00:00:12 ◼ ► I keep thinking back when we asked, remember when we first asked Marco when he thought his house
00:00:16 ◼ ► would be done and they were like, they're telling me like October? I think he said October maybe,
00:00:21 ◼ ► maybe a month after that if they run a little late. I don't remember the exact quote, but.
00:00:44 ◼ ► One week and we've got almost bathrooms. I think those are, that's an important part of the house.
00:01:02 ◼ ► We have a lot of news with regard to the show, happy news, great news even with regard to the
00:01:12 ◼ ► We've been purchased by Apple. Some happy news, we're sunsetting the show, we were given a truck
00:01:19 ◼ ► load of money. How many truck loads of money would that take? It would take a fair, well for Apple
00:01:24 ◼ ► it would be nothing, but for us it would be several truck loads. It would be a drop in the bucket for
00:01:27 ◼ ► them. Anyways, John has been the mastermind of this in the way that John is really, and I mean
00:01:35 ◼ ► this in a good way, not in a bad way at all, John has taken over the show in the best possible way
00:01:38 ◼ ► and is just making changes all over the place. And again, this may sound bad, it is not bad. I am
00:01:54 ◼ ► I think I made this joke last time, but we now have a fully operational Syracuse working only on
00:02:00 ◼ ► this show. You don't even know the right Star Wars quote to reference there, but you're doing it by
00:02:22 ◼ ► Anyway, so we now have a fully operational Syracuse working only on this show and it is
00:02:40 ◼ ► All right, so here's the pitch. We here at ATP are always trying to improve the membership program.
00:02:46 ◼ ► Why do we do that? We want people who are currently members to be happy and stay members,
00:02:55 ◼ ► now. Actually, before we get to that, just to review, last year in the membership program,
00:03:00 ◼ ► we did a bunch of stuff in the same vein. Last year, we decreased the price of annual membership.
00:03:04 ◼ ► Everybody loves that. We released more member specials. In fact, we've been doing one per
00:03:09 ◼ ► month starting since like last summer. We added gift memberships and then we added the very weird
00:03:19 ◼ ► They are weird, but they are amazing. Weird can be beautiful. Anyway, so here's our latest
00:03:27 ◼ ► wonderful change that Casey oversold to make it scary. What did you say? Like happy, wonderful,
00:03:31 ◼ ► great change. When you have that many positive adjectives, it sounds sinister. It's not sinister
00:03:38 ◼ ► Here is what we're announcing today. ATP Overtime. What the heck is ATP Overtime? Well, first,
00:03:45 ◼ ► let me tell you what it does. It solves a problem this show has had since basically the very
00:03:49 ◼ ► beginning. A problem you may not be aware of unless you listen to the bootleg, but we are
00:03:52 ◼ ► very aware of. So we've got a document that we confusingly call the show notes, which is the
00:03:56 ◼ ► document that the hosts look at that has what we're going to talk about in the show in it.
00:04:00 ◼ ► And it's just one document and we update it before each show. And it's got a list of topics
00:04:06 ◼ ► and stuff that we're going to talk about. And that list of topics essentially only grows. We
00:04:10 ◼ ► delete them after we talk about them on the show, but that topic list is always, it's pages and
00:04:15 ◼ ► pages. And when news breaks or something exciting happens or whatever, we prioritize the topics,
00:04:24 ◼ ► But what happens is there's never enough time in the show to fit all the topics. And I know
00:04:27 ◼ ► you're making jokes about how long our shows are, but I'm telling you, even with our shows
00:04:31 ◼ ► being as long as they are, there is never enough time to fit everything. And we always have to end
00:04:35 ◼ ► up cutting stuff. And some of that's really good stuff. So ATP Overtime is the solution.
00:04:40 ◼ ► It is a new segment that comes after the after show for members only. It's for stuff that we
00:04:47 ◼ ► think is, we want to talk about, but that hasn't fit into the show. The target length for this
00:04:52 ◼ ► segment, and you can feel free to quote this to us in years to come and laugh at us, is 15 to 45
00:04:58 ◼ ► minutes. That's what I say the target is. Because if it's only 15 minutes, we'll do 15 minutes. If
00:05:02 ◼ ► it's 45, but we're not making the show like 12 hours long, we'll see how we do. ATP Overtime is
00:05:08 ◼ ► main show content. So anything that would be in the main show is potential for Overtime. A topic
00:05:15 ◼ ► is the obvious choice. Because again, the topic list where we stop recording the show, there are
00:05:19 ◼ ► pages of topics below them. Some of them are super old and no longer relevant, but it's a shame
00:05:23 ◼ ► because some of them were really good back when they were relevant. We just never got to them.
00:05:27 ◼ ► ATP Overtime is not the after show. The after show is where Margo talks about getting his truck stuck
00:05:32 ◼ ► in the sand on the beach. It's where Casey talks about cracking his windshield with his iPad,
00:05:37 ◼ ► even if he talked about that in the pre-show. Anyway, Overtime is not the after show. The
00:05:40 ◼ ► after show is anything goes, total random stuff about our lives. ATP Overtime is just a bit more
00:05:46 ◼ ► ATP. It's tech stuff. Here's why we think ATP Overtime is good for members. It's just more ATP.
00:05:53 ◼ ► Presumably, if you're a member, you like ATP. How about a little bit more? But the point is,
00:05:57 ◼ ► it's a little bit more ATP. It's not another two-hour show. We're not going to bury you in ATP.
00:06:06 ◼ ► And it's stuff that we think is worth talking about because, believe me, there is never a
00:06:10 ◼ ► shortage of topics or Ask ATP or any of the things that we end up having to cut out of the show. We
00:06:16 ◼ ► are going to pick the very best ones of those and get them into Overtime. So they don't end up
00:06:20 ◼ ► getting pushed off the bottom of the topic list by a million different things Apple does with
00:06:25 ◼ ► relation to the EU's DMA, for example, for a recent example. So once again, this will be in
00:06:31 ◼ ► the members only episodes after the after show. It'll be in the edited episode after the after
00:06:36 ◼ ► show. It'll be in the bootleg after the after show. No matter how you listen to your members
00:06:44 ◼ ► there'll be a chapter marker. So that's it. We hope this will keep members happy because they
00:06:50 ◼ ► get a little bit extra stuff. And we hope this book might entice you to become a member because,
00:06:54 ◼ ► hey, you're getting a little bit more ATP. Yep. We are really excited about this. We'll see how we
00:06:58 ◼ ► do in terms of not going for 14 hours each night because Marco and I at least will probably be
00:07:03 ◼ ► snoring at the microphone if we go that long. But like Jon said, 15 to 45 minutes, hopefully closer
00:07:08 ◼ ► to 15, but we'll see what happens. I would guess closer to 15 most of the time. Yeah. But Jon, now
00:07:14 ◼ ► as a member, I'm worried what's happening to the member specials. Are those going away?
00:07:18 ◼ ► Member specials are still there. In fact, we just released the member special moments ago,
00:07:22 ◼ ► and it's a very special member special. This one is now, I think, well established that this is a
00:07:29 ◼ ► unintentional running gag instead of an actual system. The way we name member specials is we
00:07:35 ◼ ► have like a prefix that's like ATP something colon. And so if you've been a member for a
00:07:40 ◼ ► while, you know, we do like ATP tier list colon every iPhone, ATP top four colon laptops, right?
00:07:46 ◼ ► It's always some kind of prefix and then a suffix, but we have too many freaking prefixes. So now we
00:07:52 ◼ ► have to come up with a member special. Like, oh, what is the prefix for this one? It's not,
00:07:55 ◼ ► it's not a movie club. It's not an eats. It's not a top four. It's not a tier list. And so we just
00:07:59 ◼ ► keep making new prefixes and we've done it again. This one is ATP insider, Jon's windows. Longtime
00:08:06 ◼ ► fans of the show might know what that means. We had a well-loved episode way back at episode 96.
00:08:14 ◼ ► When was that in 2014? Maybe it was either 2014, I think it was 2014. Anyway in that episode,
00:08:21 ◼ ► in the after show, I believe Marco and Casey were dumbfounded to learn some of my habits with my
00:08:27 ◼ ► windows in Mac OS. And for years, people have been asking, Jon, can you show us how you use windows?
00:08:32 ◼ ► I don't quite understand. I listened to that episode and I don't understand what the heck
00:08:34 ◼ ► you're doing. And I'd always say, I can't show you how I use my windows. They'd be like, just send me
00:08:38 ◼ ► a screenshot or make a movie or send a screenshot or blur out every window. And it's like, I can't
00:08:42 ◼ ► do that. I can't show you all my windows on my screen because I got all my personal stuff in it.
00:08:49 ◼ ► But I figured out a way to do it. And you'll have to listen to the member special to see it's a
00:08:54 ◼ ► kind of silly solution. But I make an attempt to tell you and show you how I use windows on my Mac,
00:09:02 ◼ ► how I've used windows on my Mac for my entire life, starting with a nine inch screen and moving
00:09:06 ◼ ► all the way up to this gigantic 32 inch screen that I'm staring at now. There is a video version.
00:09:11 ◼ ► I would encourage you to watch it as with all these member specials that have video versions.
00:09:15 ◼ ► If you just look in the show notes for the regular audio only podcast episode, there will be a link
00:09:20 ◼ ► to the video of the episode. It's a YouTube video link. It is unlisted. Please don't share that link
00:09:26 ◼ ► with non members because the whole idea is supposed to be members only, but we have no real
00:09:30 ◼ ► way to do that. Anyway, you can listen to it audio only. We tried real hard to describe what we're
00:09:40 ◼ ► That was a heck of an audio edit by the way. Yeah. The audio version is substantially shorter than
00:10:01 ◼ ► which probably no one looks at, but on your member page or on the join page or a million other places,
00:10:09 ◼ ► should I rename membership fact to be like membership guide or membership help? Because
00:10:13 ◼ ► do people even know what FAQ means? In fact, how many times have I said membership fact on this
00:10:18 ◼ ► show and someone is saying, what is he saying? Membership fact? Membership fat? What is he saying?
00:10:23 ◼ ► FAQ. Frequently asked questions or file of answers and questions, depending on what expansion. So
00:10:31 ◼ ► anyway, it's at atb.fm/membership/faq. And I'm trying to add more and more information there to
00:10:37 ◼ ► explain all of the things that you get with membership, because I just had to add a section
00:10:41 ◼ ► for overtime, obviously. So if you have any questions, you can check that out. And if you
00:10:44 ◼ ► want to become a member, finally, if you heard all this and you're like, yes, I'm sold. I want
00:10:48 ◼ ► to become a member atp.fm/join. Thank you to all of our members that we've already, that have
00:10:55 ◼ ► already joined. Thank you to the future members. And thank you, especially to John, who has really
00:10:59 ◼ ► been spearheading a lot of the work over the last few months. And again, I am so here for it. And I
00:11:04 ◼ ► know I'm speaking for Marco when I say that. Turns out you're doing your homework, and you're doing
00:11:08 ◼ ► your research. I'm here for it. Now that you don't have another job, this is great. This is working
00:11:12 ◼ ► out great for Marco. Oh, one more thing about overtime, I forgot. We will announce in some
00:11:17 ◼ ► format that is yet to be determined, what we're going to talk about in overtime in each episode.
00:11:21 ◼ ► So each episode at some point, I forgot to talk to Marco about this, but that little thing that
00:11:25 ◼ ► you say at the end before the after show, we have to come up with a script for that where you have a
00:11:30 ◼ ► place to stick in what the overtime is going to be about. But yes, at some point in this episode,
00:11:34 ◼ ► one of us will say what the overtime is going to be about. Yeah. And to continue our sports metaphor,
00:11:38 ◼ ► we'll fumble through it till we get it right. Well done. Well done. You guys would have been very
00:11:43 ◼ ► amused at something I was doing earlier this afternoon. There was a wonderful episode of
00:11:47 ◼ ► the App Stories podcast by Mac Stories, where Federico and John were talking about app design
00:11:54 ◼ ► trends, like trying to, you know, look at, trying to look at it, look ahead, like what iOS 18 might
00:11:59 ◼ ► have in terms of app design trends. Cause this is very relevant to me right now as I'm working on
00:12:02 ◼ ► this rewrite to Overcast. I'm like, I I've already decided, like, I'm not going to release it before
00:12:06 ◼ ► WVDC because I kind of want to see whatever iOS 18 brings before I make any like, you know, set
00:12:16 ◼ ► like, they were like, oh, let's look at Apple's recent apps to see maybe some guidelines of where
00:12:19 ◼ ► I'd be going. And they were looking at Apple Sports and Journal. And you should have seen me
00:12:25 ◼ ► trying to navigate the Apple Sports app as somebody who does nothing about sports. So first of all,
00:12:30 ◼ ► you know, you download the app and I'm like, all right, I guess I got to start using this.
00:12:33 ◼ ► The app, while it looks good, is based on a lot of gestures that are not obvious. So that stumped me.
00:12:40 ◼ ► But then the biggest problem was in order to see any content in it, you have to tell it what teams
00:12:45 ◼ ► of what leagues you're interested in seeing the games for. So it presents the list of leagues.
00:12:51 ◼ ► And first of all, I'm like, uh, I guess let's see, like, what's a league I recognize? These are all
00:12:57 ◼ ► these different world sports leagues that most of which I've never heard of. So I'm like, I find,
00:13:02 ◼ ► like, you know, whatever the American ones are. And I'm like, uh, I guess baseball? Okay. Click
00:13:07 ◼ ► on baseball. I guess Yankees. Sure. I know. I know Yankees. I'm in New York. I read Daring Fireball.
00:13:13 ◼ ► So I picked that and there's like, no games. Oh, I guess baseball is not currently happening.
00:13:20 ◼ ► So then I have to like go back to the roots game. Like which of these leaks are currently playing?
00:13:23 ◼ ► Like I'm pretty sure football's over because the Superbowl happened a little while ago. Well done.
00:13:28 ◼ ► So I'm like, it took me a few tries to find like, which of these is a sport that is currently
00:13:34 ◼ ► playing. I ended up finding American soccer. So anyway, the process of me stumbling through
00:13:41 ◼ ► sports, trying to figure out the most bare minimum of like name one sport that is currently happening.
00:13:59 ◼ ► perfect name for us to use. All right. So let's do some follow up. Thank you for bearing with us
00:14:06 ◼ ► again. ATP.FM/join. Thank you so much. All right. Follow up. A call sheet is now available on
00:14:12 ◼ ► visionOS. Hey, congrats. We don't need to belabor this. I just wanted to call attention to it. No,
00:14:16 ◼ ► we should. We should slightly labor it because you shipped a new version of your app. Like,
00:14:21 ◼ ► like that is, look, I can tell you as somebody who attempted and then failed to make a native vision
00:14:26 ◼ ► version of my app, that's no small feat that that is, that is a significant workload. Bring your,
00:14:31 ◼ ► I, your iPhone app to visionOS is significantly more work than bringing most iPhone apps to iPad.
00:14:39 ◼ ► Like, cause it's just, it's a, it's such a different thing and the way it looks as different,
00:14:44 ◼ ► the way you interact with it is different. What's important is different. So this is not a small
00:14:49 ◼ ► thing. This is, this is a, at least a medium sized thing. So please continue. Indeed. You know, it,
00:14:55 ◼ ► it took a lot. Um, it's not flawless, but I'm pretty happy with it. I think it looks pretty
00:15:00 ◼ ► good. I'm still tweaking it. I was making a bunch of changes earlier today. Um, it was funny because
00:15:05 ◼ ► it went from actively embarrassing to, eh, I can at least ship this to test flight too. Okay,
00:15:13 ◼ ► I can make this public relatively quickly. Like once I got out of actively embarrassing,
00:15:18 ◼ ► then it was pretty fast for the rest of the way. Um, and I think it looks pretty good. And I think
00:15:23 ◼ ► it fits in pretty nicely. I think it's a pretty good platform citizen. It is a native app to be
00:15:27 ◼ ► clear. It is not like an iPad app running on the, on the vision pro. Uh, you can tell when that
00:15:33 ◼ ► happens because the iPad apps in compatibility mode or whatever they call it are in light mode,
00:15:38 ◼ ► whereas vision pro native revision, west native apps are in dark mode, if you will. And that's
00:15:42 ◼ ► not literally the case, but that's like the presentation, the way it looks like. And so,
00:15:46 ◼ ► yeah, this is a dark mode, if you will app. And I think it looks pretty great and I think it works
00:15:50 ◼ ► pretty well and you do not need to repurchase if you've already purchased and subscribe via one of
00:15:56 ◼ ► the other platforms, uh, it will just carry right over should do so automatically. In fact, it
00:16:00 ◼ ► shouldn't need you to, uh, you know, manually go in and restore your purchase or anything like that.
00:16:04 ◼ ► So, uh, getting your pins and all that stuff and recent searches should all carry right over
00:16:08 ◼ ► because that's all iCloud. So yeah, check it out. Uh, the one, a couple of caveats. Uh, first of all,
00:16:14 ◼ ► uh, one of the features on iOS is you can change the icon on the iOS app and the iPad app and so
00:16:20 ◼ ► on and so forth. That is literally not possible in vision. OS right now. You may only have one icon
00:16:24 ◼ ► on vision. OS. There is the API that you use as a app developer to change the icon. They tell you
00:16:30 ◼ ► to kindly pound sand. That's not available on a vision. OS. So you're stuck with the icon I chose
00:16:35 ◼ ► and the icon I chose is different. It's one of the alternate icons that my friend Steve had done,
00:16:40 ◼ ► uh, which you can change the iOS app to that same icon if you're a subscriber. Um, but the,
00:16:46 ◼ ► the default icon, which I, which I also love by our dear friend jelly, uh, that one didn't really
00:16:52 ◼ ► lend itself as well to a circle because all the icons on vision. OS are circles. And so I chose
00:16:57 ◼ ► a different one that Steve had done. And so it's a kind of blue, similar clapper board, but it's a
00:17:02 ◼ ► blue background and it looks at first glance, you might think, Oh, who's trying to steal Casey's app
00:17:07 ◼ ► and you know, steal his thunder or whatever. No, no, no. The icon is just different on vision.
00:17:10 ◼ ► OS, but yeah, it is available now. So go check it out. We'll put a link in the show notes.
00:17:18 ◼ ► I had to get it from my work. Yeah. I have to say, when we talked about you getting vision
00:17:22 ◼ ► pro and it's like, Oh, well your app is ideal for it. Imagine if you could have a video player with
00:17:25 ◼ ► call sheet right next to it. You have that already in record time. Like it was, it was just like a
00:17:30 ◼ ► musing fantasy of like, Oh, no, if you're going to be able to do that, or it might be a pain or
00:17:33 ◼ ► what is it going to be like to develop a vision OS. And here we are not too long after the release
00:17:37 ◼ ► of vision OS and one of your screenshots was like, look, video with call sheet next to it,
00:17:41 ◼ ► where I'm looking up the person who's in the video. You even beat the coming soon environments
00:17:45 ◼ ► that are in the environment picker. Indeed. I will, I will take that to my grave. So yeah,
00:17:56 ◼ ► well, a lot of people had some followup with regard to put me on aux. Most it was a smattering of,
00:18:02 ◼ ► you know, yes, this is 100% a thing. And no, that was 100% not a thing, but nobody said it was 100%
00:18:08 ◼ ► not a thing. It's just people, people who didn't know it themselves. But I think even the people
00:18:12 ◼ ► who didn't know themselves can Google and say, okay, it's a thing. That's just not a thing that
00:18:15 ◼ ► they were aware of. That's fair. But regardless, Sam wrote, my 13 year olds take on, put me on aux
00:18:21 ◼ ► quote. That's something an old person would say. I just say, can you approve me on the SharePlay?
00:18:30 ◼ ► they're on the SharePlay. But nevertheless, I am impressed that any human of any age knows that
00:18:36 ◼ ► SharePlay is a thing in the car. So I mean, I have used it once or twice and it does work well,
00:18:41 ◼ ► but I am very surprised that even the kids these days are aware of SharePlay is a thing. So I was
00:18:45 ◼ ► very impressed by that. Or at least one kid somewhere, an Apple marketer is smiling. One of
00:18:49 ◼ ► the interesting things about the feedback was where is the age cutoff between people knowing it and
00:18:54 ◼ ► not knowing it. I learned it from my 16 year old. And it seemed from most of the feedback, especially
00:18:59 ◼ ► families that had multiple kids, that the kids who were 15 or 16 knew it, but the kids who are 10 or
00:19:04 ◼ ► 11 didn't. Now here's the question. Part of the reason put me on aux is so cool is because it's
00:19:08 ◼ ► like the kids who are saying it, who are 15, most likely never actually had to use an aux cable.
00:19:13 ◼ ► Like they're living in a Bluetooth age, but they know it. So that means they got this term passed
00:19:17 ◼ ► on to them from somebody who was older probably. So when those 10 year olds turn 13, 14, 15, or 16,
00:19:25 ◼ ► are they going to then learn and adopt this phrase or is it going to be rejected? Because you'd be
00:19:30 ◼ ► like, oh, maybe this is the younger generation will never use this. But already I think the
00:19:34 ◼ ► current 15 year olds are a younger generation that must have picked it up from an older generation.
00:19:39 ◼ ► So it does have some transference. So tune in, I guess, in five years to find out if those 10 year
00:19:43 ◼ ► olds know about putting me on aux when they're 15. One of our gates is finally over. The 256
00:19:52 ◼ ► gigabyte M3 MacBook Air has two 128 gigabyte NAND chips. So if you recall prior versions of the Air,
00:20:00 ◼ ► and I think as the Pro as well, I forget exactly the timeline. I mean, maybe the M1 MacBook Pro,
00:20:05 ◼ ► it was good, but the M2 it was bad or something like that. But what ended up happening was
00:20:09 ◼ ► they decided to use a single 256 gig chip in certain circumstances. Obviously, if you're
00:20:14 ◼ ► buying a 256 gig computer, and people found that the SSD rights in particular, I believe,
00:20:26 ◼ ► kind of rating it, if you will, not in a literal sense. But you're not splitting that across two
00:20:31 ◼ ► different chips. It's one physical component. And so in teardowns, we've learned that the 256
00:20:37 ◼ ► gig M3 MacBook Air, like I said, has two 128 gig NAND chips. So as per Mac rumors, the SSD in the
00:20:43 ◼ ► M3 model achieved up to 33% faster write speeds and up to 82% faster read speeds compared to the
00:20:49 ◼ ► SSD in the M2 model. So that's pretty cool. Yeah. And this is interesting because what I had heard
00:20:54 ◼ ► about the decision to go with one chip in the M2 model was that I was, as with many of the things,
00:21:00 ◼ ► debated heavily internally. Obviously one side being that it's cheaper to get just the one chip
00:21:07 ◼ ► because of economies of scale and everything like that, but aren't you worried about the speed hit?
00:21:11 ◼ ► And the decision to go with this is based on, look, in real world tests of people doing things
00:21:16 ◼ ► that we know people do with their laptops, there's no way they'd be able to tell a difference. Sure,
00:21:19 ◼ ► if you do a disk speed benchmark, you'll see it, but in actual real world tests of doing real
00:21:24 ◼ ► things is not noticeable. But in the M3 generation, apparently there was enough complaints about this
00:21:28 ◼ ► on the internet that despite Apple's supposed determination that it didn't make a difference,
00:21:33 ◼ ► the fact is people who write articles do run benchmark tests and it doesn't look good. So
00:21:37 ◼ ► I'm glad they reversed this decision, even if it quote unquote wasn't needed. All right. We've got
00:21:43 ◼ ► an interesting bit of Project Titan rumor with regard to chips. Mark Gurman had a post up, I don't
00:21:50 ◼ ► know, sometime the last few days, and it was kind of, I have not noticed these before, but I think
00:21:55 ◼ ► they've been around for a while. It's kind of like a live blog sort of thing. And at 2 27 PM or
00:22:01 ◼ ► whatever day this was, Gurman writes, perhaps reluctantly, the Apple Silicon team was heavily
00:22:06 ◼ ► involved in the Apple car project. Remember the most important part of the car was its AI brain.
00:22:11 ◼ ► The chip Apple developed was nearly finished. It was equal to about four M2 ultras combined.
00:22:17 ◼ ► I'm sorry, what now, Jon? Yeah. So these Q and A's with Gurman, like just so many things like,
00:22:25 ◼ ► look, if you have this information, write it in an article, perhaps reluctantly, is there some story
00:22:29 ◼ ► about the Apple Silicon team didn't like the car project and, or everybody saying perhaps reluctant,
00:22:33 ◼ ► like why would they be reluctant? That's their, the Apple Silicon team, their job is to make
00:22:37 ◼ ► Silicon for Apple products that the Apple car was going to, or whatever they were doing was going to
00:22:39 ◼ ► be anyway, like just, just drop that in there, whatever. And there's not mentioned this in the
00:22:44 ◼ ► actual article, because it didn't seem interesting. So to recap, what we've been waiting for is
00:22:50 ◼ ► essentially two ultras combined into a quad because the ultra is two maxes stuck together and two
00:22:56 ◼ ► ultras would be like four maxes stuck together. And the original rumor back when the M1 was
00:23:01 ◼ ► on the drawing board and we had the rumors about Apple Silicon chips was that there would be the
00:23:04 ◼ ► Jade 4C die, which would be two M1 max, two M1 ultras or four M1 maxes combined. And that would
00:23:12 ◼ ► be in the Mac Pro. We never got that chip. In the M2 generation, we also never got that chip. In the
00:23:17 ◼ ► M3 generation, presumably we will continue to not get that chip for the Mac Pro. But this rumor is
00:23:22 ◼ ► not for two ultras, but for four ultras. And saying this chip is equal to four ultras combined is just
00:23:29 ◼ ► saying like roughly how many transistors, Silicon, whatever. Again, no details are given in this very
00:23:34 ◼ ► brief Q&A. The only thing this makes me think is, well, first of all, they said it was nearly
00:23:38 ◼ ► finished, which means it was not finished. So we're not entirely sure what this would have
00:23:47 ◼ ► a lot of those transitions I imagine have to be spent on GPU/neuro engine type of stuff. If you
00:23:54 ◼ ► look at the Silicon that's in Tesla's self-driving system, a lot of stuff with NVIDIA's machine
00:24:00 ◼ ► learning, AI stuff, it's a lot of transistors spent on stuff that is not as useful to like,
00:24:07 ◼ ► for example, a desktop computer like a Mac Pro, because presumably a Mac Pro is not being used to
00:24:13 ◼ ► do self-driving. You could squint and say, well, it's GPU power and you want a big GPU. So this
00:24:18 ◼ ► would be like a Mac Pro with a giant GPU. And isn't that what you want? Maybe, but the GPUs
00:24:23 ◼ ► and the self-driving things are not being used to render 3D scenes. They're being used to do machine
00:24:29 ◼ ► learning stuff. So I don't know what to make of this, except to say that apparently they thought
00:24:35 ◼ ► they were going to sell more $100,000 Apple cars than they were going to sell $8,000 Mac Pros,
00:24:41 ◼ ► because if they're willing to even go down the road of designing something that is twice the size of
00:24:46 ◼ ► what we wanted in a Mac Pro chip, they say, well, we can justify that because even though it's
00:24:51 ◼ ► expensive and even though we're not going to sell a lot of them, we're going to sell enough of them
00:24:54 ◼ ► in our rumored to be $100,000 car that we think this is worth doing. Or this could all just be a
00:25:00 ◼ ► misunderstanding about a test mule that had some big monster Silicon setup that was never going to
00:25:06 ◼ ► ship for real. Who knows? But anyway, the entire Apple car project was canned. So I'm glad the Mac
00:25:12 ◼ ► Pro project continues to exist. My fingers are crossed for an M something chip in the Mac Pro
00:25:19 ◼ ► that is not the same as that same M something chip in the Mac studio. But again, I don't think that's
00:25:25 ◼ ► going to happen this year. And that last rumor we got about it was that not until after the M7,
00:25:35 ◼ ► All right. So John, if you are a resident of the European Union and you use alternative app stores,
00:25:43 ◼ ► Don't worry, Apple's thought of that. If you thought they would just let you keep using your
00:25:46 ◼ ► stuff, why would they do that? Unless someone is forcing them to, we'll have more than a minute.
00:25:51 ◼ ► According to an Apple document, if you leave the European Union, you can continue to open and use
00:25:56 ◼ ► apps that you have previously installed from alternative app marketplaces. Alternative app
00:26:00 ◼ ► marketplaces can continue updating those apps for up to 30 days after you leave the EU. And you can
00:26:06 ◼ ► continue using the alternative app marketplaces to manage previously installed apps. However,
00:26:10 ◼ ► you must be in the European Union to install alternative app marketplaces and new apps from
00:26:15 ◼ ► alternative app marketplaces. It seems kind of punitive. You can't even get updates if you've
00:26:23 ◼ ► been gone for 30 days. You can't install unless you're in the EU. Why? Why? Because they think
00:26:32 ◼ ► like, oh, and they're on vacation and then use it for the rest of their life in the US? I suppose,
00:26:35 ◼ ► but we'll see how that goes. Yeah. Indeed. Since we last spoke, Epic's developer account has been
00:26:43 ◼ ► reinstated. So this is some coverage from the Verge. Epic Games will be able to open its iOS
00:26:48 ◼ ► app store in the European Union. After all, the game publisher had its developer license revoked
00:26:52 ◼ ► by Apple earlier, I guess it was last week at this point. But Epic Games now says that Apple has
00:26:57 ◼ ► reversed its decision following an inquiry from the European Commission. Imagine that. Apple writes,
00:27:02 ◼ ► "Following conversations with Epic, they have committed to follow the rules, including our
00:27:05 ◼ ► Digital Markets Act policies. As a result, Epic Sweden AB has been permitted to re-sign
00:27:10 ◼ ► the developer agreement and accept it into the Apple developer program." And then there's...
00:27:14 ◼ ► That's quite a statement given the emails we read on the last show. Following conversations with
00:27:18 ◼ ► Epic, they have committed to follow the rules. Epic just showed us an email where they said,
00:27:22 ◼ ► "We are going to follow the rules," and then you cancel their account. So I think the conversation
00:27:26 ◼ ► that you had that changed this, Apple, was probably with the EU and not with Epic, because I imagine
00:27:32 ◼ ► Epic has been saying the same thing. "We'll totally follow the rules this time. You should totally
00:27:37 ◼ ► believe us. You wanted written assurances. Here's a written insurance. Oh, you canceled our account."
00:27:48 ◼ ► whatever thing, and Apple. And as a result, Epic has its account back. And this seems like this
00:27:53 ◼ ► is the way this is going to go, where we kept wondering, "Hey, Apple's done a bunch of stuff
00:27:58 ◼ ► for DMA compliance. Does their proposed plan comply? Is the EU happy with it?" They say, "Yes,
00:28:07 ◼ ► Apple, you complied with our rule." And the answer is they're not going to come out and... So far,
00:28:12 ◼ ► they're not going to come out and say, "No, Apple, you haven't complied." Apparently, there's this
00:28:15 ◼ ► behind the scenes conversation that's resulting in Apple making changes, and we'll have more of
00:28:19 ◼ ► them in a minute. And this is such a weird way to do things, because it basically makes Apple
00:28:23 ◼ ► do stuff, and then a few days pass. Then Apple says, "You know this thing we said we're going
00:28:26 ◼ ► to do? Yeah, we're going to undo a couple of that." And then you wait a few days, and they say,
00:28:29 ◼ ► "Oh, and that other thing that we did? Yeah, we're going to undo that too." To a degree that I was
00:28:33 ◼ ► really shocked about in this next item. But before we move on to that, one tidbit Gruber posted about
00:28:38 ◼ ► this, and he linked to some confirmation that essentially Apple says that they didn't know
00:28:46 ◼ ► that Epic had gotten a dev account back. That Epic just went through the normal channels,
00:28:52 ◼ ► and even though there was a three-day gap between them, and they signed up for when they got it,
00:28:55 ◼ ► Apple didn't really know about it. Epic had assumed, "Hey, we waited three days, they gave
00:28:59 ◼ ► us a dev account. Apple has approved us." But apparently, Apple says, "We just didn't notice
00:29:04 ◼ ► that they got a dev account, and as soon as we noticed..." Well, this is the weird thing.
00:29:08 ◼ ► "And as soon as we noticed, we sent them an email that said, 'Hey, can you tell us that you're going
00:29:12 ◼ ► to comply?' And then they said, 'Yes, we'll comply,' and then we canceled their account."
00:29:16 ◼ ► That's the story Apple is sticking to at this point. It doesn't really make sense to me,
00:29:19 ◼ ► because I feel like if you didn't notice they got an account, and you didn't want them to have an
00:29:22 ◼ ► account, and then you noticed they got it, you would just immediately cancel it and not send them
00:29:25 ◼ ► an email and say, "Hey, we noticed you got an account. Are you going to break the rules? Please
00:29:30 ◼ ► tell us you're not going to." "Oh, you tell us you're not going to? Canceled." Yeah. Like,
00:29:36 ◼ ► you know, John Gruber and Ben Thompson were talking about this on Dithering. Jason Snell had a great,
00:29:41 ◼ ► you know, bit about this on Upgrade. Even if you accept Gruber's assertion that Apple did not know
00:29:48 ◼ ► about the account being created, and then when they learned about it, then Phil sent that letter,
00:29:53 ◼ ► I don't think that changes how incredibly badly Apple/Phil Schiller totally bungled this.
00:30:02 ◼ ► And I say this with a lot of love for Phil as an executive, and from what I understand,
00:30:10 ◼ ► he's done a lot of things there that I really very strongly agree with, and I love what he did with
00:30:15 ◼ ► product marketing as far as I have heard about it. But it's really hard to look at the handling of
00:30:21 ◼ ► this situation and say that he needs to lead the App Store further, because he either made a huge
00:30:26 ◼ ► mistake or was really badly thrown under the bus by someone higher than him, presumably Tim Cook,
00:30:33 ◼ ► and either way, he can't keep running the App Store. I think he can keep running the App Store,
00:30:37 ◼ ► but like, but here's the thing, he did that, and then a few days later, like between our episodes
00:30:42 ◼ ► of our show, guess what, it doesn't matter because it got reversed. And why did it get reversed?
00:30:46 ◼ ► Following conversations with Epic, Apple says? No. This is all backroom stuff that I guess the EU is
00:30:52 ◼ ► saying to Apple, these things you're doing, we don't like them. And so undo them. So despite
00:30:58 ◼ ► all this thing, oh, you accidentally made account, oh, we'll do this back and forth, oh, you're
00:31:01 ◼ ► canceled, oh, guess what, you're not canceled. And Epic, of course, is touting this as a victory,
00:31:05 ◼ ► they say, see, the EU stuff works because they did a thing we didn't like, and then yada, yada,
00:31:10 ◼ ► yada, EU talks to Apple, and now we have our account back. It's just not, it's not going great
00:31:16 ◼ ► for anybody here. Even for Epic, I feel like Epic's getting jerked around, I feel like the EU doing
00:31:20 ◼ ► things behind the scenes, like we really are simply to be settled, like you pass these new guidelines,
00:31:24 ◼ ► Apple tries to comply with them, like, I would like there to be a point where, where we come to
00:31:30 ◼ ► a stopping point where it says, okay, now the EU agrees that what Apple is doing is compliant with
00:31:35 ◼ ► the DMA, but we are not there yet, not by a long shot. - You're right about that, but I cannot help
00:31:41 ◼ ► but look at, like, the situation that just happened with this Epic account, why was it necessary for
00:31:47 ◼ ► Phil to respond or whatever, reach out to Epic, why was it necessary for that email? And wouldn't,
00:31:53 ◼ ► wouldn't this situation have been a hundred times better looking for Apple if the only response
00:31:59 ◼ ► between Apple and Epic was that lawyer's letter saying, we're not gonna allow you for past
00:32:03 ◼ ► behavior and not citing these stupid tweets, like, 'cause the way they look now, first of all,
00:32:07 ◼ ► they look a little bit ineffective because they tried to do something and very quickly got
00:32:11 ◼ ► smacked down by the EU, that's how it looks from the outside, so, you know, they look, they look
00:32:15 ◼ ► weak and bumbling, but also, they look petty and vindictive. They literally put it in writing that
00:32:22 ◼ ► they are severely retaliating against a developer for publicly criticizing Apple. Think about what
00:32:28 ◼ ► that means. Theoretically, if I was smart, I would never say anything about Apple again because my
00:32:35 ◼ ► business is an app in the App Store and what Apple has just shown is that they will, and I mean,
00:32:41 ◼ ► look, we've known this for a while, Jason did a lot, a good bit on this on Upgraded, but we know
00:32:45 ◼ ► there's a lot of people on Apple who have thin skins and they, they will, like, you know, get
00:32:51 ◼ ► mad at people and be a little bit vindictive, you know, behind the scenes, but this is them putting
00:32:57 ◼ ► it in writing in a very high-profile legal case. This is them showing the world, if you criticize
00:33:03 ◼ ► us, we will retaliate and that is, first of all, going to put them in legal hot water, I think,
00:33:09 ◼ ► very easily, like, there's no way Epic doesn't bring this up in a lawsuit, like, there's no way,
00:33:14 ◼ ► but also, think about what that does to the culture around Apple commentary and developers.
00:33:20 ◼ ► Most developers who are smart would never talk about Apple publicly. I do it because I'm an
00:33:26 ◼ ► idiot, but no, like, most people should not be doing this because we know if you insult Apple,
00:33:31 ◼ ► that can impact things like press access for, for the, for journalists, that can impact review unit
00:33:37 ◼ ► access, that can, for developers, that can impact whether you ever get featured again in the App
00:33:41 ◼ ► Store. We know this, we've heard this, we've sometimes experienced this, but for them to
00:33:46 ◼ ► actually go as far as terminating a developer account, using that as a large part of the
00:33:53 ◼ ► justification why, that is a very different level. Whoever, the highest up person at Apple,
00:34:01 ◼ ► who approved that lawyer's letter, using that as an example, should take serious heat for that.
00:34:09 ◼ ► The scale of this blunder, both legally and for reputation for developers and everyone else,
00:34:16 ◼ ► like, that's, this is not a small blunder, this is a huge blunder, and it's so unnecessary,
00:34:23 ◼ ► like, such an own goal. I don't think it's as bad as you think. I don't, I don't even think it's as
00:34:28 ◼ ► bad, I don't even think it's bad legally at all. I mean, it's embarrassing because it makes you look,
00:34:37 ◼ ► people have, you know, litigated this nuance, and I think it is true. Apple didn't say they
00:34:43 ◼ ► were terminating their account because they made mean tweets. They said-- No, they actually did!
00:34:46 ◼ ► They said that! No, no, they didn't. They said, in the past, in the past, you intentionally broke
00:34:50 ◼ ► the rules, and during the time you intentionally broke the rules, you were making mean tweets.
00:34:55 ◼ ► Now, you're asking to have an account back, and you're making mean tweets. We see that as a
00:34:59 ◼ ► pattern of behavior, where last time, you said mean things about us, and then you intentionally
00:35:03 ◼ ► broke the rules. This time, you're saying mean things about us, but you're saying you're not
00:35:07 ◼ ► going to break the rules. We don't believe you, account terminate. That is a wonderful, you know,
00:35:12 ◼ ► excuse, but the reality is they terminated the account, citing that as the reason. Right, but
00:35:16 ◼ ► they didn't say that was the reason. They said we're citing a pattern of behavior, and the pattern
00:35:19 ◼ ► of behavior is you intentionally broke the rules in the past while behaving the same way. Here you
00:35:23 ◼ ► are behaving the same way you did then. Therefore, it's basically a justification of why don't we
00:35:27 ◼ ► believe you. We don't believe you because you're doing the same thing you did last time when you
00:35:31 ◼ ► broke the rules intentionally, right? And here's the other thing. Epic didn't just say mean things
00:35:35 ◼ ► about Apple, and then so they get their account terminated. Epic has sued them, continues to sue
00:35:39 ◼ ► them, continues to battle in court, continues to appeal the things that they have done. Like,
00:35:43 ◼ ► it's way higher level than... Yeah, honestly, as they should. But I'm saying it's not the same as
00:35:48 ◼ ► like, oh, if you say mean things about Apple, they're going to retaliate and do mean things
00:35:52 ◼ ► to you. If you sue them multiple times over, that is a lot... The fact that you can sue them and you
00:36:00 ◼ ► can say you need to have a developer account at all is pretty amazing in the grand scheme of things
00:36:04 ◼ ► because in general, once one company sues another company or you have some lawsuit against Apple,
00:36:09 ◼ ► it's expected that you're not going to be their favorite person. The fact that Apple let Epic
00:36:13 ◼ ► continue to have an account at all is probably because they thought it would be bad for the legal
00:36:16 ◼ ► case. But anyway, I think this is not quite... Because people do say mean things. We say mean
00:36:20 ◼ ► things about Apple. Apple doesn't hate us, right? I mean, I don't think we're universally loved.
00:36:25 ◼ ► You know, the people you cited, Jason Snell, John Gruber, will constantly say, "We'll call out Apple
00:36:29 ◼ ► when they do something bad," and they continue to have access to Apple executives and get review
00:36:33 ◼ ► units and stuff like that. So it's not as cut and dry. I know it's bad to see the mean tweets cited
00:36:37 ◼ ► in there, but technically and legally, they were trying to cite a pattern of behavior to say, "Hey,
00:36:42 ◼ ► and why didn't they decide anything?" They could have just said, "We're terminating your account
00:36:46 ◼ ► because as established in our lawsuit, blah, blah, blah, we're allowed to terminate your account if
00:36:49 ◼ ► we feel like it." That would have been much better. I'm not so critical of them terminating the
00:36:56 ◼ ► account. I'm critical of the communication that led to it and the reasons they cited in doing it.
00:37:03 ◼ ► They really could just terminate the account because Epic violated the rules last time. That's
00:37:08 ◼ ► it. That could be the only reason. They could say, "You created this account. We found that it was
00:37:12 ◼ ► you. You're out." That's it. But first of all, the Schiller letter, which I think that back and forth
00:37:17 ◼ ► was like, "Prove to us you won't do it." They say, "We won't do it," and then they say, "Not enough.
00:37:22 ◼ ► You're out." That back and forth. That was just petty. That entire situation would have been
00:37:28 ◼ ► better if the Schiller email never happened. Huge own goal. That Schiller letter never should have
00:37:33 ◼ ► been sent if they were just going to kill the account anyway. And then secondly, when they
00:37:37 ◼ ► killed the account, the fact that they cited critical tweets that themselves were not breaking
00:37:42 ◼ ► the agreement... As a pattern of behavior. That's a flimsy, very flimsy argument. That's how they
00:37:49 ◼ ► explain it in the thing. I'm not making this up as an interpretation. That's what they literally say.
00:37:53 ◼ ► No, I know. It was flimsy when they said it too. They should not be appearing to be capricious
00:37:58 ◼ ► and petty and vindictive on personal levels like a tweet critical of them. Because, again,
00:38:05 ◼ ► think about what that communicates to not only developers, but also to regulators and lawmakers.
00:38:10 ◼ ► That's not good for them. That's the thing to me is that I think I can see a reality where
00:38:18 ◼ ► it's exactly what Jon is saying. And I think that is how they meant it is that, "Oh, this is a
00:38:21 ◼ ► pattern. This is an example of pattern. This is why we're cutting you off." But I think the thing
00:38:26 ◼ ► that I find most discouraging is that they are so devout in their belief that they are entitled to
00:38:36 ◼ ► and owed what they think they're entitled to and owed from the app store that everything else is
00:38:43 ◼ ► cloudy to them. It's so crystal clear to Apple that they are owed this. We owe them because without
00:38:50 ◼ ► their platform, we wouldn't exist. Leaving aside, that also works the other direction, but we'll
00:38:54 ◼ ► just forget about that for now. Yeah, how's the vision pro app market going there, Casey?
00:38:58 ◼ ► Right? But without Apple, we would not exist. And so we owe them. And I think that the thing that's
00:39:08 ◼ ► so frustrating to me is that they are so tunnel visioned on this. They're so myopically obsessed
00:39:14 ◼ ► with this that they can't step back and see the forest for the trees, the trees for forest,
00:39:17 ◼ ► whatever the turn of phrase is, and realize, and this is what I think you were saying, Marco,
00:39:21 ◼ ► this looks bad. Even if it was done with good intentions, like I 100% buy that they genuinely
00:39:29 ◼ ► had no idea that Epic Sweden had gotten a developer account. I bet that's all automated.
00:39:34 ◼ ► It's a little weird that it took over the weekend for it to get approved, but I can still buy that
00:39:39 ◼ ► that was an automated thing. Honestly, that's within normal parameters. Developer accounts take
00:39:42 ◼ ► multiple days to get approved often. Especially over a weekend, that's fine. And I think that's
00:39:45 ◼ ► totally irrelevant. Like that detail is totally irrelevant to the story. Yeah. And so like,
00:39:50 ◼ ► I don't think they're completely full of it. I think it's just that they are so obsessed with
00:39:56 ◼ ► this. We are owed were friggin entitled that they just can't see how gross this looks, whether or
00:40:04 ◼ ► not it's legal, whether or not it's moral. I think everyone seems to agree. It's just gross. And
00:40:13 ◼ ► that's the thing that just stinks. But you know, it is what it is. We should move along. We have
00:40:18 ◼ ► a lot to talk about. We have more options for apps distributed in the European Union. So Apple has
00:40:24 ◼ ► announced we are providing more flexibility for developers who distribute apps in the European
00:40:28 ◼ ► Union, including introducing a new way to distribute apps directly from a developer's website.
00:40:33 ◼ ► Whoa. Developers who agreed to the alternative terms addendum for apps in the EU have new options
00:40:38 ◼ ► for their apps in the EU. The alternative app marketplaces can choose to order a catalog of
00:40:43 ◼ ► apps solely from the developer of the marketplace. So you can have a Facebook only app store or an
00:40:48 ◼ ► epic only app store. When directing users to complete a transaction for digital goods or
00:40:52 ◼ ► services on an external webpage, developers can choose how to design promotions, discounts,
00:40:56 ◼ ► and other deals. The Apple provided design templates, which are optimized for key purchase
00:41:00 ◼ ► and promotional use cases are now optional. So I should back up a half step. I believe that what
00:41:06 ◼ ► Apple said was, Hey, if you're going to link out to a webpage or what have you, you must use either
00:41:12 ◼ ► this design or one of these handful of designs. I forget the details. And that is the only choice
00:41:16 ◼ ► you have. And now they're saying, well, no, you can do what you want. Although we have strong
00:41:20 ◼ ► suggestions about what you might want to do. You can do it how you want. But that is definitely the
00:41:25 ◼ ► less interesting of the two points here. What you just read through quickly is what we were talking
00:41:29 ◼ ► about back when we were wondering, how is Apple going to allow sideloading in the EU? Are they
00:41:34 ◼ ► actually going to allow people to go to a webpage and click on something and get an app installed
00:41:38 ◼ ► on their phone? And when they came out with their DMA compliance, we're like, no, they're not going
00:41:43 ◼ ► to allow it. But now here they are again, Apple saying, we're providing more flexibility. Why,
00:41:48 ◼ ► why Apple? Why are you providing more flexibility? Probably because of discussions with the EU. Let's
00:41:53 ◼ ► say again, all happening behind the scenes with no announcement from either party, but this is a
00:41:58 ◼ ► radical change to Apple's DMA compliance instead of a third party marketplace where you need to
00:42:03 ◼ ► have a million dollars and you have to post other people's apps and yada, yada, yada, blah, all of
00:42:06 ◼ ► these things that we talked about back when they first came out with their plan. Now it's like,
00:42:11 ◼ ► well, okay, that, but now you don't need a million dollars. And by the way, also you can literally
00:42:16 ◼ ► sideload from a webpage. That is a big change, a huge change, a gigantic change. Yeah. And first
00:42:24 ◼ ► of all, I think even though third party app stores did not have a lot of chance before, I think this
00:42:31 ◼ ► kills them now that the larger well-known apps can theoretically just distribute through their
00:42:37 ◼ ► websites. I don't see how there's a market for alternative app stores after this. It was already
00:42:40 ◼ ► going to be difficult, but that's it. There might be because, so the one thing that haven't changed
00:42:46 ◼ ► is the core technology fee, which really makes this whole economically dodgy and the user
00:42:50 ◼ ► experience of having a third party app store might still be desirable. We'll see how this shakes out,
00:42:54 ◼ ► but like I said, we don't know where this ends. Like, is this the end? Is this, is this what it's
00:42:59 ◼ ► going to be like? Or next week, is there going to be another change? Like they keep Apple started
00:43:07 ◼ ► compliant. And then slowly, gradually over the days and weeks, Apple has been chipping away or
00:43:13 ◼ ► the EU has been chipping away what Apple wanted to do and said, Apple, you said you wanted to do this,
00:43:17 ◼ ► that, and that. Well, no, you got to back that off and you got to back it. We don't know what's
00:43:21 ◼ ► happening on it. All we know is Apple is loosening up progressively. And this is the biggest, the
00:43:26 ◼ ► biggest one they've done, which is like that whole scheme about third party app stores. How about you
00:43:30 ◼ ► just let people install from web page? Still get to pay the core technology fee and a bunch of other
00:43:34 ◼ ► restrictions that we're going to get to. But by the time you listen to this episode, will the
00:43:37 ◼ ► restrictions we're about to list still be in effect? Who knows? Stay tuned. So what does it
00:43:43 ◼ ► take? If you want to distribute from your website, from your own website, if you want to distribute
00:43:46 ◼ ► an iOS app from your own website that goes right into people's phones, what do you have to do?
00:43:52 ◼ ► Web distribution available with the software update later this spring will let authorized
00:43:55 ◼ ► developers distribute their iOS apps to EU users directly from a website owned by the developer.
00:43:59 ◼ ► Apple will provide authorized developers access to APIs that facilitate the distribution of their
00:44:07 ◼ ► and more. For details, visit Getting Ready for Web Distribution in the EU, which we will link
00:44:12 ◼ ► in the show notes. I don't think that you need to have an account for this, like a developer account
00:44:16 ◼ ► for this, but you might. Anyway, continuing along, to be eligible for web distribution, John, you
00:44:21 ◼ ► must be enrolled in the Apple Developer Program as an organization incorporated, domiciled, and/or
00:44:27 ◼ ► registered in the EU, or have a subsidiary legal entity incorporated, domiciled, and/or registered
00:44:32 ◼ ► in the EU that's listed in App Store Connect. And/or, I'm not sure if this is an and bullet or
00:44:37 ◼ ► an or bullet, but the second bullet is... - It's an and. - Thank you. Be a member of good standing in
00:44:41 ◼ ► the Apple Developer Program for two continuous years or more, and have an app that had more than
00:44:46 ◼ ► one million first annual installs on iOS in the EU in the prior calendar year. - Apple's always
00:44:53 ◼ ► pulling out the restrictions. Yeah, so the two continuous years in good standing, that means not
00:44:58 ◼ ► you, Epic. And have an app with more than one million first annual... That's just like, we don't
00:45:06 ◼ ► want a lot of people to do this. One easy way to do that is say, hey, you need a million installs
00:45:11 ◼ ► in the EU in the prior calendar year. It doesn't say you need a million installs of any specific
00:45:16 ◼ ► app. You just need to be... They want to eliminate the vast majority of potential people. So you have
00:45:24 ◼ ► to account for two years, and you've got to have some app that's successful enough to have a
00:45:28 ◼ ► million installs in the EU only, not worldwide. That narrows the field down so much. If you think
00:45:33 ◼ ► you're a hobbyist developer, and you're like, I'm going to make an iOS app, and I'm going to
00:45:36 ◼ ► distribute it from my homepage. No, you're not. Because first you need an app with a million EU
00:45:42 ◼ ► installs in the prior calendar year. So go do that, which is real easy to do, let me tell you,
00:45:47 ◼ ► and then come back and now you can distribute an app from your webpage. So despite the fact that
00:45:52 ◼ ► Apple has given so much and say, we're going to let you distribute apps from the web. But most
00:45:56 ◼ ► people know you can't. Here's a bunch of rules. You don't need a million euros in the bank,
00:46:04 ◼ ► get a million installs in the EU from the iOS app store, even with a free app. Good luck.
00:46:11 ◼ ► Yeah, very true. Yeah, the other thing is the core technology fee, which is the thing where you owe,
00:46:18 ◼ ► what is it like 27% or no, no, no, it's a dollar or I guess not a dollar. That's the half euro per
00:46:23 ◼ ► install per year. There it is. Yeah, thank you. That is applicable over 1 million first annual
00:46:29 ◼ ► installs, right? So they're basically saying you will be paying us the core technology fee. Well,
00:46:33 ◼ ► no, because you could just cancel that app. So you could have a successful free app in the regular app
00:46:37 ◼ ► store in the previous calendar year that you get a million installs, then you just cancel that app
00:46:42 ◼ ► and don't make it anymore. And then you start like it's just arbitrary nonsensical. There's nothing
00:46:47 ◼ ► about having a popular app that makes you more worthy of being able to distribute app from your
00:46:52 ◼ ► web page or something, right? Like if anything, the things that are able to get a million installs in
00:46:56 ◼ ► the EU are probably like, you know, scam apps with big come ons that can get a lot of installs with
00:47:01 ◼ ► their free app to try to sell you their weekly subscription or something. It's just, it makes no
00:47:05 ◼ ► sense. And like, I'm hoping by next week, the EU will have talked to Apple about this and said,
00:47:10 ◼ ► yeah, that's not really what we meant. Get rid of like, why don't they just meet with each other and
00:47:18 ◼ ► you're compliant. This back and forth is ridiculous. Well, that's, I have a feeling that's
00:47:22 ◼ ► how this change happened. I mean, if you, if you look at the timeline of this, this is effectively
00:47:27 ◼ ► direct sideloading from a website. Yes, with restrictions, but it's sideloading. It's not
00:47:30 ◼ ► alternative app stores. I know, but they're, they're announcing these incremental steps. Like every
00:47:34 ◼ ► time that Apple has a concession, they announce a new thing, like just figure it all out at once.
00:47:39 ◼ ► Well, that's my theory is their first plan. They thought that this should be sufficient to satisfy
00:47:44 ◼ ► them. The EU, somebody in, in somewhere in the world, whether it was, you know, somebody like
00:47:49 ◼ ► Spotify or like Epic or somebody, or just the EU commission itself decided, not good enough. Let's
00:47:55 ◼ ► apply some pressure and negotiate behind the scenes. And Apple had to then, you know, give
00:48:00 ◼ ► this up too. And that's why it's not there yet. And it will be added later. You know, so this is
00:48:05 ◼ ► obviously what's happening here is like Apple's trying to do the least possible to comply and
00:48:10 ◼ ► make it as restricted, difficult and expensive as possible for anybody to do that. Hopefully nobody
00:48:15 ◼ ► will actually do it. And then the rest of the industry sees that and says, uh, that's too
00:48:24 ◼ ► following the law. They go lobby the EU commission stuff happens behind the scenes. Pressure is
00:48:29 ◼ ► applied in various ways. Politics happen. So that's probably what's happening here. And I think this
00:48:34 ◼ ► is going to be a process that lasts months or years, because even after Apple finally kind of,
00:48:41 ◼ ► you know, reaches some kind of stasis with the European commission, uh, then, you know,
00:48:46 ◼ ► lawsuits and lobbying will continue in the background and, you know, stuff will change.
00:48:51 ◼ ► And over time, they'll have to make more changes though, or there'll be pressure to make more
00:48:54 ◼ ► changes or new legislation will be drafted or some other regulator somewhere else or somewhere
00:48:59 ◼ ► nearby or somewhere within there will make a new regulation. Like this is going to keep going
00:49:03 ◼ ► forever until Apple relents in larger ways that they probably are not going to do unless very,
00:49:10 ◼ ► very forced. You know, so there's these really huge fines that are possible under the DMA. Like,
00:49:15 ◼ ► it's like 10% of worldwide revenue or something, which is more than Apple makes in the entire EU,
00:49:23 ◼ ► Like, but when does that happen? Like this, apparently what they tried before wasn't compliant
00:49:28 ◼ ► is this thing that they've said now compliant. Like at a certain point of going back and forth,
00:49:32 ◼ ► I feel like when, when are they in violation? I forget what the timeline is. Maybe there's
00:49:41 ◼ ► could work it out amongst themselves before announcing it. Like there's no reason Apple
00:49:44 ◼ ► needed to even announce their plan publicly. They could have given the plan to, I know it's not the
00:49:48 ◼ ► way it works. The EU wants you to, but anyway, like, I feel like it would be better if they,
00:49:51 ◼ ► two of them got together and did not leave the negotiating table until they both agreed that
00:49:56 ◼ ► what they had to announce was compliant and then Apple could just announce it once. But instead
00:49:59 ◼ ► we're doing this onesie twosie thing where slowly Apple is chipping away at its terrible plan to be
00:50:03 ◼ ► slightly less terrible. On the one side, I applaud them for embracing web distribution, but on the
00:50:09 ◼ ► other side, like you, well, I think that's what they're doing. Enabling, I guess is a better word
00:50:15 ◼ ► for it, but, um, even that's a stretch. Those videos are like you're, you're taking, you're
00:50:19 ◼ ► trying to drag a toddler off, take a nap, but they're just actually literally dragging their
00:50:23 ◼ ► limp body on the ground. That's what Apple's doing. That's fair, but no, I mean, I, I'm glad
00:50:32 ◼ ► but golly, so many gotchas, so many caveats, so many asterisks, so many daggers and double
00:50:36 ◼ ► daggers and so on and so forth. And again, it's by design. Like they really, they really don't want
00:50:42 ◼ ► anyone doing this and it shows, and that's why it's going to just keep being this back and forth.
00:50:46 ◼ ► And again, it's real, I don't want to belabor this, but it's really too bad because I feel like
00:50:51 ◼ ► if Apple had been less tunnel visioned and less, you know, we are entitled like we were talking
00:50:56 ◼ ► about a minute ago, if they had done some of this of their own volition, I really genuinely think,
00:51:02 ◼ ► and obviously there's no way to prove it, but I genuinely think that a lot of this, uh, attention
00:51:07 ◼ ► from the EU and from other countries as well, I think it would not necessarily be pointed at
00:51:12 ◼ ► Apple or, you know, it just wouldn't be a thing at all if they had given even just a few inches,
00:51:16 ◼ ► maybe not one inch, but a few inches, or I guess it's EU, a few centimeters or whatever, um, that
00:51:22 ◼ ► I feel like maybe this wouldn't have been a thing, but because they're so just stuck in their ways
00:51:28 ◼ ► and so entitled, this is where we are. And you know, you made your bed and I get to sleep in it.
00:51:33 ◼ ► Someone in the chat room just posted a link to a tweet, uh, that I don't know how they know this,
00:51:38 ◼ ► maybe they're using betas or something, but saying that it takes 15 steps to install an app from the
00:51:42 ◼ ► web using the newly proposed Apple flow, uh, and it's a thread that goes through all the steps.
00:51:46 ◼ ► So, uh, yeah, as you would imagine, doing it from a webpage is not as simple as clicking one link
00:51:51 ◼ ► and then voila, the app appears on your home screen because of course Apple wouldn't do it
00:51:55 ◼ ► that way. Um, but it's, you know, having it be possible is better than nothing, but it would be
00:52:01 ◼ ► nice if it wasn't as painful as it appears to be according to these steps. We'll link the tweet in
00:52:04 ◼ ► the show notes. We are brought to you this episode by Magic Lasso Adblock, an ad blocker that's easy
00:52:11 ◼ ► to set up and with pro features that are really, really great. Magic Lasso Adblock is the ad blocker
00:52:16 ◼ ► designed for you. It doubles safaris loading speed. It's an efficient high performance and
00:52:21 ◼ ► native ad blocker for iPhone, iPad, and Mac simply blocking all intrusive ads, trackers,
00:52:28 ◼ ► and annoyances in Safari. They have over 5,000 five-star reviews. So it's simply the best Safari
00:52:34 ◼ ► ad blocker. They block intrusive trackers. So this also delivers a more secure browsing experience
00:52:39 ◼ ► and stops you from being followed by ads around the web. This keeps your browsing history from
00:52:44 ◼ ► being visible to ad networks. And it even blocks a lot of YouTube ads, including all pre-roll,
00:52:49 ◼ ► banners, search, and suggested product ads. With Magic Lasso's pro features, you can also craft
00:52:59 ◼ ► can do things like prevent playback of media, stop personalization of certain websites, or just
00:53:03 ◼ ► streamline really heavy websites. You can customize pages to your liking. One of my favorite ways to
00:53:08 ◼ ► do this is a feature called tap to block. You can just simply tap any element and it disappears and
00:53:13 ◼ ► it will never show up again on that site. And also similar elements will also not show up. So
00:53:20 ◼ ► performance insights, where you can visualize exactly the speed, energy efficiency, and data
00:53:26 ◼ ► savings that Magic Lasso ad block is delivering for you. So you can see things like how much
00:53:30 ◼ ► carbon you've saved with the energy that you've saved. It's really a pretty eye opening feature
00:53:35 ◼ ► and you are saving huge amounts of data and time and everything with Magic Lasso ad block. So
00:53:40 ◼ ► join over 300,000 users and download Magic Lasso ad block today from the app store. Unlike other
00:54:03 ◼ ► to receive one month free access to all features. Our thanks again to Magic Lasso ad block for
00:54:08 ◼ ► their support of our show. Some potentially happy news about error network change, isn't there?
00:54:17 ◼ ► Yeah. People have been sending me information about the Chrome error that has been happening to me for
00:54:23 ◼ ► I guess months now and not just me happening to many people in the world. And I've been monitoring
00:54:27 ◼ ► the bug reports in the various Chromium bug issue tracking systems since we first talked about it.
00:54:34 ◼ ► In fact, when we first talked about it, I think I linked to like four separate bug reports,
00:54:38 ◼ ► some of them many years old talking about this. But now it's been getting a little bit of traction.
00:54:43 ◼ ► More people are encountering this. Some of the bugs have been consolidated into a particular
00:54:47 ◼ ► issue that we will link in the show notes. Here's the news though. Well, first people were excited.
00:54:51 ◼ ► I think the last week of the week before they're like, "Oh, its priority has been changed from P2
00:54:55 ◼ ► to P1." And in most bug tracking systems, P1 is like the most important, highest priority bug,
00:55:00 ◼ ► like must get fixed. This is a serious issue, right? In practice, the only ones that ever get
00:55:04 ◼ ► fixed. Yeah, pretty much. P2 was like, "Ah, maybe we'll get to it." Anyway, it got changed to P1,
00:55:08 ◼ ► but then somebody changed it back to P2. And I was like, I didn't talk about it. I'm like,
00:55:12 ◼ ► "Well, whatever." But something happened a few days ago that is worth reporting on. And that is
00:55:18 ◼ ► that someone, the bug was assigned and someone had attached a patch to the bug report saying,
00:55:24 ◼ ► "I think this fixes the issue." Here's some text from that report. "Chromium uses the dynamic store
00:55:33 ◼ ► API to monitor IP address changes. It also resets TCP and QUIC connections whenever it receives a
00:55:39 ◼ ► notification of a network state change from the API. This logic was introduced over 10 years ago.
00:55:48 ◼ ► This causes an issue in Mac OS 14 where a UTON interface," we've heard about this in past
00:55:52 ◼ ► episodes, "for communicating with iOS devices is regularly created and destroyed." And by the way,
00:55:57 ◼ ► commentary here, lots of people are like, "Oh, if you just don't have any iOS dev devices,
00:56:01 ◼ ► if you just do this, if you just do that, it fixes the problem." Every single thing anyone has
00:56:04 ◼ ► suggested to me, I have tried. And a lot of them make the problem occur far less frequently,
00:56:09 ◼ ► but none of them have eliminated it, including not having iOS devices, including not even having
00:56:13 ◼ ► Xcode installed, all that stuff. It helps, but the problem, the root problem is they're using
00:56:19 ◼ ► this thing that looks for network changes. Whenever there's a network change, Chrome flips out.
00:56:23 ◼ ► And Mac OS 14 has more network changes. Those UTON interfaces add more network changes, but even if
00:56:29 ◼ ► you do none of those things, there are still some network changes occasionally. And when you get
00:56:33 ◼ ► one, it blows up Chrome and makes all the HTTP requests fail. Anyway, here's the change that
00:56:39 ◼ ► was introduced to fix this. This change introduces a new flag, "Reduce IP address change notification."
00:56:45 ◼ ► This is like with the Chrome flag system where they have a feature and it's not enabled by
00:56:49 ◼ ► default, but you can enable it. I'm not sure when this will land, if at all. I think the patch needs
00:56:54 ◼ ► to be approved or whatever. When it does land, it will arrive as a flag. In Chrome, you can go
00:56:59 ◼ ► to chrome colon slash slash flags, I believe, and see all the flags and look for this one.
00:57:04 ◼ ► When this feature is enabled, Chromium on Mac ignores notifications from the dynamic store API
00:57:10 ◼ ► if the network interface Delta, meaning the change between the old set of interfaces and the new one,
00:57:14 ◼ ► is a non-primary interface with a local IPv6 address, because those UTON ones are local
00:57:20 ◼ ► IPv6 addresses. I don't think this is a great patch. I'm just saying, "Okay, still freak out,
00:57:26 ◼ ► except when it's a non-primary interface with local IPv6 address," because they're basically
00:57:31 ◼ ► kind of micro-targeting these UTON things because the people reading the bug think that's the only
00:57:36 ◼ ► issue. I don't know why they think that. There's tons of bug reports of people saying, "Hey, you
00:57:40 ◼ ► keep telling me that if I get rid of my iOS dev device, this won't happen. Well, I don't even have
00:57:45 ◼ ► Xcode installed and I'm not a developer and it's happening to me." So I'm not super optimistic,
00:57:50 ◼ ► but is it exciting that someone has proposed a patch to fix this? Stay tuned if they actually
00:57:55 ◼ ► did. Good luck, Jon. Good luck to all of us. I see it all the time now. I don't use Chrome.
00:58:01 ◼ ► Well, I don't use Chrome except when I'm doing stuff related to this very program and analog.
00:58:07 ◼ ► And my goodness, I see it all the time now. It's constant. I mean, you are literally using an iOS
00:58:12 ◼ ► dev device in Chrome or whatever, but again, that shouldn't be a reason that Chrome should
00:58:16 ◼ ► stop working. But there's reports from people who do not have Xcode installed and are not developers.
00:58:29 ◼ ► and there was a big brouhaha about how... I don't know, how can we summarize this well? I guess
00:58:35 ◼ ► when you had Google Gemini, which is their chat GPT equivalent, when you asked it to generate
00:58:41 ◼ ► things from history, like images of history, where take the founding fathers of the United States,
00:58:48 ◼ ► for better or for worse, founding fathers of the US, pretty much all white dudes. And so when you
00:58:53 ◼ ► asked Google Gemini to generate a picture of the founding fathers, you would see a racially diverse
00:58:58 ◼ ► set of people. And everyone started scratching their heads and saying, "Huh? That doesn't seem
00:59:03 ◼ ► right at all." So there's a Verge article about this, which we'll put in the show notes. Oh,
00:59:07 ◼ ► it was actually late February, a collie time. How does it work? Anyways, Google apologizes for
00:59:13 ◼ ► missing the mark after Gemini generated racially diverse Nazis. That's fun. Google has apologized
00:59:18 ◼ ► for what it describes as "inaccuracies in some historical image generation depictions," with its
00:59:23 ◼ ► Gemini AI tool saying its attempts at creating a wide range of results missed the mark. The
00:59:30 ◼ ► statement follows criticism. It depicted specific white figures like the US founding fathers or
00:59:33 ◼ ► groups like Nazi era German soldiers as people of color, possibly as an overcorrection to
00:59:38 ◼ ► long-standing racial bias problems in artificial intelligence. A quote from Google, "We are aware
00:59:44 ◼ ► that Gemini is offering inaccuracies in some historical image generation depictions," says
00:59:48 ◼ ► the Google statement posted this afternoon on Twitter. "We are working to improve these kinds
00:59:54 ◼ ► of depictions immediately. Gemini's AI image generation does generate a wide range of people,
00:59:59 ◼ ► and that's generally a good thing because people around the world use it, but it's missing the mark
01:00:03 ◼ ► here." Whoops. So this is an older story, and I know that we didn't get to this. I see ATP
01:00:09 ◼ ► over time. We didn't get to this when it was happening because other things were happening,
01:00:11 ◼ ► but I wanted to bring it up still on the show because I think it is an interesting problem that
01:00:19 ◼ ► a lot of people, including Apple presumably this year, will have to wrangle with with these large
01:00:24 ◼ ► language models. So these large language models are trained with a tremendous amount of data. They
01:00:30 ◼ ► need a lot of data to be useful, and a lot of that data comes from scraping the internet, but it's
01:00:37 ◼ ► not data that these companies produce themselves. They go look out into the wider world and say,
01:00:41 ◼ ► "Where can we get data?" There are lawsuits about this, but where can they get data? They can get
01:00:45 ◼ ► it from publicly accessible web pages. They can get it from the New York Times. They can get it
01:00:49 ◼ ► from YouTube. They can get anywhere they can find data available, and there's a lot of it out there.
01:00:53 ◼ ► They want to feed that in to train these models to make them do useful things, and they even
01:00:58 ◼ ► hire large numbers of people to do question and answers, like to do as part of the training where
01:01:05 ◼ ► they'll have a human being write a question and then answer it yourself and feed that into
01:01:10 ◼ ► our model so that it understands the whole question and answer thing. They just need lots of data,
01:01:15 ◼ ► right? And the problem with that is you don't get to pick, with the exception of the people you hire,
01:01:20 ◼ ► to do the question and answer. You don't get to pick what the data that you're training on is.
01:01:25 ◼ ► The internet is the internet. The data is out there, right? If you want it, you have to consume
01:01:31 ◼ ► it, and then if you do consume it, you're consuming every part of that data. And that's why a lot of
01:01:38 ◼ ► the companies that invent these things have a problem. They're like, "Hey, we trained this on
01:01:42 ◼ ► the internet. Now ask it a question." And someone would ask it a question, and it would come back
01:01:47 ◼ ► with some terribly racist response. You'd be like, "Where did that come from?" And the answer is,
01:01:52 ◼ ► the internet. That's where it came from, because it has been trained on the world of human knowledge
01:01:58 ◼ ► and the world of human information. The world of human information is filled with terrible things.
01:02:03 ◼ ► All of the good and bad things about us are contained in all the things that we produce,
01:02:06 ◼ ► all our biases, all our racism, all our ignorance, all our right and wrong answers about everything.
01:02:12 ◼ ► That's all on the internet. And it's not like, "Okay, well, why don't you just train on the good
01:02:15 ◼ ► data?" Well, it's not that easy to tell what the good data is. And it's just so far we haven't
01:02:22 ◼ ► cracked that problem. So they are trained in that. And now why is that a problem? Well, one thing is,
01:02:26 ◼ ► in the beginning when they had these large language models, they'd say offensive things.
01:02:30 ◼ ► You'd write something, they'd say something offensive back, and people would get... They're
01:02:33 ◼ ► like, "Well, I never..." They get upset by it. People don't want that. So they're like, "Well,
01:02:38 ◼ ► okay. Well, how about we try to make our model not do offensive things by either eliminating data
01:02:47 ◼ ► that we don't want it to be trained on from the training set or after the fact trying to massage
01:02:51 ◼ ► it in the right direction by doing..." It turned out to be very tricky, but there are some things
01:02:56 ◼ ► that you can do in that regard. With this specific one, what Google was trying to deal with was this
01:03:00 ◼ ► problem. Not that their model shouldn't be able to generate pictures of white people or whatever.
01:03:11 ◼ ► like an image generator, say, if you ask for a picture of the founding fathers, you just want it
01:03:15 ◼ ► to be a picture of the founding fathers, and the founding fathers are all over the internet,
01:03:18 ◼ ► and it's probably easy to find pictures of them and generate a picture of them, it's fine.
01:03:20 ◼ ► But if you say something like, "Give me a picture of a doctor," that's where these models run into
01:03:28 ◼ ► problems. Because you've just asked for a picture of a doctor, but as far as this thing trained on
01:03:32 ◼ ► the entire world is concerned, most doctors are men, white men, whatever is the dominant thing in
01:03:40 ◼ ► the training set. Sometimes a training set could be accurate. I don't know what the actual ratio of
01:03:45 ◼ ► men to women doctors in the world is or whatever, but sometimes you could say, "Okay, well, but
01:03:48 ◼ ► I made sure that this model was trained on real accurate data, and X percent of the doctor pictures
01:03:53 ◼ ► are men, and Y percent are women." It's exactly what it is in the real world. Still, if you have
01:03:59 ◼ ► a model that says, "Give me a picture of a doctor," and X percent of the time it shows you a man,
01:04:06 ◼ ► and Y percent it shows you a woman, and those percentages exactly match what's in the real
01:04:09 ◼ ► world, you'd be like, "Victory, our model is perfect." And we would say, "No," because we think
01:04:15 ◼ ► the ratios of men and women doctors are still out of balance in the real world. We don't want it to
01:04:21 ◼ ► be 60/40 or 70/30 or whatever imbalance it is in favor of men. I don't know what the actual ratios
01:04:26 ◼ ► are or whatever. We would like it to be to match the ratio of men and women population in the world,
01:04:31 ◼ ► which is roughly 50/50, because there's nothing inherently inherent about being a man or a woman
01:04:38 ◼ ► that makes you more or less able to be a doctor. So the model should not reflect the biases that
01:04:44 ◼ ► are resulting in this, you know, this biased set of actual doctors in the world. But the problem is
01:04:50 ◼ ► all the training data reflects either biases that are in the world or the reality in the world,
01:04:56 ◼ ► and the reality in the world is the result of unequal opportunity, for example, you know what
01:05:01 ◼ ► I mean? That's a tricky problem to deal with, because you have to train on large amounts of
01:05:08 ◼ ► data, but all that large amount of data, best case, accurately reflects the world. Worst case,
01:05:14 ◼ ► accurately reflects the biases of all the humans in the world. And that is an extremely tricky
01:05:20 ◼ ► problem. And the way—I don't know if this is true or not, because I couldn't chase this down to see
01:05:24 ◼ ► if someone was just making this up or whether this is actually how it's working—but one thing I saw
01:05:29 ◼ ► was that, like, hey, if you type in "show me a picture of a doctor," what Google Gemini would do,
01:05:37 ◼ ► So you're not actually saying "show me a picture of a doctor," you're saying "giant lung preamble
01:05:45 ◼ ► - Yeah, this was covered on Stratechery, I believe, so we'll put a link to that. And I think it was a
01:05:54 ◼ ► but I don't know if that was confirmed that's what they're doing. But it's actually—you can think
01:05:56 ◼ ► of it that way. And the preamble text was like, "make sure it's racially diverse, show this,"
01:06:01 ◼ ► you know, like, it was a bunch of text that essentially asked for what you were getting.
01:06:04 ◼ ► So when you would ask for "show me a picture of the founding fathers," the actual text that
01:06:08 ◼ ► the Large Language Model would get is "show me a group of racially diverse figures, don't just show
01:06:13 ◼ ► only white men, show me the US founding fathers." And if you fed that into a Large Language Model,
01:06:19 ◼ ► you would not be shocked to get back a bunch of founding fathers who were not just white men,
01:06:24 ◼ ► because that's what you asked it for. But you can see what Google Gemini is trying to do. It's like,
01:06:36 ◼ ► or the bias of the opinions of humans. And it's not good, like, it's not—this is not like some
01:06:43 ◼ ► kind of political project or whatever. We don't want our thing to say, "show me a doctor and it
01:06:47 ◼ ► shows a white man," you know, the majority of the time. We don't think that's the right thing to do.
01:06:52 ◼ ► So how do we stop it from doing that? And the way they chose to try to stop it from doing that
01:06:56 ◼ ► essentially made it impossible for it to do useful things for anybody, because anything you tried to
01:07:02 ◼ ► do, it was shoved so hard and in such a strange, clumsy way in the direction of avoiding that,
01:07:10 ◼ ► that it just stopped being useful for anything. And I don't—like, this is embarrassing for Google,
01:07:15 ◼ ► right? And they should have tested it before they put it out, and it's a dumb way to do it.
01:07:18 ◼ ► But I think the most interesting thing about this story is everybody who's got one of these
01:07:25 ◼ ► large language model image generator type thingies is going to face this exact problem.
01:07:30 ◼ ► And everybody has so far has been trying to put Band-Aids on it somehow, saying, "the model is
01:07:36 ◼ ► trained on the data. The data is biased and bad, but the model can do useful things. So how do we
01:07:42 ◼ ► let people use the model without telling them how to build pipe bombs or saying racist things or
01:07:50 ◼ ► showing that all doctors are men and all nurses are women?" And that is a really hard problem.
01:08:06 ◼ ► head into WWC 2024 and Apple has its own large language models that is working behind the scenes,
01:08:12 ◼ ► one of the things I'm going to be looking for is how has Apple, a historically very—a company
01:08:21 ◼ ► And the final tidbit here on this Gemini story is—this is actually breaking news from, I think,
01:08:26 ◼ ► today—is that Google's Gemini now refuses to answer election questions. According to Reuters,
01:08:32 ◼ ► Google has restricted the chatbot from answering questions about the upcoming US election and
01:08:36 ◼ ► instead will direct users to Google search. I guess that's one way to do it is to look at
01:08:41 ◼ ► the prompt and say, "Hey, if you ask me anything, anything related to the election, I'm not going to
01:08:48 ◼ ► do it. I'm not going to feed it to the LLM. I'm going to say, 'Eh, no, sorry. I'm not answering
01:08:52 ◼ ► any election questions because I can't be trusted to answer election questions because I am filled
01:08:58 ◼ ► with lies and bias and terrible things, so go elsewhere.'" Of course, what you actually have
01:09:06 ◼ ► Now, right, there's all these hacks to get around, but this is a very difficult problem,
01:09:13 ◼ ► and this problem is not like, "Oh, Google's bad at their jobs." This problem is inherent in things
01:09:21 ◼ ► any smarts and are really just fancy, fuzzy, autocomplete summarization compression engines.
01:09:27 ◼ ► It's basically kind of like doing a Google search or an index lookup on the world's information,
01:09:32 ◼ ► and you're shocked when what pops out is a summary of the world's information. You're like,
01:09:35 ◼ ► "I want you to do useful things, but don't actually show me a summary of the information you have
01:09:42 ◼ ► because sometimes I think that's bad." When you're saying stuff like that, it's like, "Well, who are
01:09:46 ◼ ► you talking to?" Because what you've got is a fuzzy autocomplete summarization engine. There's
01:09:51 ◼ ► no actual intelligence or entity there for you to converse with, so maybe think of a different
01:10:02 ◼ ► inclusive, but you also want to be historically accurate. And what are you going to train if not
01:10:07 ◼ ► all the things on the Internet, which as you already covered is filth? So what are you going
01:10:12 ◼ ► to do? It's tough. But the good stuff is in there too. That's the whole thing. You want it to be
01:10:17 ◼ ► useful, right? If you ask for George Washington, you probably want George Washington. That's useful.
01:10:28 ◼ ► I don't know what they're going to do about this. I guess they can continue to try to fine-tune and
01:10:32 ◼ ► tweak it, but them saying, "Hey, we're just going to not do election stuff," that's probably also
01:10:38 ◼ ► not the answer because if you use that approach, pretty soon this thing can't be used for anything.
01:10:42 ◼ ► It's too dangerous to have election stuff, but I guess you can ask it politics questions,
01:10:47 ◼ ► and I guess you can ask it how to make a pipe bomb. It's like, "Well, soon the list of things
01:10:52 ◼ ► that it's not allowed to answer is going to be so long that no one's going to want to use it."
01:10:54 ◼ ► And that's where you have to start thinking, "Maybe the approach of training this on the
01:11:01 ◼ ► Okay, let's do some Ask ATP. Eric Neu writes, "In the PC world, very, very old programs typically
01:11:07 ◼ ► continue to run well on new versions of the OS. This does not hold up in the mobile world. Among
01:11:12 ◼ ► other things, that makes buy-once pricing much less viable on mobile, as Marco has explained
01:11:17 ◼ ► at great length. Any idea why? Is it that mobile upgrades are more likely to be unintentionally
01:11:21 ◼ ► breaking, or is it that mobile is just a much more dynamic environment with a faster pace of change,
01:11:28 ◼ ► so that either breaking is an intentional side effect, or if not breaking, then obsolete enough
01:11:33 ◼ ► that the old version is just not functionally competitive? Or is it something entirely different?
01:11:37 ◼ ► Like, if the PC world had grown up with over-the-air updates in app stores, would it behave
01:11:42 ◼ ► the same as mobile?" There's a lot to unpack here, but as a quick anecdote, Declan has been playing
01:11:49 ◼ ► a fair bit of Super Smash Bros. on the Switch recently, in between the times that he's playing
01:11:54 ◼ ► Minecraft, of course. And so he had been going and doing single-player mode in order to get new
01:12:01 ◼ ► characters that he can use in the game, and he eventually stumbled upon, like, he had to fight
01:12:06 ◼ ► Mega Man and beat Mega Man so that Mega Man would be added to his roster of playable characters.
01:12:11 ◼ ► And he's like, "Dad, who's Mega Man?" And I was like, "Oh! Oh! Just buckle up, because we got a
01:12:16 ◼ ► history lesson to do!" And of course, you know, 10 minutes later, I'm upstairs with him using,
01:12:21 ◼ ► what is the name of the emulator on the Mac? OpenEMU, which I think is not exclusive to the
01:12:26 ◼ ► Mac at all. But anyways, we were playing with the, whatever you would recommend it, I think, Marco,
01:12:31 ◼ ► the, what is this, the 8-bit Do SN30 Pro, and we're playing, you know, we connect that to the
01:12:36 ◼ ► Mac, and we're playing Mega Man, and, you know, he plays it for a few minutes first, and he says,
01:12:41 ◼ ► in so many words, "Oh my god, this is so hard." And then I would, and this was Mega Man 2, sorry,
01:12:46 ◼ ► Mega Man 2, and I was like, man, I remember it being hard, but I don't remember it being unplayable.
01:12:51 ◼ ► Like, whatever version they came out with, I think on the Wii, was straight up unplayably hard.
01:12:56 ◼ ► But anyways, Mega Man 2, I played the snot out of that game as a kid, and sure enough, you know,
01:13:03 ◼ ► he hands me the controller, and I wouldn't say I'm in, you know, 10-year-old Casey form, but
01:13:08 ◼ ► I've still got it, baby! It was quite funny, and it was a very happy dad moment watching him be
01:13:12 ◼ ► like, "Wow, dad, you're really good at this." Anyway, it just, but it did strike me as we were
01:13:17 ◼ ► doing this. You know, this is a game, I don't remember exactly when it came out, it doesn't
01:13:24 ◼ ► Declan's 9 right now, so I was roughly his age, and granted, I'm not playing on an original NES
01:13:30 ◼ ► like Marco could, and perhaps does, but I am playing the original NES game on my computer
01:13:36 ◼ ► that's, you know, less than six months old, and that's pretty freaking cool. That's pretty awesome.
01:13:44 ◼ ► I think it comes down to two major differences. So, first of all, mobile in its earlier days,
01:13:51 ◼ ► you know, say from, you know, 2006, you know, even free iPhone, like some of the early smartphones,
01:13:57 ◼ ► like, so that from that era through maybe 2015, 2016, it was just a very young industry. You know,
01:14:04 ◼ ► the personal computer and video game consoles went through a lot of the same levels of change,
01:14:11 ◼ ► and incompatibility over time in their earlier days, too. That just mostly happened before the
01:14:16 ◼ ► time we're talking about. So, while, yes, you can easily, well, not easily, but I don't know
01:14:20 ◼ ► how easily, but you can today, probably with an Intel-based PC especially, like, you can probably,
01:14:36 ◼ ► But DOS and old Windows software, like, you know, you could do that, but those were already
01:14:42 ◼ ► a pretty far way into the personal computer age. There were a lot of personal computer systems that
01:14:47 ◼ ► came before those, and even the very early versions of those, that might be harder to run today,
01:14:53 ◼ ► because enough stuff had changed. You know, what we saw in the phone era, you know, in its first
01:14:58 ◼ ► 10, 15 years, we saw it go through a huge amount of change in just the hardware, the software,
01:15:06 ◼ ► 32 to 64-bit, that was a big thing, different types of compilation, different types of distribution
01:15:18 ◼ ► these different things that all kind of came in the first five or six or seven years of the iPhone,
01:15:23 ◼ ► maybe a little more than that. So, there was all that to fight through with the mobile era.
01:15:28 ◼ ► And so, a lot of what we have a hard time using now in mobile is because there was one of those
01:15:34 ◼ ► transitions that happened that probably won't happen very soon after this now. You know,
01:15:41 ◼ ► some of those transitions are kind of one-time things. Once you make your app run, instead of
01:15:46 ◼ ► running on like one exact screen size, now it can run on flexible screen sizes. Well, that makes it
01:15:52 ◼ ► easy to make it run on any screen size in the future, because they will probably just keep
01:15:55 ◼ ► getting bigger, not smaller. Also, once you go from 32-bit to 64-bit, there's not a lot of reason,
01:16:01 ◼ ► probably within our lifetimes or at least a very long time, to go to 128-bit, for instance.
01:16:06 ◼ ► A lot of these are kind of changes that they happen more frequently when industry or platform
01:16:13 ◼ ► is young, and then over time, you're not likely to see that rate of change again, because the whole
01:16:18 ◼ ► industry matures. The second thing that's different is that the early PC and video game
01:16:25 ◼ ► software that we're talking about here was made to be self-contained. It did not communicate with
01:16:32 ◼ ► external services. It cannot have its TLS certificates expire and not be able to communicate
01:16:37 ◼ ► with modern ciphers to its servers or whatever else. It was not distributed with a bunch of
01:16:47 ◼ ► being installed on a new device very easily 10 years after it came out. It does not rely on
01:16:54 ◼ ► operating systems that themselves have code signing and locking to hardware and stuff like that.
01:16:59 ◼ ► The ecosystem that we have to provide or emulate for that old software, not only is the hardware
01:17:09 ◼ ► much simpler to emulate, but we could get there eventually with modern hardware. We can already
01:17:13 ◼ ► emulate early iPhones fairly well, performance-wise. We could get there with hardware performance,
01:17:25 ◼ ► mobile software needs to run in is so much more complicated than whatever parts of DOS we would
01:17:33 ◼ ► need to emulate, like a 386 that we would need to emulate to run some old DOS software. Or if you're
01:17:40 ◼ ► running a Nintendo game, like you were saying, Casey, it doesn't take much for modern hardware
01:17:51 ◼ ► Nintendo didn't have a BIOS to speak of, really. They didn't have an OS to speak of. Everything
01:17:59 ◼ ► was so much simpler back then. Today, all these mobile apps, we have DRM, we have code signing,
01:18:11 ◼ ► ciphers that go out of date. There are so many more factors now that make it much harder for
01:18:18 ◼ ► software to have much longevity now. Then down the road, once that software is no longer supported,
01:18:30 ◼ ► different dependencies. John, anything to add? Eric says, asking about old programs continuing
01:18:36 ◼ ► to run well on new versions of the OS, and then says, among other things, this makes the "buy once"
01:18:40 ◼ ► pricing much less viable on mobile. Buy once pricing is also not particularly viable in the PC
01:18:46 ◼ ► world. The difference is that in the PC world, because it was not entirely controlled by a
01:18:50 ◼ ► gatekeeper like the App Store, we had a thing called upgrade pricing. That was an essential
01:18:56 ◼ ► part of making a buy upfront application work. Now, you could say, okay, if you bought, back in
01:19:01 ◼ ► the old days, a version of Photoshop, it would run on new versions of the OS better than mobile ones
01:19:05 ◼ ► do for the reasons that Marco just outlined. But inevitably, eventually, that version of
01:19:09 ◼ ► Photoshop you bought would not run well on the new version of the OS. And the reason you could
01:19:13 ◼ ► pay for it once, at least back in the day, not today, obviously, was that Adobe would sell you
01:19:18 ◼ ► a new version of Photoshop at a cheaper price. That's upgrade pricing. Because you bought an
01:19:22 ◼ ► earlier version of Photoshop, you could buy a new version at a discounted price, and that would keep
01:19:26 ◼ ► you paying them money at increments when eventually, either you wanted the new features or eventually,
01:19:36 ◼ ► The other thing that I'll add is a sad fact about the desktop world today. And this is related to
01:19:41 ◼ ► another thing that Marco has talked about, like, what does it take if you have an iOS app
01:19:45 ◼ ► and you just want to keep that app running? Keep that app running on new versions of the OS. Every
01:19:51 ◼ ► year, there's probably going to be something you need to do. Compatibility with new hardware,
01:19:55 ◼ ► fixing bugs, cases you didn't think of before, working around new behaviors and old APIs that
01:20:02 ◼ ► you were using. There's some minimum amount of work you have to put into an app that is
01:20:06 ◼ ► "done." Nothing more needs to be done to the app. Why doesn't it just keep running forever? Well,
01:20:12 ◼ ► you have to do some stuff to it. And then on top of that, which we've also talked about a lot on
01:20:16 ◼ ► the show, is like, okay, but now when the new version of the OS comes out, all your customers
01:20:20 ◼ ► are going to leave one-star reviews in the app store. But it's like, why doesn't it have widgets?
01:20:24 ◼ ► Widgets just came out. I want widgets. Why doesn't your app have widgets? Your app would be great
01:20:27 ◼ ► with the widget. And you're like, so there's new features in the OS that adds that your customers
01:20:31 ◼ ► expect you to put into the app to sort of just keep up with the pace of the OS. Because it
01:20:36 ◼ ► doesn't support multitasking well, doesn't support slide-over. There's so many things that have
01:20:40 ◼ ► happened in the mobile world that you have to keep up with. Well, in the desktop world,
01:20:45 ◼ ► that is also true and used to be even more true. If you want to have a good desktop app,
01:20:51 ◼ ► you will have to update it each year to deal with bugs in the new version of macOS and API changes.
01:20:56 ◼ ► And you should update it to support widgets or whatever other new macOS features. It's just
01:21:00 ◼ ► that we've been so trained to have such low expectations of desktop software that we're
01:21:05 ◼ ► just happy if it continues to launch and run. We're not out there leaving one-star reviews
01:21:09 ◼ ► saying, why doesn't your Mac app have widgets on day one that the new version of the Mac
01:21:12 ◼ ► operating system includes widgets? It used to be when the Mac was the big show and iOS didn't exist
01:21:17 ◼ ► yet that we did have those expectations of desktop apps. And then we were excited when a new version
01:21:21 ◼ ► of the OS came out to see our desktop apps gain the new features from the new OS. And we certainly
01:21:25 ◼ ► wanted them to run bug-free on the new version of the OS and that all required work from the desktop
01:21:30 ◼ ► app developer. Now we've been so beaten down that we're just like, well, I just hope this electron
01:21:34 ◼ ► app continues to run in the new version of the operating system without crashing. Sad reality.
01:21:38 ◼ ► But yeah, I think we've covered most of the major factors that are contributing to this. But the
01:21:43 ◼ ► bottom line is that the PC world is different than the mobile world in many, many ways. Some of them
01:21:48 ◼ ► depressing, some of them good, but all of them conspire to make the experience of being a user
01:21:53 ◼ ► of software on those platforms very different. Ryan Maloney writes, would Apple be better off
01:21:58 ◼ ► giving up all but about 3% of its App Store commission and increasing its hardware prices
01:22:03 ◼ ► $50 to $100 to keep total profits the same? Is the App Store revenue uniquely corrosive or is
01:22:08 ◼ ► the problem just that Apple wants to capture as much money as possible? Porque no los dos. But
01:22:13 ◼ ► I don't think Apple would be better off doing that because, first of all, they're already expensive.
01:22:20 ◼ ► Apple already makes expensive stuff. But more importantly, the App Store is a ton of recurring
01:22:25 ◼ ► revenue. That happens constantly. And Apple wants a piece of that. Just as much as Marco or me wants
01:22:32 ◼ ► a subscription so we have recurring revenue, well, Apple wants that too. And I don't think it would
01:22:36 ◼ ► be better for Apple, no, to give up the App Store commissions. That's why they're so steadfast in it.
01:22:42 ◼ ► To a degree, I get it. I think it's excessive, but I get it to a degree. But that's why. I mean,
01:22:48 ◼ ► it's a lot of money. I think it would be good for Apple in the long term. What they would see is,
01:22:53 ◼ ► "Oh, look at all this money we're not getting." But in the long term, it's not uniquely corrosive,
01:22:57 ◼ ► but it is corrosive. We talked about this before. Getting money by taking a percentage of stuff that
01:23:03 ◼ ► other people do is corrosive in that your main motivation is to get more people to do things
01:23:11 ◼ ► that you can get a cut of money from. And if you notice a bunch of people getting money that you're
01:23:15 ◼ ► not getting a cut of, you have to figure out how to get a cut of that. And you're just trying to
01:23:19 ◼ ► encourage people to do your thing that makes money, but make sure you do it through the channels where
01:23:24 ◼ ► we get a cut. And that is not a good motivation for you doing good work. It's why in the last
01:23:30 ◼ ► episode when I was talking about Apple should try to compete with the DMA stuff, make the App Store
01:23:35 ◼ ► the best place to get apps. Compete based on how good it is to get apps from there. Compete for
01:23:40 ◼ ► developers to put their apps in there because you want to give the developers the best experience.
01:23:44 ◼ ► Compete for users to make it the safest place. Having it be by your rule the only place is the
01:23:51 ◼ ► definition of not competing. Right? That is having the motivation to satisfy your developers and your
01:24:00 ◼ ► users will make you make the App Store better. As opposed to having the App Store be the only
01:24:05 ◼ ► game in town, then the only thing you need to worry about is how do we make sure we get
01:24:10 ◼ ► a cut of everything that goes through there. Right? Keep the developers happy enough that
01:24:14 ◼ ► they don't actually leave the platform, but keep in mind that if they leave, there's no alternative
01:24:26 ◼ ► that's the right word, corrosive motivations. Because they lead Apple to do things that are
01:24:30 ◼ ► not in the interest of their users or the developers and ultimately not in the interest
01:24:36 ◼ ► of Apple in the long run because as we see them going through this EU DMA stuff, it's not making
01:24:41 ◼ ► things better for anybody really. It's a mishmash of different rules in different regions with
01:24:50 ◼ ► competition we talked about last time. If Apple really just did open it up and competed based on
01:24:53 ◼ ► the merits, that would be a virtuous cycle. They're not doing that. They're dragging their feet,
01:24:58 ◼ ► so now it's just like the worst of all possible worlds. It's not one single system that's simple
01:25:02 ◼ ► for users and also Apple continues not to be motivated to compete because they set all the
01:25:06 ◼ ► rules to make it so that the other people can't make things that are even half as good as what
01:25:10 ◼ ► they're already doing. So yeah, it's raising the prices and hardware to make up for this,
01:25:16 ◼ ► they don't need to do that. If they actually competed, like I said, they could run the app
01:25:19 ◼ ► store at break even and still make plenty of money to run their business by selling hardware. The
01:25:25 ◼ ► margins of their hardware are huge anyway. That is the virtuous cycle that led Apple to where it is.
01:25:29 ◼ ► That services revenue is the new rocket ship that they're taking, but only because they maxed out
01:25:33 ◼ ► the other one. And they maxed out the other one and became the biggest company in the world
01:25:41 ◼ ► what more do you want? We're the biggest company in the world, but we're not growing anymore,
01:25:46 ◼ ► therefore we're going to die tomorrow? No. There's plenty of other places for you to make money.
01:25:53 ◼ ► We don't know what their margins are, but on each individual product, but the idea that they need
01:26:02 ◼ ► growth at all costs is another thing that's corrosive, because once you're selling an iPhone
01:26:06 ◼ ► to everyone who can afford one, what's left to do but to sell them subscriptions? It's not great.
01:26:12 ◼ ► But yeah, I don't think they'd be better off raising their hardware prices, but I do think
01:26:16 ◼ ► they would be better off giving up on taking a cut of everyone else's money as their main growth
01:26:24 ◼ ► driver. I want to be very clear, because people often think I'm arguing something that I'm not.
01:26:31 ◼ ► I am not saying that Apple should not have a cut of any App Store stuff. The idea here that Ryan
01:26:39 ◼ ► asks of giving up all but about 3%, which would be the credit card fees of the App Store commission,
01:26:45 ◼ ► I would never say Apple should do that. That is kind of ridiculous, because Apple's payment system
01:26:51 ◼ ► is providing value. I pay 15% for my stuff because I qualify for the small business program with
01:26:57 ◼ ► Overcast. Honestly, even before the small business program, my average rate was only about 20%,
01:27:03 ◼ ► because they had already been doing the thing where subsequent years after the first year of
01:27:07 ◼ ► a subscription, you'd only pay 15% instead of 30. And I have a lot of repeat customers for
01:27:11 ◼ ► Overcast Premium, so I was already around 20% before that, now I'm at 15%. And I would say
01:27:17 ◼ ► Apple actually provides enough value that I am satisfied personally with my app for that.
01:27:24 ◼ ► The problem is not that Apple's cut is totally ridiculous in everybody's opinion, that's not true.
01:27:29 ◼ ► The problem is that Apple's cut, first of all, as John said, creates some corrosive incentives,
01:27:37 ◼ ► and that's part of Ryan's question. And it's not just that they are incentivized to squeeze
01:27:43 ◼ ► everybody as much as possible, and also they don't have to compete to make their system good and
01:27:49 ◼ ► to prove their value. But also that means that they are incentivized to maybe not necessarily promote,
01:27:59 ◼ ► but at least not discourage some pretty dark patterns in app monetization, let's say. You know,
01:28:08 ◼ ► we've joked before with the phrase "casino games for children." If you look at how a lot of the
01:28:13 ◼ ► App Store money is made, it's made in ways that I don't think Apple would be very proud to talk
01:28:19 ◼ ► about. It's made with manipulative games, it's made with scammy, you know, weekly overpriced
01:28:25 ◼ ► subscriptions that they trick people into buying, it's made with a lot of dark patterns, a lot of
01:28:30 ◼ ► addiction mechanics, a lot of psychological tricks. It's made in ways that do not fit the Apple brand.
01:28:37 ◼ ► Apple position themselves as a high-end, socially responsible, good quality brand. And if you look
01:28:44 ◼ ► at how a lot of the App Store money is made, it's not those things. Meanwhile though, because this
01:28:49 ◼ ► is such a big part of their major new growth area of services, they are continually incentivized to
01:28:55 ◼ ► keep stepping on the gas in those areas that make a bunch of money that are basically casino tricks
01:29:01 ◼ ► for children. So it is not a great way for them to make money. It does bear bad incentives like that.
01:29:09 ◼ ► And that's in addition to them not really having an incentive to compete and everything. But then
01:29:13 ◼ ► also, again, I already joked about it this episode, again, you look at the Vision Pro launch. Here
01:29:23 ◼ ► hardware business. And in the beginning of a new hardware business, there is not much sales volume,
01:29:28 ◼ ► especially something that costs a lot of money and has a fairly narrow market like the Vision Pro.
01:29:33 ◼ ► So there's not a lot of numbers there in the user base to convince developers to make software for
01:29:38 ◼ ► it. So you kind of have to rely on developers loving the platform, maybe using it themselves,
01:29:45 ◼ ► and wanting to develop for it because they love it. And you look at the Vision Pro, and I know
01:29:51 ◼ ► there's a few people out there like Casey, bless your heart, holding up the software library there.
01:29:56 ◼ ► But I'll tell you, I keep browsing the apps on my Vision Pro, and there's not much there. Even now,
01:30:02 ◼ ► like over a month after launch, and it's just a ghost town for software. And I'm seeing my own,
01:30:08 ◼ ► you know, when I look at the usage of Overcast on the Vision Pro, yes, admittedly, it's an audio
01:30:14 ◼ ► only podcast app. It's an iPad mode, so it kind of sucks. But the usage of Overcast on the Vision Pro
01:30:18 ◼ ► just keeps going down every day I look at it. And it's a pretty small number to begin with.
01:30:23 ◼ ► We're talking low hundreds of people, and that's not a substantial portion of my user base.
01:30:34 ◼ ► it's a new platform, not a lot of users, again, you're relying on enthusiasm of developers who
01:30:39 ◼ ► love it anyway, who will make apps for it because they know, everyone knows they're probably not
01:30:44 ◼ ► going to make a ton of money on it, but they do it because they love it. And a great counter example
01:30:49 ◼ ► of this is the Panic Playdate. Look at the community around the Panic Playdate. Look at
01:30:55 ◼ ► games for systems, not even systems necessarily, but games for platforms like Pico 8 that are kind
01:30:59 ◼ ► of like, they started as kind of hobbyist things or kind of, you know, fun toy things. People make
01:31:05 ◼ ► software for platforms where they don't stand to make a ton of money, but they make it anyway if
01:31:11 ◼ ► they like it or it tickles some itch they have in intellectual curiosity or they just want to play
01:31:18 ◼ ► with something. They do it for fun or they do it for the love of the platform. And you look at
01:31:23 ◼ ► something like the Playdate and I'm sure you can make more money making a game for the iPhone than
01:31:28 ◼ ► you can for the Playdate, but people make games for the Playdate anyway because they love it and
01:31:33 ◼ ► it's fun and they feel good and they just, they want to tinker or they want like a big part of a
01:31:39 ◼ ► small market, whatever the motivations are. It's a very small market that doesn't make a lot of
01:31:44 ◼ ► sense financially on paper, but they do it anyway. And when Apple launches a new platform like Vision
01:31:49 ◼ ► Pro, that's what they're relying on. They're relying on that kind of developer interest to
01:31:55 ◼ ► get that software library started and then maybe down the road they might build up towards a decent
01:32:00 ◼ ► amount of sales volume for the product and then the numbers can start justifying themselves for
01:32:05 ◼ ► people who take a more, you know, numbers-based approach to the question of whether they should
01:32:08 ◼ ► develop for it. But they're not there yet and to get from where they are now to that point,
01:32:14 ◼ ► they need those enthusiastic developers who just love it and want to develop for it because they
01:32:18 ◼ ► love it and or because they want to play around or experiment, not because they're going to make
01:32:22 ◼ ► money on paper. And what Apple has done with these fairly cynical, you know, developer policies and
01:32:28 ◼ ► treatment over the years, what they've done is eroded all of that attitude that people have of
01:32:34 ◼ ► this is a fun new thing. Now, yes, there are people who feel that way about Vision Pro,
01:32:38 ◼ ► but there's not a lot of them anymore. That community of who would do that is much smaller
01:32:43 ◼ ► than it used to be now. A lot of that is due to the attitude Apple has had towards developers
01:32:48 ◼ ► and a lot of that is due to their addiction to the services revenue cut. So when you look at things
01:32:55 ◼ ► in this very small picture of should Apple, getting back to Ryan's question, should Apple
01:33:00 ◼ ► give up their cut because it's too high or whatever, the answer is I don't think so. No,
01:33:05 ◼ ► I think their cut, especially if you're getting the 15% for most or all of your income, I think
01:33:11 ◼ ► their cut is fairly reasonable for what it is. But the thing is, we don't know their cut is
01:33:15 ◼ ► reasonable until we see someone competing with them because like what the market would bear
01:33:19 ◼ ► would be okay, well, there's another store and here's the cut that they take and here's the
01:33:22 ◼ ► services they provide. And I think you would need that competition to sit for the water to find its
01:33:27 ◼ ► level for actually competing for the for the the affections and for the software of developers.
01:33:33 ◼ ► I don't know what that number is, but I bet it's not what Apple is currently setting because there's
01:33:37 ◼ ► no competition now. And yes, Apple has lowered it. But I have to think that the number they've
01:33:41 ◼ ► currently lowered it to is not the quote unquote market value because there is no market the App
01:33:50 ◼ ► perfectly commoditized. I don't think it needs to be like bringing a bunch of people and whoever
01:33:55 ◼ ► can offer the lowest price wins. No, it's not not commoditized in that way. Because you're still
01:33:59 ◼ ► competing based on like features and user experience and like all the things that Apple
01:34:03 ◼ ► would excel at many of these. I'm not saying that they're they wouldn't be better. But like with
01:34:07 ◼ ► the total lack of competition, Apple is allowed to continue to mistreat developers and neglect parts
01:34:12 ◼ ► of its platform. Because like where are you going to go? You have no choice. Your choices either
01:34:15 ◼ ► you're on the platform or you're not. And third party app stores with a reasonable with a reasonable
01:34:20 ◼ ► way to compete without rules that essentially make them not able to be any better than the App Store,
01:34:25 ◼ ► which I think they're still trying to do with the DMA compliance that would help them to find the
01:34:29 ◼ ► level and that would, to my point earlier, would help Apple realign incentives to stop worrying
01:34:34 ◼ ► about how they're how they're going to get as much cut as they can without pissing people off too much
01:34:38 ◼ ► and start saying how do we make our App Store the place where developers want to be? Because if they
01:34:44 ◼ ► make it way better than everyone else, they can charge a higher cut. Like it's not like they have
01:34:49 ◼ ► it's not a race to the bottom. They can get a higher cut if they provide a better experience.
01:34:59 ◼ ► actual percentage and the idea of people coming in and making alternatives. I don't care so much
01:35:05 ◼ ► about that. I care about the behaviors that Apple is incentivized and enabled to do with their
01:35:12 ◼ ► current system of them being the only game in town by force. And then also what that does to the
01:35:18 ◼ ► ecosystem in general, really souring a lot of developers on development for the platform. And
01:35:22 ◼ ► again, back to the Vision Pro problem that they have, I as a user and as a lover of Apple products,
01:35:30 ◼ ► I want their new platforms to succeed. I want there to be a large amount of great software
01:35:36 ◼ ► on the new platforms that they launch because I want to use them. I want to enjoy them. That's
01:35:41 ◼ ► part of being a fan of this stuff. I would love to see like in the tech business we have hardware
01:35:48 ◼ ► coming at our asses. We have such a just oversupply of cool hardware. The scarce resource for most
01:35:54 ◼ ► platforms in tech is great software. We have way less of that than cool hardware. Cool hardware
01:36:02 ◼ ► comes out great. That's nice. We have a lot of that. We talk about it. It's fine. But great
01:36:07 ◼ ► software is what actually really makes a difference. So any kind of practice or policy or
01:36:20 ◼ ► I think is a massive strategic problem for a company. That's where I keep criticizing Apple
01:36:24 ◼ ► for their app store mishandling and the way they keep trashing developers very publicly and souring
01:36:32 ◼ ► so many people on them and developing for their platforms. The reason I criticize this is because
01:36:37 ◼ ► I love Apple products and I want them to succeed in their software ecosystems because that matters
01:36:43 ◼ ► so so much. And I just keep seeing over and over again Apple doing these own goals that try to
01:36:50 ◼ ► maybe save a relatively small percentage of their income over here at the expense of the success of
01:36:58 ◼ ► their very profitable hardware platforms and the software ecosystems that develop for them which
01:37:03 ◼ ► matters so much more to the company. Thank you to our sponsor this week Magic Lasso Adblock and
01:37:09 ◼ ► thank you to our members who support us directly. You can join us at atp.fm/join and one of the new
01:37:14 ◼ ► benefits members now get is the ATP overtime. This is an extra topic that we're going to be talking
01:37:19 ◼ ► about just for members after the after show. This week it's going to be the Rabbit R1. We didn't get
01:37:25 ◼ ► a chance to talk about it in the main show because there was just too much news during those weeks.
01:37:54 ◼ ► because it was accidental. It was accidental. And you can find the show notes at atp.fm.
01:38:05 ◼ ► And if you're into Twitter you can follow them at c-a-s-e-y-l-i-s-s. So that's Casey List m-a-r-c-o-a-r-m
01:38:19 ◼ ► anti-marco-r-m-n-s-i-r-a-c-u-s-a-c-r-a-q-s-a. It's accidental. They didn't mean to. Accidental. Tech podcast so long.
01:38:52 ◼ ► There you go. So you've had six months with it and despite my best efforts it's still working.
01:38:56 ◼ ► But given that you've had it for six months I guess it's time to get a new one right? So
01:39:01 ◼ ► getting an R2 or an R3 or an R3X? Oh man yes so Rivian this past week they had a live event
01:39:13 ◼ ► By the way on this numbering are they the only car company that makes the number bigger where
01:39:18 ◼ ► the car gets smaller? So BMW the bigger number the bigger the car. 3 Series is smaller than the 5 is
01:39:24 ◼ ► smaller than the 7. Mercedes has the the the letters I guess but then just the engine designation.
01:39:30 ◼ ► Audi A4 is smaller than the A6 is smaller than the A8. But in Rivian the R1 not the Rabbit R1.
01:39:36 ◼ ► The R1 is the biggest and like if they ever make anything bigger than the R1 is it the R0? Like I
01:39:41 ◼ ► think they might have painted themselves into a corner here. They can go negative. It's like
01:39:45 ◼ ► how Canon names their big cameras like you know like that like the 1D was the really big one.
01:39:50 ◼ ► Yeah. And then you know the 5D was a little bit smaller and then like you know the the 50D you
01:39:53 ◼ ► know like they they would make them they would make the number bigger as the camera got smaller.
01:39:56 ◼ ► Yeah Evolvo also XC30 is smaller than the XC90. I don't know I'm not telling Rivian how to do
01:40:02 ◼ ► their naming I just it's just weird to me to see that the R3 is smaller than the R2 is smaller than
01:40:06 ◼ ► the R1. Yeah anyway basically this is like their you know Model S to Model 3 kind of moment. The
01:40:12 ◼ ► Rivian R1 series is like larger more expensive bigger batteries and then the the R2 platform
01:40:19 ◼ ► XDR3 is also built on is their like smaller more affordable version that they will probably sell
01:40:25 ◼ ► in much higher volume and it's not out yet but it will be coming out 2026 they said. Yeah this is
01:40:30 ◼ ► quite a pre-announcement. This is like a Tesla style pre-announcement. Look at these new models
01:40:34 ◼ ► you can pre-order one now when am I get it 2026 how do you feel about that? Normal car companies
01:40:41 ◼ ► usually don't announce cars more than one model year in advance but Rivian is doing that now
01:40:47 ◼ ► because I kind of feel like Rivian wants people to know we're not dead we're still making good
01:40:52 ◼ ► things and I think that's good because I think these two cars do get people excited about the
01:40:57 ◼ ► brand but who's excited to wait two years for your pre-order? Honestly I think what they announced
01:41:02 ◼ ► was pretty exciting I don't know how long I mean if you look at like the Rivian's own models if you
01:41:06 ◼ ► look at Tesla's models like I think waiting about one to two years for a brand new model if you
01:41:12 ◼ ► place a pre-order I think that's fairly common. I am excited to see this because first of all I
01:41:16 ◼ ► don't think anybody assumes they're dead I think I think the R1 S is I don't know how many T's are
01:41:21 ◼ ► selling but the R1 S I think is selling extremely well I think they're doing just fine. It's not
01:41:26 ◼ ► because of their cars aren't desirable it's because the company does not make money yet. Well right.
01:41:31 ◼ ► And they're burning through their remaining cash everyone's looking at their bank account going
01:41:35 ◼ ► we love your cars you're selling them as many as you can make but you're burning through cash and
01:41:39 ◼ ► that's what people are worried about. Yeah fair enough but anyway so I so what they announced I
01:41:45 ◼ ► think looks really good I mean I'm not sure that it's what I'm going to buy in the future but at
01:41:50 ◼ ► some point I'm going to replace the R1 S that I have and love I don't know when I'm not in any
01:41:55 ◼ ► rush to replace it but when I do need to replace it I'm going to take a really serious look at I
01:42:00 ◼ ► think maybe the R3 or the R3 X depending on you know how those end up I think these look you know
01:42:08 ◼ ► so the R2 looks like the R1 but smaller it is still a big rectangle you know it's it's a
01:42:14 ◼ ► utilitarian mid-sized SUV and again I love my R1 S like I so I I think they're gonna I think the R2
01:42:22 ◼ ► being a slightly smaller significantly cheaper R1 S basically I think they're gonna sell a ton of
01:42:28 ◼ ► them and and they're gonna do very much I assume they're gonna do very very well if they survive
01:42:32 ◼ ► until then if they survive long enough to ship this thing I think they're gonna do very very
01:42:36 ◼ ► well. Yeah because the R1 S is big like it's not problematically big but it's a big car. I think
01:42:43 ◼ ► it's almost problematically but especially for Marco who let's be fair his family of three people
01:42:48 ◼ ► and three people are not big they're you're small people and the thing is the place where you live
01:42:54 ◼ ► has very very narrow quote-unquote roads. Oh gosh yeah we talked about this I almost died when I was
01:43:01 ◼ ► in the passenger seat watching Marco navigate on Fire Island. I know and so like it's like you it's
01:43:06 ◼ ► not like you need that space because you're three tiny little people rattling around the giant
01:43:10 ◼ ► interior and you have really narrow roads so I feel like the R2 at the very least you should
01:43:14 ◼ ► consider you know not there's anything wrong with your own right now the only thing I worry about is
01:43:19 ◼ ► off-road capabilities because I feel like the R1 may be more off-road capable than these little
01:43:25 ◼ ► ones certainly it has more ground clearance so you have to check into that but like I feel like the
01:43:29 ◼ ► width is not helping you on that R1. The R1 it is a large vehicle for sure but when you look at what
01:43:36 ◼ ► other people on the Beaks Drive there's a lot of like the like the even larger like Chevy whatever
01:43:40 ◼ ► SUVs and the Ford whatever SUVs. Are they clearing the path for you by hitting all the branches?
01:43:45 ◼ ► Yeah you go through they're all knocked down. You know I'm used to it now I would love a smaller
01:43:50 ◼ ► vehicle for maneuverability on those streets but ultimately I'm used to this and it is far from the
01:43:54 ◼ ► largest vehicle there. Everyone's driving these giant pickup trucks they're driving like the
01:43:58 ◼ ► Expeditions and whatever the Yukons all these other things that are that all seem larger to me
01:44:02 ◼ ► I don't know if they actually are if I haven't looked it up but I don't think it's that crazy
01:44:06 ◼ ► big. Anyway anyway yes Marco could do with a smaller car but that's not why Marco's looking at
01:44:10 ◼ ► this you know why Marco's excited about the R3 because the back of it is not a right angle and
01:44:14 ◼ ► his little R1 mind is exploding with the possibilities of that styling choice. I think it
01:44:19 ◼ ► looks okay so yeah so the again the R2 is basically a smaller box the R3 takes the R2 platform and
01:44:26 ◼ ► makes it look more like a giant hatchback. By making one change which is the back windshield
01:44:31 ◼ ► is at 45 degrees instead of 90. I think it looks really cool I mean it's going to be a long time
01:44:39 ◼ ► coming out in 2026 they gave no date for the R3 they just said it's going to be after the R2
01:44:45 ◼ ► but it looks fun it's still going to be like you know SUV or I think they described it as a
01:44:50 ◼ ► crossover so it's going to be it's going to be a large vehicle but it looks kind of like those old
01:44:55 ◼ ► like 80s little like like a Volkswagen Rabbit somebody said that yeah that we're in only
01:45:00 ◼ ► massively larger yeah yeah it's quite a bit bigger it's the size of it's a size like an IONIQ 5 like
01:45:06 ◼ ► that's the new size for like or the Honda CR-V the new what would previously in previous generations
01:45:11 ◼ ► be a massive car these are now the small SUVs and when you rake the back windshield it's like it
01:45:15 ◼ ► looks like a Rabbit only twice as big yeah and honestly I think that's a fun look for you know
01:45:20 ◼ ► because if you look at like modern you know modern American especially sensibilities for cars I loved
01:45:27 ◼ ► the Honda E that they launched in Europe and I think just discontinued that actually is significantly
01:45:32 ◼ ► smaller than this I think yeah oh yes oh definitely that's why they could it's so small they couldn't
01:45:36 ◼ ► even sell it in the U.S. that was the problem like I thought the Honda E looked so cool it looked
01:45:41 ◼ ► like a cool like you know 70s 80s throwback but new and electric and modern inside yeah agreed
01:45:45 ◼ ► I have like an 80 mile range though so yeah and but yeah it was too small for the American market
01:45:49 ◼ ► so they didn't even they didn't even launch it here and I thought that was such a shame I think
01:45:53 ◼ ► the R3 is probably as close as you can get to that and still have a chance of selling it in the
01:45:59 ◼ ► American market which is not even close size wise well I mean Casey's driving one he's driving a
01:46:04 ◼ ► Volkswagen Golf which is you know it's not as small as a Rabbit back in the day but it's still
01:46:09 ◼ ► significantly smaller than the R3 yeah I would assume I haven't looked at numbers but I believe
01:46:13 ◼ ► that to be correct and yeah I mean I think you could make a strong argument that that you start
01:46:19 ◼ ► from the Rabbit and you know you diverge one one path of the family tree goes to my car and another
01:46:25 ◼ ► path of family tree ends up in an R3 I actually it's despite what you would think I don't have
01:46:30 ◼ ► any specific love for a hatchback like I think it's fine if I could have gotten a sedan version
01:46:38 ◼ ► but they definitely don't make that car in the US definitely don't make that by the way I disagree
01:46:42 ◼ ► strongly now that I've had hatchback or lift backs like the Model S I'll never go back to a sedan
01:46:48 ◼ ► style again unless it has a lift back you know like like the Model S like no more like regular
01:46:52 ◼ ► trunk for me this is what Europe figured out forever ago and we're still trying to learn
01:46:56 ◼ ► well I mean Europe Europe did it their cars were actually small and they because they have
01:47:00 ◼ ► even narrower roads well I want to say narrower roads than Marco but narrower normal roads than
01:47:04 ◼ ► Marco does like far on doesn't have normal roads and you know short cars getting into parking spots
01:47:10 ◼ ► but there's the you know those pictures you can find on the internet of the original Mini versus
01:47:13 ◼ ► the current Mini and they have the same problem they have car inflation car inflation is everywhere
01:47:19 ◼ ► oh yeah yeah because Minis are now basically little SUVs like they're not really Mini it's
01:47:24 ◼ ► comical when you see the original one next to it you either think that you really think it's like
01:47:28 ◼ ► one of those power wheels toys for for kids or you think the new one is somehow massively inflated
01:47:33 ◼ ► with photoshop but no this is a real picture I wonder if I could somehow import a Honda e
01:47:37 ◼ ► can how how hard is it to import a car that's not made for your market I think it's pretty hard
01:47:41 ◼ ► right it's easy if you're if you're willing to wait another 23 or 24 years because I think it's
01:47:45 ◼ ► right five years then then it becomes easy you should concentrate on on building up your stock
01:47:51 ◼ ► of backup i3s probably yeah it's probably easier with the put them with the keyboards no I think
01:47:57 ◼ ► the r3 looks great there's an r3x which is apparently a performance version which I haven't
01:48:01 ◼ ► really looked much into but you say performance and I'm paying attention I think this looks
01:48:06 ◼ ► awesome I really admire Rivian a lot I think they're they're taking a lot of the path that
01:48:15 ◼ ► Tesla had trailblazed but learning from it and doing it better from what I can tell and more
01:48:21 ◼ ► maturely from what I can tell I'm still grumbly that carplay is not a thing and doesn't appear
01:48:28 ◼ ► that it ever will be and I will forever be grumbly about that but leaving that aside these these look
01:48:33 ◼ ► really good I mean I don't think that I would want an r2 only because I really like having the third
01:48:39 ◼ ► row in Aaron's car we don't need it often but there are times we kind of need it I guess need
01:48:44 ◼ ► is a strong word but would really really really want it and so an r2 is only two row only two row
01:48:51 ◼ ► which is a bit of a bummer and an r1s is extremely expensive so just so expensive that it's not of
01:48:56 ◼ ► our price range but I mean I I really dig these in principle and I think if we were to look at a car
01:49:04 ◼ ► for Aaron tomorrow we would probably end up in just a brand new or you know lightly used xc90
01:49:09 ◼ ► but I would try to hold on for the ex90 which is the electric version of Aaron's car that's coming
01:49:15 ◼ ► in a year or two I think as well I would at least try to find an r1s that was maybe used enough that
01:49:21 ◼ ► it made it affordable for the list family but I don't know there's there's a lot of great options
01:49:27 ◼ ► here and I'm really pleased the Rivian seems to be you know they're not resting on their laurels
01:49:31 ◼ ► they're making forward progress oh this these have nacs instead of ccs and they're in the back
01:49:36 ◼ ► instead of the front yeah they move the charge port and it's much smaller now and it doesn't
01:49:40 ◼ ► have the big annoying swing up door one of the things I think they're worried about with these
01:49:43 ◼ ► ones though is because they're announcing these two years ahead of time but the competition has
01:49:48 ◼ ► similar size similar market things like the Volvo I think the Volvo has their electric whatever their
01:49:53 ◼ ► 30 series xc30 or ex30 or whatever it is but that you can like buy that now or very soon and it is
01:50:00 ◼ ► basically the direct competition for the r3 and the r3 doesn't have a date or a price and so I
01:50:06 ◼ ► think Rivian is a little bit behind the market with these like that their competitors have cars in
01:50:12 ◼ ► these same segments at same price points that are also evs from reputable brands that are doing
01:50:16 ◼ ► pretty well now I think Rivian has the edge because I just think Rivian is doing a better
01:50:20 ◼ ► job of evs than for example Volvo or Polestar but it's but it's close depending on your tastes
01:50:26 ◼ ► and to have to wait two years while your competitors essentially gain a foothold in this
01:50:29 ◼ ► market of 50 grand ish evs with decent range and small SUV size has got to you know be difficult
01:50:37 ◼ ► for Rivian to see but that's just you know you can't they can't do everything at once and
01:50:41 ◼ ► it takes a long time to make a new car so it's I think it is important for them to get these
01:50:47 ◼ ► cars out ASAP they even said as part of the announcement like we're going to release them
01:50:50 ◼ ► earlier than we even thought we would it's because I think they're feeling the pressure of like get
01:50:54 ◼ ► these things out the door because to Marco's point earlier yeah the big r1 is great but those cars