, I know. But I feel like what I want to do is make it like$9 a year, but I still think that I agree with what you were saying earlier. I feel like there's some weird divide between$8 and$9 that you would expect to see between$9 and$10. But I feel like$8 feels cheap in a happy way, whereas$9 does not feel cheap in a happy way. It feels like it's getting, it doesn't feel cheap at all, really. It feels like it's getting expensive. And the thing, what you have to consider here is, you know, it's easy to look at these numbers and be like, well, geez, I'm setting this up so early.

2:33:10And if I make this nine instead of eight, I'm going to make X percent more money. But if the conversion rate is affected even a little bit by these pricing differences, you can quickly erase that margin so if you know if if you convert a decent number more people at eight than you would at nine you will make more money total at eight yeah and and it's and it doesn't take that many more people to make up that kind of difference and the same thing is true of the monthly level you know whatever it is like if you're setting it at you know a dollar fifty a month versus two dollars a month i think you'll get way more people in at a dollar fifty than you would at two and therefore I think you'd make up that difference in volume.

2:33:54This is the kind of stuff that it's one thing to have these kind of gut feelings this is how we think it is. It's really hard to know in advance and it's even harder to be able to make this kind of decision without just testing it and then being willing to change the pricing in the cumbersome ways that we have to do that. One final thing we didn't talk about much but I'm curious both of your opinions. Maybe we'll start with John. do I do some sort of one-time only, you know, like standard style in-app purchase? I'm going to call it a lifetime unlock during this conversation, but I would not refer to it that way.

2:34:29Like I would refer to it as like a one-time purchase or something. Do I do a lifetime unlock? No. Hey, I agree. That's where I've landed too, but tell me why. I mean, because you can always roll it out later and you can roll it out to people who have already paid you a bunch of money, but just don't want to deal with it anymore, but really love the app and are willing to pay a lot and have had time to fall in love with it. But I don't see how anyone is going to see this app, use it enough without doing one of the subscriptions to decide they want to pay a price that's going to be worth your while for a one-time unlock.

2:35:02Keep it in your back pocket. Save it for three years from now from this app when this app is wildly successful, but all the power users are so pissed about paying subscriptions that you can charge them$100. Then we can talk about it again. But now, absolutely not. So for many reasons, the you know whatever even if you didn't call it lifetime unlock that's how people will perceive it they will perceive this as i buy this and now i quote own it and they're going to a expect that it will work forever which b involves you updating it over time because when we build modern software we're building it on quicksand it there is no such thing as software that stays working forever without updates anymore, especially iOS software accessing a web service.

2:35:45Like there's so many moving factors here. Like that's so, you know, you're building on quicksand. This is going to require constant updates, you know, not like every day, but, you know, at least every couple of years. Like suppose you suppose you worked on this, decided, you know, this app's kind of done and abandoned it. And people were and you wanted people to still be able to use it. You would have to put in a certain amount of time every couple of years just to keep it working on the latest versions of iOS and the latest devices. so you know there there's going to be some degree of maintenance over time so you need ongoing revenue from people it's simple as that you need some way to make money from ongoing use of your app to fund ongoing updates of your app which is what everybody expects you to have so you know ongoing revenue simple as that if you don't if you don't have ads then it's that's it like that's those are two options that we know about or you know creepy user data which you don't want to do and i don't blame you so so like that's it like you have to make ongoing revenue from your app somehow to justify considering to update it so that and and lifetime any kind of lifetime or flat rate or however you would brand it it it gives users the impression and the expectation that and the entitlement that they will feel that they deserve access to this app for an indefinite period into the future because they paid to unlock it they paid to own it but that's not the reality of the modern software environment they can't own it they can't like have it forever without ongoing work from you so it's better to not even try to sell that to people don't even sell them on the idea of that anymore because that doesn't exist for an ios app yeah all right brain trust has spoken i appreciate it fellas