The Talk Show

403: ‘150 Million Calculator Apps’, With Quinn Nelson


00:00:00   this better be worth it. Sure enough. It was really cool. It

00:00:04   was super well done. The live stream went off without a hitch.

00:00:07   There was no weird quality issues or any major frame skips.

00:00:11   And it really did feel like you were in the audience, which was

00:00:14   pretty awesome.

00:00:15   The funniest thing was that Adam was just off stage. All he had to

00:00:21   do is turn his head and he could have watched the whole show from

00:00:24   right where he was behind the curtain and he watched the whole

00:00:26   thing in the vision.

00:00:27   In the vision pro?

00:00:28   Because he was like, I just wanted to see that it worked.

00:00:32   And then he was like, well, and then I got into it. And I was

00:00:35   like, well, I'll just keep watching. So he was literally

00:00:38   right off the stage, watching Envision Pro. So I, I have to

00:00:42   say I'm pretty I mean, again, I cannot take credit for any of

00:00:45   the technology. It was all Adam and Spatial Gen and his

00:00:48   colleagues at sandwich. And all I did was say, Okay, let's try

00:00:52   it. You know, all I did was grant permission. So I don't

00:00:54   want credit for it. But I have to say it is a minor thrill that

00:00:59   my show to my knowledge is the first live thing ever broadcast

00:01:05   for vision Pro.

00:01:06   I think that's right. Yeah, that's cool.

00:01:09   And it's kind of weird. Oh, well, you'd think Apple would

00:01:12   do it first. But Apple doesn't really do anything live anymore.

00:01:16   Right? Their keynotes ever since COVID are all pre recorded. And

00:01:19   even like at WWDC each year, when they show it to 1000 or 2000

00:01:27   people on the lawn outside Apple Park and Tim and Craig

00:01:31   Federighi come out beforehand, they don't broadcast the little

00:01:34   three minute introduction. So I don't even know what Apple would

00:01:38   do live other than sports, of course, which sports fans are

00:01:42   all waiting for, but which is a whole nother huge, humongous

00:01:47   can of worms and certainly isn't going to be the first thing that

00:01:50   happens something like complexity. Right, something

00:01:55   like my show, here's a stage you can just point static, right at

00:01:59   it. And yeah, static subjects is kind of perfect for it, even if

00:02:03   not exciting. I thought it was great. It was really fun. So I

00:02:08   just want to say again, what a thrill. I'm glad it worked very

00:02:12   well by thanks to sandwich and everybody and it's I guess we'll

00:02:16   do it again next year. And the other thing I was braced for were

00:02:20   complaints from people who don't have vision Pro Well, how come

00:02:23   you don't just broadcast do the live thing flat too so I can

00:02:27   watch and nobody complained. Which again, I'm like, are they

00:02:33   in my spam? I'm like looking at my spam. It's like, nope,

00:02:36   everybody just took it in stride. And I was like, pretty

00:02:39   cool. Thank you for everybody for not complaining. I guess we

00:02:43   should start with Apple intelligence. I wrote about it

00:02:47   finally over the weekend or finally finished it up. You know,

00:02:50   some years WWDC has a Oh, that was the year blank. And now

00:02:56   we've got two years in a row last year was the vision Pro got

00:02:59   introduced this year obviously will be remembered as Oh, this

00:03:02   is the year that Apple intelligence was announced. It

00:03:06   does seem like a big deal. And I worry that I buried it by

00:03:09   putting it the conclusion of my article because I do think it's

00:03:13   like the single most important thing people can take away. But

00:03:16   and it felt like it belonged at the end. But I always worry about

00:03:19   putting things at the end. Maybe you feel the same way with

00:03:21   videos to where it's I worry about the people who don't get

00:03:23   to the end. I'd like to think that everybody who reads my

00:03:27   articles reads them to the end because they're so compelling.

00:03:30   But I know that people start reading and then they drift

00:03:32   away. But the main takeaway I want to emphasize is we don't

00:03:36   know if any of this stuff works. We don't know even if it does

00:03:40   work if it's going to scale to where Apple needs it to scale. I

00:03:44   don't know that even true experts in setting up data

00:03:47   centers ever really know if something's going to scale until

00:03:50   you flip the switch and go live. And it just seems like if it

00:03:55   wasn't this huge moment of pressure from users who seem to

00:04:01   want these features and Wall Street investors who definitely

00:04:04   seem to want Apple to to have something to say on generative

00:04:08   AI, I honestly think almost all of this would have come next

00:04:13   year, not this year. I feel like we're finally get here's what

00:04:16   Apple would do if they told us what they're doing a year in

00:04:18   advance.

00:04:19   I tend to agree if anything, because all the features that

00:04:22   are the most compelling features don't even have a date or or

00:04:27   have been suggested to come sometime next year into the

00:04:30   future. I think they did a pretty good job at explaining

00:04:34   what Apple intelligence is supposed to be and what it will

00:04:37   be. That's part of the problem, right? This is such a new thing

00:04:41   to so many people, that not only do they have to explain what it

00:04:45   is, and how it works, and why it's different from other

00:04:49   generative AI stuff that exists on the market today, but then

00:04:52   they have to lean into the okay, now that we have the foundations

00:04:55   of what it is, what does it actually do? And I felt a little

00:05:00   bit, I don't want to say disappointed. But that's where I

00:05:04   feel we didn't get a ton of information about what it will

00:05:08   do beyond the basic stuff that was expected, like summaries and

00:05:13   rewriting tools and generative images, which I didn't really

00:05:17   expect, but it's not a novel or new thing. That's where I think

00:05:21   it's, it's difficult and maybe weird is they came out guns

00:05:25   blazing with here's what it is. And here's how it works. But we

00:05:29   still are kind of waiting for the okay, but what, how will it

00:05:32   help me? What is it supposed to do into the future? We just have

00:05:36   the little kind of principal foundations that are good. But I

00:05:40   don't know that any of them are all that unique as compared to

00:05:43   what already exists today on the market.

00:05:45   And it's it was very telling to me not in a surprising way. But

00:05:51   in a Oh, this is very early days. And they are really giving

00:05:55   us a sneak peek behind the curtain that at Apple Park, for

00:06:01   those of us who got media invitations, they had briefings

00:06:04   after the keynote Monday afternoon. And then back on

00:06:07   Tuesday, I was back there all all morning Tuesday for demos of

00:06:12   Mac OS and iOS and iPad OS and stuff like that. And none of

00:06:18   those demos, not only are those features not in the developer

00:06:21   beta that shipped last week, and not coming, I don't know, maybe

00:06:25   even while we're recording this podcast today, supposedly, today

00:06:29   is the day second developer betas are dropping. But the

00:06:32   Apple intelligence features aren't coming yet. But they,

00:06:35   they didn't even let us play with it hands on, which is pretty

00:06:39   unusual for Apple. But again, in this case, it's not surprising.

00:06:43   So I got to see Apple representative in a briefing,

00:06:49   writing an email, just exactly like the keynote, here's an

00:06:52   email for a job, I'm applying for a job, and I've attached my

00:06:56   resume. And I've written the email, select the text of the

00:07:00   email and hit this little blue thing and say, make this sound

00:07:03   more professional. And I got, we got to watch that happen. And I

00:07:07   asked, is this determinant? I know you're doing these demos

00:07:11   for these are small groups of media three or four at a time.

00:07:14   And then the same people every 40 minutes are doing it for

00:07:17   another group. I was like, do you get the same answer every

00:07:20   time? And they said no, we there are slight differences. And

00:07:24   because they've been rehearsing the demo and doing it for

00:07:26   groups, he even told me, here's, here's the sentence

00:07:30   where sometimes we get slightly different words when we say make

00:07:33   it more professional. That's how LLM's work, right? That's

00:07:36   the proof that it was a live demo of what they actually have

00:07:40   working behind the scenes. But it also also shows why they

00:07:43   didn't let us type our own email and, and say make this

00:07:48   funnier or make it whatever the options are for that. We didn't

00:07:53   get to make our own Genmoji. We didn't get to they didn't even

00:07:57   show the image playgrounds in action. So I mean, there's a lot

00:08:00   of stuff that they didn't even demo to us live. I mean, that's

00:08:04   how early it is. The Xcode stuff they definitely showed live. And

00:08:08   it seems again, I keep mentioning this on dithering to

00:08:13   it just is wild to me, but I get it. But it just seems crazy to

00:08:17   me that LLM's are particularly good at writing computer

00:08:21   programs. Whereas for humans, writing computer programs is

00:08:27   generally considered a much rarer skill and harder than

00:08:31   writing prose, right? It's just the nature of LLM's and the

00:08:35   nature of code that it works with the robotic brain for lack

00:08:39   of a better and they showed us that live demo of the code

00:08:43   completion and Xcode and the much more impressive Swift

00:08:47   assist where you can sort of just type a natural language

00:08:51   query, like I've got a collection of images, I'd like

00:08:55   to present them in a carousel that users can swipe

00:08:59   horizontally. Boom. And then here's some code that would say,

00:09:03   okay, given a collection of still images, here's how you

00:09:06   here's some code that would generate a Swift UI view of a

00:09:10   swipeable carousel of those images. Very, very impressive.

00:09:14   But again, they did not let us I don't know how much help it

00:09:17   would be letting the media write code, but we could write

00:09:21   natural language queries, though, right? I mean, if

00:09:24   anything, that would actually be a demo we could have done, but

00:09:27   very, very early days. And it's kind of exciting, right? To be

00:09:32   following Apple and to actually see them reveal something so

00:09:36   early.

00:09:36   Yeah, it's very atypical. On the one hand, I agree. It's it's

00:09:41   definitely for optics, they need to let investors know that it's

00:09:45   something that they've been working on and something that's

00:09:47   planned. I think their focus on running many of the models that

00:09:52   they have on device is proof, I guess that the NPUs and the

00:09:59   neural engine as it as Apple calls them as that they've been

00:10:02   putting in basically all of their a series chips for several

00:10:05   years now. And in the M series chips as well, that that there's

00:10:09   a point to that there's been a point in the past. I mean, NPS

00:10:13   have been doing all sorts of stuff on Apple Silicon for

00:10:16   years, but nothing so definitive as in, this is why these chips

00:10:20   keep getting larger and larger and more powerful, especially in

00:10:24   a market where you know, with something like the Snapdragon X

00:10:27   elite, the new, you know, alleged competitor to the M

00:10:30   series chips, they they keep talking ad nauseum about how

00:10:35   amazing their NPU is and why neural engines and application

00:10:38   specific silicon is going to be super important into the future.

00:10:41   And Apple's kind of just Yeah, we've been doing this for a

00:10:45   decade now. And this is the proof that it matters. And, and

00:10:48   being able to do a lot of the stuff that right now with other

00:10:53   AI models is done in the cloud on device is a really big deal

00:10:57   being able to write, rewrite or to proofread or to make your

00:11:01   email Sam more friendly when you're on an airplane with no

00:11:05   internet connection. It's it's going to be faster. It's going

00:11:09   to be there's no downsides really, it's pretty amazing.

00:11:12   Again, not surprising, it almost would have it would have been far

00:11:16   more surprising if it wasn't in a personal context, right?

00:11:21   Because it really does fit with Apple's long standing under Tim

00:11:27   Cook sort of doctrine, as Horace Dedue has put it that the Apple

00:11:32   seeks to own and control the core technologies, most

00:11:36   important to its products, blah, blah, blah. But there's also, I

00:11:41   don't know, a couple years ago, there was sort of a mantra, only

00:11:44   Apple things only Apple could do and that that semantic index of

00:11:50   your personal information put into a stored securely on

00:11:56   device, cryptographically, and the key demo of the whole keynote

00:12:01   was that when does my mom's flight arrive? Question, right?

00:12:06   And presuming that it worked that you'll be able to do

00:12:09   something like that, and that it'll work even close to as

00:12:13   well as Apple is suggesting it works, is, yes, that's what

00:12:18   we've been thinking in the back of our heads ever since the

00:12:22   first version of Siri, right, that you'd just be able to ask

00:12:26   a question like that and get it. But it, it really is the nature

00:12:31   of the question. You don't have to get into one of these

00:12:35   comparisons. I know with Anthropic, and they just

00:12:39   announced the Claude 3.5. So nada, whatever their new one is,

00:12:44   and they've got these benchmarks, and they show these

00:12:46   comparisons to chat GPT four o and perplexity and whatever else

00:12:51   the leading Gemini whatever the leading other ones are, and they

00:12:54   come up with scores like 64 to 65% on this issue or 83 versus

00:13:01   79% on something else. Put aside where Apple's models compare in

00:13:08   a benchmark like that against these rivals, only apples is

00:13:12   going to be able to get the personal context of that right,

00:13:14   you could go we could go to the future and chat GPT five isn't

00:13:20   going to be able to answer the question when is my mom's flight

00:13:23   arriving because it doesn't even know who you are, let alone your

00:13:26   mom or have access to your email. So the focus on stuff that

00:13:31   only Apple can do that because it's stuff that's only on your

00:13:35   device and that you trust to be on your device is potentially I

00:13:42   know and it sounds like a bold adjective but radical, right?

00:13:45   This could change the way we use our devices.

00:13:48   And and that's what I think is so compelling about Apple

00:13:52   intelligence. Most of the stuff that they showed this year is

00:13:57   stuff that no doubt will be helpful. I think that rewriting

00:14:01   tools are going to be used by tons of people, I think image

00:14:04   generation for emojis and memes will definitely be a thing.

00:14:09   These are all things that are coming and they're cool. But I

00:14:13   don't know that we'll look back at tools like this in five years

00:14:16   and go remember when that changed the world. But I I do

00:14:20   think that reasoning as it's kind of referred to in in ml and

00:14:25   AI nomenclature, which I don't really like because it's such a

00:14:28   wide encompassing term. So I invented one that's not really a

00:14:31   real thing. But like this idea of an intent engine, something

00:14:35   that understands the input that you as a user are asking it,

00:14:39   knows how to find the data related to that and then provide

00:14:43   a result that you expect. That's the promise that so many of

00:14:48   these kind of AI specific hardware devices have sold

00:14:52   themselves on. But nobody's come even close. I think like the

00:14:56   humane AI pin is a good example. I reviewed that thing. It's it's

00:15:00   not a great product. I didn't recommend anybody buy it. But

00:15:03   one of the things that did do that was quite cool is there was

00:15:09   one time where I was standing in front of a bus stop. And I said,

00:15:13   When is the next bus supposed to arrive? It took a picture of the

00:15:17   sign. And I guess via searching the web told me that the next

00:15:22   bus was supposed to arrive in eight minutes. And it was off by

00:15:25   a little bit, but that could have been the buses for all I

00:15:27   know. And then I at the same time after said, Let my wife

00:15:32   know when I will be arriving home. And it used its reasoning

00:15:37   to find when the bus would get to my house. I don't know if

00:15:40   that was done via traffic data or whatever. But it was really

00:15:44   cool because I was able to just say a thing and in the

00:15:47   background, it was able to go okay, he needs to get from point

00:15:50   A to point B, he's standing in front of a bus stop. So he's

00:15:52   going to take the bus. Who is his wife? When he says let her

00:15:56   know that's probably a message. How should we do that? How do we

00:15:59   craft the message? And it was one of the few instances that I

00:16:02   use the AI pin where I was like, this actually could be a game

00:16:06   changer. And I think that is what the promise of Apple

00:16:10   intelligence is.

00:16:11   John Greenewald In your case, and that is a great demo. I

00:16:14   mean, just a great example. Did it have to intuit that you would

00:16:19   be riding the bus that you just asked about? Like you didn't

00:16:22   say, I'm going to get on the bus.

00:16:25   Yeah, I actually said and you have to use the look command.

00:16:28   But I said, look at where I am, and tell me when the next bus is

00:16:32   coming. So it took a picture of the scene. And I don't know what

00:16:36   kind of models that used and Apple is a little different

00:16:39   because they're doing most of this with their own stuff. If

00:16:43   you look at something like the human AI pin, they need to this

00:16:48   engine basically that understands intent. The only

00:16:51   thing they're doing is basically saying, okay, what does the user

00:16:54   want? Because we don't have the models, they don't have their

00:16:57   own on device models, they don't have image recognition. So they

00:17:01   just try to understand what the user wants. And then they go out

00:17:03   to Claude or to the perplexity or to chat GPT, or any one of

00:17:08   these AI services. It's called the AI bus. That's what humane

00:17:12   calls it. And it's actually a pretty good name. Because it

00:17:15   that is what it does. And basically, what's the what's the

00:17:18   input? And how do we get an output and make sure that that

00:17:21   output responds to the input from the user. And that's what I

00:17:25   think is so interesting about Apple intelligence. If that demo

00:17:29   that they gave in the keynote of Craig, where he said, Oh, I have

00:17:33   a meeting later, or excuse me, they want to change the time of

00:17:36   a meeting. Can we change the time and but I know my daughter

00:17:40   has a concert. And I didn't put that in my calendar, but she did

00:17:42   send me a PDF that has details of the performance. That kind of

00:17:46   stuff is a game changer. If that works, where I think there's

00:17:51   caution that will need to be had is that my main frustration with

00:17:56   the AI pin and the rabbit and all of these devices when

00:17:59   compared to something like Siri, they're vastly more capable. But

00:18:03   I ran into more and more limitations, because there was

00:18:07   this expectation of Oh, well, you can do anything. And that's

00:18:10   where I think it will be interesting to see if Apple can

00:18:13   pull this off is right now there's a lot of things that

00:18:17   people know Siri can't do. Siri's been to put a charitably

00:18:21   not very good for a long time, has gotten better. But most

00:18:26   people have just they know what Siri does. They asked her to set

00:18:29   alarms and timers and home automation stuff, but they're

00:18:31   not going to ask questions because they know it can't

00:18:33   respond. When it starts to be able to do those things. That's

00:18:37   exciting. But the more and more capable it gets the higher the

00:18:42   expectation from the user that it completes the thing that the

00:18:45   user wants it to do. And if it gets to the point where it can't

00:18:48   do that, or it doesn't do reliably, or in the case of the

00:18:52   AI pin, it'll do some things, but then five minutes later, it

00:18:54   will go, Oh, I can't do that when when it actually can.

00:18:57   That's where I think there's a risk of frustrating the end user

00:19:02   because they're kind of over promising of a the world's your

00:19:06   oyster. But then you're dealing with the realities of reasoning

00:19:11   in AI models that are still relatively young and small and

00:19:14   inherently limited because this is early days of AI. I don't

00:19:18   know. I keep thinking about and I go back to just writing silly

00:19:27   programs and basic in the 80s in like grade school and it's you

00:19:32   think I'm gonna we're gonna make a computer program that's so

00:19:35   smart, it can answer questions. And as a little kid, you just

00:19:38   wind up with if this else, if this else if this else and you

00:19:44   start, you just quickly realize, oh, you really can't build up

00:19:48   enough if F's of else if else to cover everything, even for a

00:19:53   limited domain used you what seems like Oh, we'll just build

00:19:57   a bunch of those. No, that that's not how you build

00:19:59   something that can answer questions arbitrarily. It just

00:20:05   always springs to mind in these LLM's really sort of are the

00:20:08   first time we're seeing computers sort of encroach on

00:20:13   human intuition, right? And and it's so easy to overlook how

00:20:20   good human beings are, by nature at intuiting things like context,

00:20:28   right? If you're talking like how far away is this bus? If you

00:20:33   were just talking to a person who's standing right next to

00:20:36   you, who you think knows the bus schedule. And then they say, oh,

00:20:40   it's probably about eight minutes away. And then you said,

00:20:43   could you text my wife and tell her when I'll get home? The

00:20:46   person would know. Oh, because we're me and Quinn are getting

00:20:50   on the bus, right? And we're heading home. You don't have to

00:20:54   say, I'm going to get on this bus with you. This is why we're

00:20:58   waiting here at the bus stop.

00:21:00   Yeah, exactly. Yeah. And there's just a zillion of those little

00:21:07   fill in the gaps that humans do intuitively that computers

00:21:13   don't and, and now sort of with LLM's they kind of do except you

00:21:18   have to double check. And that's it. So it's the intuition of

00:21:23   filling in gaps that really is the step change.

00:21:26   It is. And that's the hardest part. I mean, to bring in

00:21:28   another example of something that's also kind of ml and AI

00:21:32   driven, self driving cars, right? This has been something

00:21:35   that's been right on the horizon for years and years and years

00:21:39   and years. And certain individuals have promised that

00:21:41   it's right around the corner for years and years and years. And

00:21:44   what we've discovered over time is that despite these systems

00:21:49   becoming vastly more capable, they're still quite bad at

00:21:54   driving. Because there are instances in which humans do

00:21:57   stuff intuitively, that goes against traditional convention,

00:22:01   they're able to perceive things that go against the traditional

00:22:05   rules of driving or the road or they're, they're aware of

00:22:08   weather conditions or someone that might be inebriated in

00:22:11   front of them. And there's so many factors of when we drive or

00:22:14   when the lane markings are disappeared, or it's a, it's a

00:22:17   weird intersection that has this strange construction layout. And

00:22:21   the first time you were kind of confused by it, but you've

00:22:23   you've driven through it once before. And so now you know,

00:22:25   that's the kind of stuff that these AI models have to deal

00:22:29   with. And the reality is that least with self driving, these

00:22:34   roads and this infrastructure has been built for humans. And

00:22:37   so once the kind of self driving problem is solved, it's gonna be

00:22:45   easy. I mean, that's, once all cars self drive, it's gonna be

00:22:50   simple. The issue right now is they're driving with people on

00:22:53   roads designed for people. And so the hardest it's ever going

00:22:57   to be to develop self driving software is right now. And I

00:23:00   think the same is true with AI and LLM's for stuff like this on

00:23:05   on device and on platform. Apple has the benefit of being the

00:23:10   owners and carers of the system. And so they have access to a

00:23:15   bunch of information that third party applications just don't.

00:23:18   It's a huge competitive advantage. The question is, like

00:23:22   self driving cars, are they going to understand the intent

00:23:26   and the reasoning that a person would want with vague input,

00:23:32   like, humans talk to other people? I mean, there are entire

00:23:36   pieces of software that these reasoning engines, their entire

00:23:40   AI companies that exist to basically rewrite your prompts,

00:23:44   so that you can get the output from chat GPT that you want,

00:23:48   right? And it's because humans are inherently not very good and

00:23:52   not very descriptive at this stuff. But that's where the AI

00:23:56   is going to have to come across the aisle to get to the hardest

00:23:59   part is going to be right now. And unfortunately, once the

00:24:03   infrastructure is built out for cars, and once every car is

00:24:06   self driving, well, then it's no problem. But until that happens,

00:24:09   it's incredibly difficult. And that's not going to happen until

00:24:12   AI itself has been able to overcome those things. So it's,

00:24:15   it's almost like a catch 22. But that's where the promise of this

00:24:19   is incredible. And I'm actually surprised that Apple showed us as

00:24:23   much of their hand as they did, like you mentioned, because this

00:24:26   is the kind of stuff that they don't traditionally talk about.

00:24:29   But at the same time, it's the kind of thing where I go, nobody

00:24:33   else has really been able to figure this out. And it's still

00:24:35   pie in the sky. This is full self driving kind of stuff. Yeah.

00:24:38   And will Apple be able to come out guns blazing in a year and

00:24:42   have solved the issue? I don't know.

00:24:45   Yeah, because it's like, we all know how it's supposed to work.

00:24:49   And that's the other thing too, like, what, when does when does

00:24:55   my mom's flight arrive? You know exactly what you mean. It's

00:24:59   your mom's next flight. And you know that you must be you must

00:25:03   trust that your your mom has either texted you or emailed you

00:25:07   with the flight plan so that it's there somehow you wouldn't

00:25:10   be asking your phone this question if you thought there's

00:25:15   no chance that the information's even on your phone.

00:25:17   Totally.

00:25:18   Given our experience with Siri over the last what, 14 years?

00:25:23   How shocked will we be if the answer is about a flight your mom

00:25:29   took six months ago? Right? It's, you know, we're going to be

00:25:35   disappointed. It's not going to get reviewed very well. But I

00:25:39   don't know that if that's what happens, or if that's how it

00:25:42   happens out of the gate. We're not going to be shocked. Get me

00:25:47   to the fainting couch. Siri just gave me my mom's flight

00:25:51   information from 2019. It was there in your email.

00:25:55   Yeah, or even harder. What if the flight is changed? Or there's

00:25:59   a cancellation and the flight number changes? That's a much

00:26:02   more real world example of it. Serious. Well, we just grabbed

00:26:06   the flight from a day ago. We didn't know there was another

00:26:09   one that appended this and it's, it's gonna be hard.

00:26:13   It really is. And it really speaks to, like I said, just how

00:26:20   intuitive actual human beings are. Like when you hire a actual

00:26:26   personal assistant who is a human being, you just make all

00:26:31   sorts of assumptions that they're going to get if you tell

00:26:36   your assistant to get you, can you get me three tickets to the

00:26:42   Phillies game Friday night? You kind of know that they're going

00:26:46   to ask you certain reasonable questions like, well, how much

00:26:50   are you looking to spend or double check with me before you

00:26:55   spend an exorbitant amount of money or get me a seat all the

00:26:59   way in the third deck way too far away to see anything I know

00:27:03   you don't or you might expect your if you're have a you know,

00:27:09   if you've been working with the same human assistant for a

00:27:11   while, they might know something like, Oh, I know john likes to

00:27:15   sit behind first base or home plate usually not third base.

00:27:19   They know stuff like that. Totally all of these things. You

00:27:23   just expect these computer assistants to pick up but will

00:27:26   they I don't know. And we're not going to be satisfied until they

00:27:30   do. Right.

00:27:31   And what level of information do they retain over time? That's

00:27:35   another thing too. I mean, right, you can tell someone,

00:27:37   hey, I think I'm going to the jazz game on Thursday. And then

00:27:41   two days later, you can come back and ask a person, hey, did

00:27:44   you get tickets to the to the game? Or an AI model might go,

00:27:48   Okay, tickets to game? What game? There's too many sporting

00:27:51   franchises here until they are. And maybe they meant on the TV.

00:27:55   Well, here's an Apple TV suggestion of where you can get

00:27:57   this. It just so wide. And I think in the beginning, it is

00:28:02   going to be limited, and that'll be okay. But they have to do the

00:28:06   things that users expect them to do however limited very well.

00:28:10   That's where I think it will become frustrating because with

00:28:15   and again, I'm coming back to the AI pen, it's gonna be way

00:28:17   better than the AI pen. But one of the things the AI pen would

00:28:21   do was when I asked it to do something that couldn't do I'd

00:28:23   be like, okay, like it was worth a shot. I wasn't that frustrated

00:28:27   because it was never really a promise. But there were things

00:28:30   that I knew the pin could do that it would have done

00:28:33   previously. And then two hours later, it was just like, Oh, I

00:28:36   can't do that. And I'm thinking to myself, yeah, you can you

00:28:39   just did it a couple hours ago, or you did this task that's so

00:28:43   hyper similar that why would you not be able to do task B, it's

00:28:47   almost the same thing. And that's where it gets frustrating

00:28:51   is, is what's the level of expectation they set from day

00:28:55   one? And how good is the baseline? I think it will

00:28:59   satisfy me and most people, if it's good at just indexing

00:29:04   information. So if you say, Hey, when was my daughter's concert

00:29:09   supposed to be and it wasn't in your calendar, and it goes and

00:29:11   finds the PDF and then does OCR on the PDF and finds that's good

00:29:16   enough. I don't need it to do all the other stuff. But if it

00:29:19   can't even do that, or it fine, that's where it's gonna be this

00:29:23   failed promise kind of thing where the potential is

00:29:26   theoretically so high, but in practice, it's just not there. I

00:29:29   am excited because I think if anyone can manage this, it's

00:29:32   Apple, they're running the models on device, the model seem

00:29:35   quite powerful. They have full control of the operating system

00:29:38   that nobody has control of. If anyone can do it, it's them. But

00:29:42   that's also like saying it's NASA. Yeah, it's Yeah, they can,

00:29:46   but it doesn't mean that it's easy. You know what I mean? So

00:29:48   yeah, yeah.

00:29:50   All right, let me take a break here and thank our first

00:29:53   sponsor. It is our good friends at trade coffee, whose fine

00:29:57   product I just finished consuming earlier in this

00:30:00   recording. Trade coffee has a great deal for listeners of the

00:30:05   show 30% off your first month of coffee that's 30% off by going

00:30:11   to drink trade.com slash the talk show. Trade coffee offers

00:30:18   tailored plans for coffee delivered to your home. And it

00:30:24   is without lock in without any kind of rigidity. Or once you've

00:30:29   signed up, you're not getting enough coffee or you're getting

00:30:32   too much coffee or something like that they're tailored plans

00:30:35   can be adjusted on the fly as you see fit. For example, during

00:30:39   the summer months when you go on vacation, it is super duper easy

00:30:43   to just log in to the website and pause the delivery just skip

00:30:47   one delivery because you know you're not going to be home or

00:30:50   something like that. Or if you've got guests coming in to

00:30:53   stay with you over summer for a week or two and you know you

00:30:56   need more coffee than usual. You can get extra coffee for one

00:31:00   shipment in August or something like that. Anything you'd want

00:31:04   to do like that you can do. But the best part are just like the

00:31:07   normal weeks of the year where your coffee just shows up at

00:31:11   your door. At the pace you consume coffee and it's you

00:31:16   never run out of extremely fresh, extremely delicious

00:31:20   specialty coffee. Trade doesn't roast their own beans. They are

00:31:25   a sort of parent umbrella over a bunch, a whole slew of

00:31:30   independent roasters all over the country who they pair with

00:31:33   to and they make recommendations and you get new coffee from new

00:31:37   small family owned roasters all over the country on a regular

00:31:41   basis. The variety is terrific. But also, you just give a thumbs

00:31:45   up and thumbs down as your shipments arrive and they kind

00:31:49   of dial in your taste profile for lack of a better term. It's

00:31:52   just terrific. I recommend them very highly and it is super

00:31:56   duper flexible. And once again, special deal for listeners of

00:31:59   the show 30% off your first month just by going to

00:32:02   www.drinktrade.com/the-talk-show. So go try Trade Coffee if you

00:32:08   drink coffee. I recommend it highly and drink it myself. What

00:32:13   do we think about the I feel like if there's the biggest

00:32:16   sour note of that whole Apple intelligence announcement was the

00:32:20   cutoff for iPhones, right? It's for iPads and Macs. It's any M

00:32:27   series device and that feels completely expected, right? And

00:32:33   for the Mac in particular, it goes back at least to what the

00:32:36   end of 2020. Yes, feels that that's a pretty good backwards

00:32:40   window. It's better than I thought that there would be.

00:32:42   Yeah, I have cut out M one honestly. Yeah. Because there's

00:32:46   so many of those M ones hanging around that are just still

00:32:49   kicking so good. So it's good on Apple like that, like a

00:32:53   legendary class of devices, the M one max, because they debuted

00:32:58   to spectacular, too good to be true reviews. And here we are,

00:33:02   almost four years into it. And they're still kick ass machines.

00:33:07   I see it all the time where people are like, you know, I

00:33:10   was thinking about upgrading, but I've got this M one that I

00:33:12   bought the first generation and it I've got no complaints with

00:33:16   it. I've long said the base model M one Pro 14 inch MacBook

00:33:20   Pro will go down in history as one of Apple's best machines

00:33:24   ever. I mean, that's legendarily good, especially for the price.

00:33:28   I mean, you could get those things for 1500 bucks on sale

00:33:32   within a year of them releasing and it's still just a killer

00:33:37   machine. That made really, really good. And so I would say

00:33:40   that the backward compatibility for Apple intelligence on Mac

00:33:43   and iPad, very fair. Now I realized the difference is with

00:33:48   a Mac, pretty much all well, effectively all Mac books. And

00:33:53   that means most Macs that people have bought since the end of

00:33:57   2020 have been Apple Silicon, right? That it's some number of

00:34:01   people bought a Mac Pro for reasons in 2021. But for the

00:34:06   most part, anybody who's bought a new Mac in the last three

00:34:10   years has one that qualifies for this. Whereas the story for

00:34:13   iPhones is for iPads. And there have been M one iPad pros for a

00:34:21   while, but it's only in the pro line, I guess, right. Or is

00:34:24   there an iPad air now?

00:34:25   Yeah, there was an M one air and that's, yeah.

00:34:29   So there's a couple, but it's, you know, it's pretty broad and

00:34:32   it's, you know, it's the ones that cost 800 bucks or more. They

00:34:36   qualify from the last few years, but that, that sad trombone is

00:34:40   the only iPhone that qualifies is the iPhone 15 pro, which is

00:34:45   nine months old. And even the regular non-pro iPhone 15, which

00:34:50   debuted just last September, doesn't qualify.

00:34:53   Yep. Brand new phone. You can still go buy in store. No go.

00:34:57   Yeah. It's a real bummer. JG on your show basically said, we

00:35:02   could, in theory have gotten Apple intelligence to run on.

00:35:06   And he said, very old devices was, I think the phrase he used.

00:35:10   Yeah. However, the performance would have been so poor that it

00:35:15   wouldn't have been worth doing. And I actually believe that I

00:35:17   think that much of what makes these models compelling is their

00:35:21   speed. And if it can't be done well on device, if you've got

00:35:24   memory limitations that just makes it not feasible, or you

00:35:27   have to load stuff in chunks, it's not ideal. It's, it's not a

00:35:30   good experience. The disappointment is that. Yeah.

00:35:33   There's one single iPhone that does this thing. And, and the

00:35:38   iPhone 15 is not a very old device. It is a current

00:35:41   generation device that is for sale today.

00:35:43   I wonder if his thinking on that, and it probably wasn't the

00:35:48   most judicious way of him phrasing it, but it's one of

00:35:52   those, it was one of those moments on the show where I was

00:35:54   like, Ooh, I was worried they wouldn't answer that question at

00:35:56   all. And that was actually a pretty straightforward answer

00:35:58   that yes, the RAM matters, right? Apple doesn't like to

00:36:01   talk about RAM and iPhones. And they more or less came right out

00:36:04   and said, yeah, well, it's, it's everything. It's the whole

00:36:06   device. It's the neural engine size, but RAM definitely matters.

00:36:10   And infamously, at least in tech circles, iPhones have been

00:36:15   relatively starved of RAM, which it's more complex than just Tim

00:36:20   Cook bean counting, right? Oh, add another $10 worth of RAM to

00:36:24   each iPhone and it affects 1% of the profit margin or something

00:36:28   like that. But every gigabyte of RAM you add to a device consumes

00:36:33   electricity. Once you put RAM in a device, it always consumes

00:36:36   electricity. More RAM adversely affects battery life. I don't

00:36:40   know by how much like for six versus eight gigabytes. It's

00:36:44   probably true that if like the A16 chips had two more gigs of

00:36:50   RAM, it really wouldn't affect battery life that much. But you

00:36:55   know, every little thing counts for battery life, right? Like

00:36:58   I've spoken to people who work on the team at Apple, more or

00:37:02   less the battery life team, which can kind of swoop into any

00:37:05   project within the company. We've identified that this is a

00:37:09   battery drain. It's hundreds and hundreds of little things that

00:37:13   contribute to battery life and RAM is one of them. But basically

00:37:17   it's it's undeniable at this point that it kind of looks like

00:37:20   a little bit of an own goal that Apple didn't move iPhones to

00:37:24   eight gigs of RAM a generation earlier at least.

00:37:27   Well, yeah, and I don't think that it was malicious intent. I

00:37:32   think it was frankly, probably, I mean, Apple's an enormous

00:37:35   company. I do think that if everything is to be believed on

00:37:39   how this kind of came to be that this Apple intelligence stuff

00:37:42   has happened fairly recently. And it's likely that the iPhone

00:37:47   15 generation of phones was already pretty much finalized by

00:37:51   the time this really started getting steam, the teams are not

00:37:54   the same team. So they're not really talking about the

00:37:57   requirements. And so it kind of just came to be that I have no

00:38:00   doubt that by the time the iPhone 15 went on sale last

00:38:03   September, they knew it wasn't going to run Apple intelligence.

00:38:06   But I'm not confident that they would have known that when they

00:38:08   started development on the iPhone 15. So I think come this

00:38:12   fall, if we see an iPhone 16, that still doesn't have eight

00:38:15   gigs of RAM and can't do Apple intelligence. That's gonna be a,

00:38:19   we should get mad at them. Yeah, yeah.

00:38:22   What, what the hell's going on? Right? Yeah, yeah. Do you think

00:38:27   that it's sort of maybe not a comedy of errors, but a sort of

00:38:32   just unfortunate timing that it seems like Apple ran into not a

00:38:43   like a hiccup with going to three nanometers at TSMC with

00:38:48   the eight. So the a 17 Pro and the M three chips are Apple

00:38:53   silicon on the first generation three nanometer process from TSMC

00:38:57   and out over my skis here talking about this you but it's

00:39:02   good that you're here much more you're you have much more

00:39:05   informed opinions about this to me. But basically, that first

00:39:09   generation three nanometer process is expensive. And it's

00:39:13   probably why the eighth, the regular iPhone 15 are using the

00:39:21   a 16 chip from the year before. And maybe I that's what I was

00:39:26   getting at with JJ's answer that maybe from his perspective, he

00:39:30   doesn't really see from a software perspective, the iPhone

00:39:33   15 isn't nine months old, it's two years old, because it's

00:39:37   that's the silicon. Yeah, it's it's possible. I think. Yeah,

00:39:44   it's tricky. Apple has been the kind of number one purchaser of

00:39:49   silicon from TSMC. Much of it is you've got to have the money and

00:39:53   you got to get in line to get the ticket deals and the kind of

00:39:57   the fabbing room that that you need. So I think they kind of

00:40:01   had to move to n three e by necessity. However, it is a bit

00:40:07   of a bummer kind of a black guy, in the sense that not that it

00:40:11   was a bum process, but it was expensive. It didn't really

00:40:16   yield the expectations and performance that people had come

00:40:20   to expect. And we don't really know the yields. TSMC doesn't

00:40:24   talk about it. Apple certainly doesn't talk about it. But I

00:40:27   would suspect they're not very good. And every everything that

00:40:31   you know, all the reports suggested it wasn't very good.

00:40:33   And so it's hard to say that it was cynical. But I do think that

00:40:42   they that it was unsurprising to me that they were not quick to

00:40:47   put the a 17 Pro in the non pro iPhone just because of the

00:40:52   amount of volume that the iPhone does. There literally may not

00:40:56   have been the overhead to be able to do that. Now with the M

00:41:00   for being on the new second generation and three process

00:41:03   from TSMC, I really expect that to come to the to the iPhone

00:41:08   this year. I'm still not confident that some of the

00:41:12   features of the a 17 Pro I mean, they gave the chip name Pro

00:41:17   right that they'll come down to some of the iPhone 16 stuff. So

00:41:21   one example is like the a 17 Pro has an onboard USB three

00:41:25   controller. The iPhone 15 famously does not support USB

00:41:29   three or Thunderbolt, it's just USB two over USB C, I will be

00:41:35   unsurprised, let's say, if a USB three controller does not make

00:41:39   its way into the base model iPhone even with newer silicon.

00:41:42   So I think they'll still find a way to kind of not fragment but

00:41:46   to cost cut in order to hit those lower prices on the kind

00:41:51   of the base model iPhone. But I do think that memory limitations

00:41:55   and then kind of just generalized fab stuff is going to

00:41:59   start to go away because TSMC new process is really good.

00:42:03   Apparently, the yields are pretty good. And I would expect

00:42:05   that to come kind of down to the rest of the lineup. I mean, look

00:42:08   at the look of the M for its killer, right? Yeah, it was way

00:42:11   more performant than anyone expected.

00:42:12   Yeah, so hold that. Yeah, hold that thought. But let me let me

00:42:16   just ask you to repeat that, though. Do you see you're

00:42:19   saying, come September, when we get the iPhone 16, and the 16

00:42:24   pros, you're not going to be surprised if the 16 get new

00:42:29   silicon, which is almost certain whether it would be shocking if

00:42:32   they're still on the ASX. I can't. But so yeah, but that

00:42:35   their silicon won't have the quote unquote, pro level USB

00:42:40   three. USB three. It feels funny to call pro Thunderbolt. You can

00:42:45   call pro. Okay, sure. Sure. But you think that maybe it won't

00:42:48   ship with that? I could see that too. Where they might not.

00:42:51   Yeah, I think they'll both get an a 18 because the a 17 pro

00:42:55   that's that's dead. That's old fat. They're not going to keep

00:42:57   the name. But I wouldn't be surprised if we see an a 18 and

00:43:01   then an a 18 Pro. And some of the distinguishing features are

00:43:05   on die memory, which has always been a characteristic of the

00:43:08   higher end chips, and then maybe goofy stuff like that USB speeds

00:43:13   or Thunderbolt speeds. I don't think they would extend that to

00:43:16   Wi Fi, but I suppose that's maybe a possibility to or maybe

00:43:19   the pro chip gets Wi Fi seven and the other one sticks with

00:43:22   60. I don't know. I wouldn't think so. But who knows?

00:43:24   Yeah, but they definitely see the Thunderbolt stuff as a pro

00:43:29   configuration. And they're all in and you've been documenting

00:43:32   this with each keynote ever since they've started shooting

00:43:35   them on iPhones last year. That's there. They really are

00:43:40   dogfooding the iPhone 15 Pro camera as their camera. They've

00:43:45   they've they've ever since they started, they've shot all of

00:43:48   their keynotes using their own devices, which is kind of

00:43:52   flabbergasting given what we know of their production quality

00:43:55   standards.

00:43:56   Yeah, especially with new apps like Final Cut camera and Final

00:44:01   Cut for iPad two, which is the worst name ever. Yeah. I've had

00:44:07   to that is a Microsoft name if I've ever heard.

00:44:10   All right, this is this is we have to have an aside on this.

00:44:13   The name of the product is Final Cut Pro for iPad two, which

00:44:18   really sounds like it's made for 2011 iPad two. And we know

00:44:24   again, it speaks to the human intuition that I was saying,

00:44:28   we're not robots. So we know that Apple in 2024 did not just

00:44:35   release Final Cut Pro for a 2011 iPad. But that sure is what

00:44:41   this what it sounds like. It's like the those those grammar

00:44:44   books, like eat shoots and leaves. It's, it needs

00:44:47   rewriting. I would suggest Final Cut Pro for iPad version two.

00:44:53   But yeah, I would go back in time and name it something fun,

00:44:57   like Final Cut Go or something like that. Yeah, maybe I don't

00:45:01   know. Because it's not really Final Cut Pro. Right. You tried

00:45:05   using it, right? Yeah. I mean, I know. We're sort of over the

00:45:09   place. And I couldn't really. So I bought the one terabyte and

00:45:13   four iPad. The footage that we shoot here in studio is ProRes

00:45:18   422, very large footage, the kind of compression ratio, it's

00:45:23   very, very little. So it's this enormous amount of data. And

00:45:27   you've got to manage the library size and the iPad in some

00:45:31   instances wants to make proxy footage. It doesn't with ProRes

00:45:34   because it plays back ProRes like butter and anything can

00:45:37   now ProRes is a pretty awesome codec, if not very highly

00:45:41   compressed. And I literally could not use Final Cut because

00:45:45   even though I had a one terabyte iPad with almost nothing on it,

00:45:48   I couldn't fit a single video's worth of video footage on the

00:45:51   iPad, you couldn't use an external drive. Now with Final

00:45:54   Cut Pro for iPad two, you can, however, it's still limited in

00:45:59   some of the stuff that it can do. It doesn't have the entire

00:46:03   feature set of the Mac variant of Final Cut. Now it does come

00:46:07   pretty close, closer than you would think. But I think you

00:46:11   would be hard pressed to find anybody that's seriously editing

00:46:14   in Final Cut, that does not have a bunch of extensions and tools

00:46:19   from third parties inside the NLE. And that's something that

00:46:23   the iPad obviously does not support. So in that way, it's

00:46:29   pretty much a no go still for me. But I think it's close

00:46:33   enough that you can get 95% of the video done. You can keep it

00:46:37   stored on an external drive now. And then when you get back to a

00:46:40   Mac, you plug it into the Mac, you finish up the last 5%. The

00:46:43   same story is kind of true with DaVinci Resolve for iPad, where

00:46:48   Resolve is very odd, is they disabled a few of the tabs.

00:46:53   DaVinci Resolve has a bunch of tabs, they have like cut tab, a

00:46:56   color tab, a fairlight tab, which is their like compositor,

00:47:00   their 3d rendering engine thing, actually, fairlight might be

00:47:03   Adobe, in any case, it doesn't matter. They've got a bunch of

00:47:06   different environments that you can use inside of this app. And

00:47:09   in typical DaVinci, I should say, maybe say a non Apple

00:47:13   fashion, this is something Microsoft would do, it seems

00:47:16   that they ported the entire app. So all of the functionality of

00:47:20   DaVinci Resolve theoretically, is in the iPad version. But then

00:47:24   from like the from a build level, they went in and disabled

00:47:30   a bunch of those pages that didn't work very well on the

00:47:32   iPad. But what they didn't disable was keyboard work

00:47:36   arounds that you can use to pull those pages back up. So you can

00:47:39   open up DaVinci Resolve for iPad and open up panels that they

00:47:43   specifically closed because they don't run very well on the iPad,

00:47:46   because it is the full desktop app on the iPad, it's the whole

00:47:49   thing. It just doesn't do certain stuff, because it's

00:47:53   missing components because of the way that iOS sandbox

00:47:56   required for it to work. So the 3d kind of the modeling engine

00:48:01   where you can do VFX stuff, that's all available, it just

00:48:04   doesn't do anything, you can go in and push all the buttons and

00:48:07   just nothing happens because the underlying code is not there on

00:48:10   the iPad and or can't run. But I was able to re enable a bunch of

00:48:14   the tabs that by default were shut off. I don't know why.

00:48:17   Because the iPad and I'm guessing it was for overhead on

00:48:21   older iPads, but now with the M four worked great. And I edited

00:48:25   a full video doing everything I normally would have done on a

00:48:29   Mac on the iPad, like legitimately 100% worked

00:48:32   awesome.

00:48:33   Huh? That's really interesting. And it is true, that is very much

00:48:37   sort of a taste of traditional PC software development on on

00:48:43   the iPad of Yep. You know, and I am not a DaVinci Resolve user in

00:48:48   any way, but I've seen it in demos. And it is sort of, you

00:48:53   know, it's obviously written in some kind of cross platform way

00:48:56   where it doesn't run as a regular Mac app, it runs sort of

00:49:00   in a rectangle, it has its own menu bar. So like the actual Mac

00:49:04   menu bar just has two commands like quit. Yeah, it's gotten a

00:49:08   lot better. But originally, like I started using resolve with

00:49:11   DaVinci resolve 15, I think, which probably came out at this

00:49:14   point, five years ago on the Mac. And resolve was originally

00:49:18   an app for Linux, because they use it in Hollywood in these

00:49:22   air gapped machines that are running like Red Hat Linux that

00:49:25   you know, it's so they did this initial kind of port to Mac OS

00:49:30   and Windows. But they were such rudimentary reports that in

00:49:34   order to quit the app, there was no system level quit command. So

00:49:37   you couldn't command queue, you had to hit the close button in

00:49:40   the top right of the window, of course, because that's where

00:49:43   where Linux and Windows had put it. And so it was really janky.

00:49:48   It's gotten much better over the years, but you still open it and

00:49:51   you think to yourself, yeah, this is not a Mac or an iPad

00:49:55   first app, that's for sure.

00:49:56   Right. And the the diehard, native Mac asked Mac app

00:50:02   proponent in me, objects, objects to that on the surface,

00:50:06   but I totally get how at the extreme ends of professional

00:50:10   software, that's where you tend to run into this type of app.

00:50:14   And I do appreciate that it if it really is not really a Mac

00:50:19   app, and they're not deeply or even shallow, we use using app

00:50:24   kit, the Mac only application framework, you really can sort

00:50:28   of just recompile it for iPad, and it just runs in a rectangle

00:50:31   on the iPad.

00:50:32   Yeah. And to be fair to resolve, they've spent a black magic has

00:50:36   spent a bunch of time making sure that it uses the video

00:50:42   engines that are available on Apple silicon, it's not running

00:50:44   some janky ported OpenGL that Adobe did for a long time, like

00:50:49   this is native uses metal, and all of the API frameworks that

00:50:54   Apple supports just in an app that's quite ugly and on Apple,

00:50:57   but it's one of the reasons why you can export a video at crazy

00:51:01   lightning fast speeds in super high resolution from really high

00:51:04   resolution source footage on a freaking iPad, and it works

00:51:08   great. It's so cool.

00:51:10   It's also a reason why Apple likes to show it to the media as

00:51:14   whether it's for the Mac or the iPad, even though it's not you

00:51:17   typically, oh, this doesn't really look like a great Mac or

00:51:21   iPad app is a big red X against his Apple going to use this as a

00:51:25   demo. But when it does amazing things with amazing performance,

00:51:30   because it's taking advantage of metal and really using the

00:51:34   Apple silicon GPUs to the full extent that they're possibly can

00:51:37   be used, then Apple's like, yeah, we'd like to show this off

00:51:40   even if don't look at it.

00:51:42   I don't know if there's a better example of a pro app than

00:51:44   resolve.

00:51:45   On the iPad, do you get to use the menu bar because the menu

00:51:50   bar is their own?

00:51:51   Yes and no. So they've put some of the menu bars into more

00:51:55   iPad esque sub menus. So they have moved the UI around a

00:52:00   little bit, but not much. There's like this home button

00:52:04   that takes you back to the main project window that's that's

00:52:06   available inside of the desktop version of resolve. But yeah,

00:52:10   you can tap a project. There's a little kind of triple dot

00:52:13   thing. And if you click that with your mouse cursor, a big

00:52:15   sub menu pops up and you can go into a full windowed settings

00:52:19   menu pane that you would see on the desktop. So effectively,

00:52:23   yeah, I mean, you can, you can get access to, and because much

00:52:27   of what they've done to make it feel more iPad esque is they've

00:52:31   disabled some of the stuff that is ugly and they've moved them

00:52:35   into more iPad looking sub menus, but all of that stuff is

00:52:37   still there because the code base hasn't changed. And so

00:52:40   you can use hotkeys that you're used to using on the desktop

00:52:43   and it'll pull up some janky looking menu that doesn't fit

00:52:47   well at all on the iPad, but it's there and it still is

00:52:50   functional. So anything you can do on resolve on the Mac, save

00:52:54   for kind of the, the compositing stuff you can do on the iPad.

00:52:56   And it might be ugly, but it works and it's pretty cool.

00:53:00   Yeah. It seems sort of like a poster child for some of these

00:53:03   arguments that we collectively have been having about the iPad

00:53:07   as a pro machine. It's well here, if you took this pro software

00:53:10   and just sort of let it be its desktop self to the fullest

00:53:15   extent possible, here's what it would be like. And it's kind of

00:53:19   weird and kind of ugly in some certain ways, but maybe one of

00:53:23   the most capable apps period on the iPad. Yeah. And I think

00:53:27   there is an argument to be had about, I think if Apple had

00:53:31   their way, they would go to Blackmagic and say, Hey, we

00:53:33   love the functionality and the power of this app. Let's spruce

00:53:36   it up visually a little bit. Let's maybe move some stuff

00:53:38   around. So it doesn't feel there's even like weird scaling

00:53:41   issues where some of the text is really small. Cause again,

00:53:44   it's just some lazy, I don't want to say lazy port, but it

00:53:46   is, it's kind of a lazy port, but I feel like part of what

00:53:50   makes it so compelling for people that are using this type

00:53:53   of app is that it is DaVinci resolve. It's not some weird

00:53:55   iPad version of DaVinci resolve. It's the app that I use

00:53:58   on my Mac. And so I don't have to relearn a new user

00:54:01   interface. I don't have to think to myself, where did they

00:54:03   hide that thing? That's not in the sub menu it's normally in.

00:54:06   Can the iPad even do that? It just, if you can do it on the

00:54:09   Mac, you can do it on the iPad. And for certain apps like that,

00:54:12   that might be the way to go, honestly.

00:54:14   Yeah. And that it just to sort of wrap up the point that the

00:54:19   M4 appearing last month in the new iPad pros was definitely a

00:54:23   surprise. And I, sir, I infamously wrote like when

00:54:27   Gurman reported on it, I don't know, this doesn't sound right

00:54:30   to me. It just based on their schedule, but it really in

00:54:33   hindsight, it makes sense because of what we were talking

00:54:35   about with the TSMC fabrication and that if it does, if the

00:54:40   yields were too low, and therefore yields being low makes

00:54:43   it more expensive. If it was expensive to start, and the

00:54:47   yields were low, it explains why only one iPhone got the A17

00:54:51   Pro. It explains why Apple was more or less in a rush to go

00:54:55   from the M3 to the M4. They are a by all accounts, because

00:54:59   WWDC has come and gone. They are going to skip the M3

00:55:04   generation for the Mac Studio and yep, no ultras for the M3

00:55:10   and wait for the M4. And it's really about getting to that

00:55:14   second generation three nanometer process. And so yeah,

00:55:18   I would be flabbergasted if all the new iPhone 16s aren't

00:55:23   using, I guess I'm guessing the names will be A18, maybe

00:55:28   A18 Bionic, I don't know. Maybe. And A18 Pro for the Pro chips.

00:55:34   And maybe they'll just be binned that there'll be more GPUs.

00:55:38   I don't know what the differences will be, but I would

00:55:41   anticipate that they didn't just introduce the A17 Pro name

00:55:45   for one year. So I guess A18, A18 Pro. But I think the other

00:55:51   thing is also going back to all of this Apple intelligence

00:55:56   being revealed by Apple almost unprecedentedly early by Apple

00:56:02   standards means that maybe this device cutoff doesn't matter

00:56:07   that much anyway, because a lot of this stuff they're already

00:56:12   saying won't even come in developer betas till later this

00:56:15   summer. All of the features in Apple intelligence are at the

00:56:19   earliest later this year. And they've already said even come

00:56:25   fall when the new OSs are out, the Apple intelligence features

00:56:30   will still be called beta. That once iOS 18 comes out and you

00:56:35   can just install iOS 18 on your phone, at whatever pace they

00:56:40   roll out the features of Apple intelligence, they're going to

00:56:43   ship as beta apparently, even when you're not running a beta

00:56:47   version of the OS. And some of the features they're even saying

00:56:50   next year, so it won't come till 18.3 or 18.4 early next year or

00:56:57   something like that. So by the time a lot of these features come

00:57:01   out, it won't just be the A15 Pro, there'll be all of the A16

00:57:05   phones too. And if we really think that this is all a year

00:57:10   ahead of time, mostly, we're really talking about like for a

00:57:14   lot of people, it may not be till the iPhone 17s come out.

00:57:19   A year and a half from now, when people really start, you

00:57:24   know, again, to overuse the metaphor, but there's kind of

00:57:27   skating to where the puck is going, not where it is right

00:57:30   now. And this, I don't think this will be that big a deal

00:57:34   that the cutoff is so limited right now in terms of hardware.

00:57:38   I agree. I think that Apple intelligence is very much not a

00:57:43   singular product, it's a platform, and that we're going

00:57:46   to continue to see new features added and new capabilities for

00:57:50   years to come. I mean, this is, I wouldn't be surprised if in

00:57:53   future WWDC keynotes, we almost get an Apple intelligence

00:57:56   section independent of, you know, as if it were its own

00:58:00   operating system kind of platform in many ways. Yeah, in

00:58:03   many ways, it somewhat will be. The other thing, I don't

00:58:06   actually think this is going to happen. But I, I thought it

00:58:09   would be interesting, perhaps to postulate. Apple still

00:58:14   doesn't really have any usage data on any of this stuff. They

00:58:16   don't really know what users are going to be most interested in

00:58:20   or what will be used the most. I'm sure they have ideas

00:58:23   because many of these features still exist, or many of these

00:58:25   features do exist in other, you know, AI company products. But

00:58:30   I wouldn't be surprised perhaps if this private cloud compute

00:58:34   functionality makes its way to some older devices via iCloud

00:58:40   Plus, if there's a way for them to use services, or to increase

00:58:45   services revenue by bringing maybe some, it's not going to do

00:58:48   everything, but some of these simpler AI features that don't

00:58:51   run on device to older devices, and it just comes with the

00:58:54   gotcha of, well, it's going through the cloud, it's going to

00:58:57   take a minute, but it can still do the thing. I wouldn't be

00:59:01   surprised if that happens next year, sometime into the future.

00:59:04   Yeah, I, because that's, and I mentioned it in my write up that

00:59:09   I published last night, that perhaps the most common question

00:59:15   I've been asked after WWDC, and I think it's a super

00:59:18   interesting, very, very good question is, okay, if they've

00:59:23   got stuff on device and the on device stuff is limited to the

00:59:27   iPhone 15 Pro and later, but they've also got private cloud

00:59:31   compute, why not just do everything through private cloud

00:59:35   compute on older phones? And I think it's, I think, I, I wish I

00:59:43   had asked a question along those lines on stage, because I,

00:59:47   because I think it's actually a really good topic. And I think

00:59:50   they might have said some interesting things about it. But

00:59:53   I also kind of think they may not want to talk about some of

00:59:56   the reasons and I, I, I, but like you said, I can't help but

01:00:00   think that maybe they will expand that and that they might

01:00:04   expand both things, right? Obviously, as time goes on,

01:00:07   they're going to do more stuff on device five or six years from

01:00:10   now, they'll be doing way more on device than they are now,

01:00:14   because the devices will be five or six years into the future in

01:00:18   terms of performance and memory and stuff like that. But I also

01:00:21   wouldn't be surprised once they get confidence and scaling and

01:00:26   know how much computational power private cloud compute

01:00:30   takes, that they'll expand what private cloud compute can do,

01:00:34   too, maybe by bringing it to other devices. But I can't help

01:00:39   but think that part of the reason for this iPhone 15 Pro

01:00:41   cutoff, in addition to the RAM and the neural engine size and

01:00:44   everything is they just want to minimize the initial onrush of

01:00:51   traffic. They, they don't know how it's going to scale. They

01:00:55   obviously have a, they have a theory as to how it's going to

01:00:57   scale. But I don't think anybody really knows how anything will

01:01:01   scale until you flip the switch and go live.

01:01:03   And I think maybe one technical reason that could be given this

01:01:07   seems like the answer maybe jaws would give. I don't know if it's

01:01:12   a technical reason or but it's one that I've heard postulated

01:01:15   that does make sense is Apple intelligence isn't just the AI

01:01:21   model that actually does the thing you want it to do. It's

01:01:23   also this kind of, you know, intent engine that's supposed to

01:01:28   understand, okay, what's the thing that wants to be

01:01:30   performed. And if there's some reason, perhaps technically,

01:01:33   that devices with not enough memory or whatever can't even do

01:01:35   that part, where they don't even know what is being asked of it.

01:01:40   And we don't want to just send everything to the cloud,

01:01:42   because that's the whole point of Apple intelligence that it

01:01:44   only sends what's required and that it's deleted right after if

01:01:48   it can't even do that part of what does the user want to have

01:01:51   happen, then how can it do the rest of the AI stuff that maybe

01:01:55   is one rationale. But yeah, I don't know, it does seem like an

01:01:59   opportunity to get a bunch of services revenue out of people

01:02:03   into the future because people buying. I mean, Apple knows the

01:02:06   average upgrade cycle of iPhones, it's several years now.

01:02:09   And there's a lot of people that have bought iPhone 15 that might

01:02:12   subscribe to iCloud plus or something if image generation,

01:02:17   for example, could be done in the cloud, but they're not going

01:02:19   to go buy a brand new phone, they'll just wait three or four

01:02:22   years till they go buy their next upgrade, you know, so

01:02:26   I had to double check, but it is true. And I guess it makes

01:02:29   sense, given that it is LLM generative image generation, but

01:02:35   Genmoji is part of Apple intelligence and therefore

01:02:38   subject to these device limitations, right? And and yeah,

01:02:44   it is, again, I don't blame Apple because they had so much

01:02:47   to announce. But in the structure of the keynote,

01:02:51   everything in the first hour isn't part of Apple

01:02:54   intelligence until they mentioned Apple intelligence,

01:02:56   like an hour and five minutes into it, and started the second

01:03:00   half of the keynote. All those previous features are available

01:03:03   to other devices, the whole math notes thing, which is so fun

01:03:07   and super cool, goes back to every device that's eligible

01:03:11   for iOS 18. So whatever the earliest cutoff is for an iPhone

01:03:15   that can run iOS 18, you can do math notes on it. The

01:03:19   handwriting match feature, they call it smart match, I think,

01:03:23   or smart script. Smart script is their name for it, which in

01:03:26   my testing so far isn't really doing a great job at my

01:03:30   handwriting. It is making my handwriting more legible, but

01:03:33   it's also to me not recognizable as my handwriting. But still,

01:03:37   I think any improvement over my really bad handwriting is,

01:03:42   but that's coming to every iPhone or iPad, obviously,

01:03:47   because it requires a pencil, A14 or later, and I forget

01:03:51   where the cutoff, but that's pretty reasonable. That's a

01:03:53   couple of years worth of iPads. But they definitely, I think,

01:04:02   left people confused because they announced so much at once

01:04:05   as to which devices are going to get any of these features

01:04:08   and which devices aren't. And I just don't know how much

01:04:12   better job they could do by trying to have one two hour

01:04:16   keynote that covered all of this.

01:04:17   Ben

01:04:26   features come and part of me suspects that even they are not

01:04:31   fully entirely sure yet. I mean, they've Yeah, what exists in

01:04:34   the beta today? That's obviously, you know, that's part

01:04:37   of iOS 18 or iPad OS 18. But much of the kind of Apple

01:04:41   Intelligency stuff, I wouldn't be surprised. I mean, so one

01:04:44   other example, like this image playgrounds app, where you can

01:04:47   go in and make whatever photo you want, that's all running on

01:04:50   device, right? So that's not in the cloud right now, as far as

01:04:53   I'm aware. But I don't know. I really thought that they're

01:04:58   done on device. I think they're I think Jen Moji might be on

01:05:02   device, but the image playground is not. But they also

01:05:06   they have been made to their credit. They behind the scenes

01:05:11   and briefings when I asked, Are you going to show what's local

01:05:14   what's remote? And they said, there's going to be a log you

01:05:18   can look at. And we'll show you when it went to private cloud

01:05:22   compute, but they're not going to show it in the interface,

01:05:24   whether it's going remote or not. And they're not talking

01:05:28   about which features are which because they just fully admitted

01:05:30   we don't know yet. And we reserve the right even after we

01:05:35   ship and come out of beta, we're going to be constantly all the

01:05:38   time. Sure. Okay, so well, then that's maybe a better argument.

01:05:42   Much of the kind of image playground stuff is context

01:05:46   based, right? So they showed in the keynote, oh, hey, it's mom's

01:05:50   birthday. And yeah, who knows that. And so it generated this

01:05:53   image in the messages app. Don't bring that because that requires

01:05:56   Apple intelligence. But why couldn't you just bring a basic

01:05:59   image playgrounds app where you can go in and there's no

01:06:01   contextual, no smart suggestions that they have. It's just like,

01:06:05   Hey, picture of Brenda on top of a mountain or whatever. And it

01:06:10   could do that all in the cloud, you would think that would be

01:06:12   something fairly easy, because that's something that, well,

01:06:15   basically, every image generation service already does.

01:06:19   And maybe that would be Apple's reasoning as well. Just go use

01:06:22   something else. But seems like money left on the table. I don't

01:06:25   know how

01:06:25   funny would it be if Genmoji is the feature that drives people

01:06:29   to like, okay, I was gonna wait another year or two to get an

01:06:32   iPhone, but I've got to get a new iPhone because it may got it.

01:06:34   It cracks me up. And I know this for a fact that Apple, I

01:06:40   forget if Apple has a say in world emoji J day, or I know

01:06:45   Apple is, they're influential across the industry, but they're

01:06:49   particularly influential in the emoji world. Because they're

01:06:53   Apple and that's where people they're sort of Apple's emoji

01:06:56   set is sort of the de facto baseline standard that all other

01:07:01   emoji sets are compared against. But I know that Apple times

01:07:07   their release of new emoji, the like here, the inner the Unicode

01:07:14   consortiums new emoji of the year are out and we're going to

01:07:18   unveil our versions of the new emoji. They always wait until

01:07:22   iOS 7, usually right. Point one or point two. And it's, it's the

01:07:26   one that they think, Oh, this is the one that's really stable.

01:07:30   And that's when they start hitting the button that pushes

01:07:33   the pushes the updates to people, right? So if you're just

01:07:36   a typical iPhone user with default settings, which include

01:07:41   okaying to automatic app store updates and automatic OS updates,

01:07:48   hey, Apple. And I think that's the right way to go. I turn that

01:07:52   stuff off. I like to upgrade all my stuff manually. I'm sure you

01:07:55   do too. Right. But I get why they want typical users to say,

01:08:02   let, let Apple handle when my device will update. They don't

01:08:06   update people to iOS 17.0 automatically. They usually wait

01:08:11   till point one, sometimes point two, but that's also when they

01:08:14   do the new emoji because they want people to upgrade and they

01:08:18   know that people I've spoken to people at Apple. They're like,

01:08:21   yes, emoji prompts people to update they, and especially once

01:08:25   they get a text message from somebody else who's updated

01:08:28   using one of the new emoji, they're like, wait, I can't see

01:08:31   that. Square. Yeah. You have to update and they're like, okay,

01:08:36   I'm going to go update right now. Yeah. Be very funny. And

01:08:39   even Jen, the driver as an, it very well could be as silly as

01:08:44   it was like an emoji when it shipped with the iPhone 10 that

01:08:48   was mocked by tons of people and everyone thought this is so

01:08:51   stupid. And that was a huge driver. I think in interest of

01:08:56   the iPhone didn't mean everyone went out and bought an iPhone 10,

01:08:58   but it was, I mean, it was probably my vote, my most viral

01:09:02   piece of content I've ever made. I made a, with an iPhone 10, I

01:09:05   made like a quote unquote Apple ad in the same style that you'd

01:09:09   find on TV. And I did an animoji dub to Gasolina from daddy

01:09:14   Yankee. And I think that got, I think that got like 250 million

01:09:20   views on Twitter or something. Oh my God. It was like the

01:09:23   number one tweet for a couple of days. And so I think that kind

01:09:26   of stuff as goofy as it is legitimately does drive real

01:09:30   interest. I don't know that it sells a bunch of phones, but it

01:09:33   definitely sells some of them. Yeah. I will say before we leave

01:09:37   the subject, I will say that I would say the least impressive

01:09:41   thing Apple showed in all of Apple intelligence to me was the

01:09:45   image playground app, just because it looked like every

01:09:50   other image, Jen thing I've seen. It, it, it seemed like the

01:09:55   one thing in all of the announcements that didn't Apple

01:10:00   wasn't bringing anything to the table in my opinion. Yeah. No.

01:10:03   And I actually think that it looked worse than many services.

01:10:08   It looked like several versions, old of Dolly. And part of that

01:10:11   is because of, and it was an intentional decision, this kind

01:10:15   of cartoonish art style they wanted. They don't want photo

01:10:18   realism, but on the same hand, we didn't look at those and go,

01:10:23   Oh, this is Pixar level. You're like, yeah, that's a, that's a

01:10:26   weird AI image. And even in some of the images they showed in the

01:10:29   keynote, which are obviously handpicked. I mean, there's a

01:10:32   probably the best that they produce. There's still weird

01:10:35   stuff. If you look close at them, there's like weird finger

01:10:37   stuff going on and hair that goes through parts of the torso

01:10:41   at some point. I mean, it's still very much a high image.

01:10:45   And there's a picture of a woman that they've used in

01:10:48   several, I might even be in the newsroom post. It's like a white

01:10:52   woman probably in her in real life, forties or fifties. And

01:10:56   it's kind of unflattering to her. It's, she's got crows feet

01:11:00   around her eyes. It's almost the opposite of what you would want,

01:11:04   right? You kind of, if you're the boardwalk caricaturist,

01:11:08   you're supposed to make the person look exaggerated. Like

01:11:11   they're going to give me a big nose because I have a big nose,

01:11:14   but not in an unflattering way, right? That's the trick. It's,

01:11:18   Oh, we'll give the John's caricature is going to have a

01:11:21   real big nose, but we're still going to make them look pretty

01:11:23   good. And a little younger than he really is, right? You don't

01:11:26   make somebody look older. Yeah. Like I didn't wear your eye

01:11:31   cream the night before. Yeah. And you wait, that's, there's

01:11:33   going to be controversy behind that because they're just

01:11:36   grappling with the stuff that everybody's grappled with. So

01:11:38   even within these, these walls and parameters that they've

01:11:40   set of, we're not doing photo realism, we don't, there's

01:11:44   definitely going to be a bunch of keywords that it just

01:11:46   refuses to generate images on, but you wait, I mean, it's

01:11:49   going to mess up people's skin tones and it's gonna, it's just

01:11:53   going to do the stuff that AI images do. And Apple's not

01:11:56   immune to that. And proof of that is kind of the images they

01:12:00   demoed, which again, like you said, we're awesome. Yeah.

01:12:03   Yeah. And with the words, the language stuff, they're not

01:12:08   letting you prompt, right there. So you don't get to say,

01:12:13   write me a cover letter for a job I want to apply for. You

01:12:17   write a cover letter and then you select the text and say,

01:12:20   make this sound more professional or proofread it, right?

01:12:23   I'm real, I personally, I'm very intrigued by the proofreading.

01:12:26   I don't, I, I'm stubborn and I think of myself very well as a

01:12:33   writer, so I'm not really looking to have my own prose

01:12:36   rewritten in a different tone, but the proofreading I would

01:12:41   love if it works well. I certainly appreciate, I have long

01:12:44   appreciated spell checking, automatic red underlines of

01:12:48   Save My Bacon, I was going to say thousands, probably tens of

01:12:52   thousands of times since I've been writing Daring Fireball.

01:12:56   So I would love the proofreading part as a one person show who

01:12:59   does his own proofreading, but you don't get to say, write me a

01:13:03   cover letter or write me a very inappropriate story about X, Y,

01:13:09   and Z. You don't get to say that. So what they're saying is

01:13:13   they don't, and again, what they're saying to me is the

01:13:16   exact attitude I would like Apple to have, which is we're

01:13:19   not going to give you a full prompt where you can ask for

01:13:24   inappropriate stuff, but if you write something yourself and

01:13:28   ask the system to proofread it or to change the tone, it will.

01:13:34   So if you write an inappropriate story, an R-rated or X-rated

01:13:39   story, select it and ask for it to be made funnier, it should

01:13:43   still do it. That's what they're saying, which I think is the

01:13:47   right way to go, in the same way that you could always just

01:13:51   open text edit and use text edit to create the instructions for

01:13:56   making a homemade pipe bomb, right?

01:13:58   Pete: Sure, totally.

01:13:59   Pete; You know, though, that there's going to be stories,

01:14:01   you know. I wrote an X-rated story about a kindergarten

01:14:06   teacher and look what Apple's AI engine helped me polish it

01:14:11   up, right? It's kind of on you, but with the image playgrounds,

01:14:14   it is just a prompt. I don't know.

01:14:17   Pete; You can type anything in there.

01:14:19   Pete; So I feel like, I don't know, again, Apple doesn't talk

01:14:23   about their internal debates, but I have to think that that

01:14:26   was sort of a, hmm, do we want to, do we want to even do this

01:14:31   and let alone announce it now or not?

01:14:32   Pete; Yeah, and it'll be interesting to see what happens

01:14:37   because there's definitely things that you can, an example I

01:14:42   used in one of my videos is, and this is much of what's hard

01:14:46   is, again, understanding the intent of the language. So

01:14:49   putting into image playgrounds, emoji with Hitler mustache is

01:14:53   probably going to result in, yeah, no, we're not doing that.

01:14:56   But if you did emoji, little black square above lip, is that

01:15:01   something that the system will recognize and say, no, no,

01:15:04   we're not doing that. Or will it just go ahead and, it's going

01:15:06   to be fascinating and you wait, there are going to be people

01:15:09   that are going to try to break it in any way they can to get

01:15:12   it to do stuff that, you know, but…

01:15:14   Pete; I forget the comedian, but there's a comedian who has

01:15:17   a great bit. It might be John Mulaney, but it doesn't matter,

01:15:21   but it's about the Michael Jordan Hanes commercials from

01:15:25   like the 90s where Jordan had a, quite frankly, sort of

01:15:29   Hitler-esque mustache. And they're like, Michael Jordan was

01:15:32   so big that Hanes didn't even say anything when he was

01:15:35   supporting. So I can see like, here's a picture of Michael

01:15:38   Jordan from 1997 Hane commercial with, give me an emoji

01:15:43   with a mustache like his here. And now you've backwards

01:15:46   engineered your prompt way into a Hitler mustache.

01:15:49   Pete; Right. So that's where they're going to have to battle

01:15:52   with stuff. And I think, and this is one thing that I believe

01:15:57   Craig said on your show was that they had established a bunch

01:16:01   of kind of principles of what are we going to let users do?

01:16:05   What's the system going to do? And one of the things that

01:16:08   they really had a big focus on according to them was that they

01:16:11   were going to respect the user's intent. Isn't that the way

01:16:14   they phrased it or something? Which, I mean, that's a big

01:16:17   deal. And that would suggest that maybe it will do stuff

01:16:21   that is unsavory if you are unsavory and that's what you

01:16:25   intend for it to do. But yeah.

01:16:27   Pete; Well, and then it raises other uncomfortable questions

01:16:31   for Apple because when you think of inappropriate topics,

01:16:34   you obviously think of stuff that would get a movie to be R

01:16:37   rated or X rated. You think of violence, you think of

01:16:41   sexuality, but then what about writing about Tiananmen Square?

01:16:46   It would, to me, be absolutely outrageous if they give you,

01:16:51   "Sorry, I can't help with that," outside mainland China for

01:16:56   asking for help writing a report on the Tiananmen Square

01:17:01   massacre. Obviously, though, if,

01:17:04   and when, Apple intelligence comes to mainland China, you're

01:17:07   obviously not going to get help asking for rewrite proofreading

01:17:13   help on a report about Tiananmen Square, right? There's

01:17:16   other topics like that that are sensitive to Apple in ways

01:17:22   because of their business interests in China in particular

01:17:25   that presumably that they're not going to, you're not going

01:17:29   to notice that as somebody who isn't in mainland China, but

01:17:33   it's something I look forward to trying as soon as I get my

01:17:36   hands on this?

01:17:36   Yeah, it'll be interesting. That's just the whole

01:17:41   boondoggle of AI, I suppose.

01:17:42   All right, let me take another break here and thank our

01:17:45   second sponsor of the show, and it is our good friends at

01:17:49   Squarespace. You can go, I'll spoil the ending right here,

01:17:53   go to squarespace.com/talkshow and you can save 10% off your

01:17:58   first purchase of a website or domain. Look, what is

01:18:01   Squarespace? Squarespace is the all in one service for

01:18:07   building your presence online. Everything from domain name

01:18:12   registration to templates to choose from for designing your

01:18:16   website to adding features to your website, hosting a

01:18:20   podcast, hosting a blog, selling stuff on online, whether

01:18:24   it's like t-shirts or other goods that you ship to people or

01:18:28   selling your time as a professional, like a consultant

01:18:31   or a trainer or something like that. You can do it all on

01:18:34   Squarespace. They handle everything from the commerce to

01:18:37   the SEO to the analytics to see how people are getting to

01:18:42   your website. They've got checkouts that are made as

01:18:46   seamless as possible, powerful payment tools. You can accept

01:18:49   everything from credit cards to PayPal, Apple Pay, of

01:18:52   course. And in eligible countries, you can even offer

01:18:55   stuff to buy now, pay later with features like Afterpay

01:18:59   and ClearPay. The Fluid Engine is their next generation

01:19:03   content website editor. I still call it new. It's from within

01:19:07   the last year, but it's a major rewrite of the system that

01:19:11   lets you design your website on even on your phone. You can

01:19:15   actually just go on your phone, log into Squarespace and do

01:19:18   the drag and drop stuff to rejigger your website right on

01:19:20   the phone. It all works great. Email campaigns, they've got

01:19:24   that built into the system too. So for email marketing tools

01:19:27   or for setting up your own newsletter, you can do it all

01:19:29   through Squarespace. If you can do it on the internet, you can

01:19:32   do it through Squarespace. Go to squarespace.com/talkshow.

01:19:36   You get a free trial for 30 days, no watermarks, no limits,

01:19:40   just the full Squarespace experience for 30 days. And

01:19:44   when you're ready to launch, just remember that same

01:19:46   domain, squarespace.com/talkshow, and you will save 10%

01:19:50   off your first purchase of a website or domain. My thanks to

01:19:54   Squarespace. Maybe lightning round, dig through the

01:19:58   announcements for the platforms, which I think it's fair to say

01:20:03   and Adam, this isn't even really a complaint, but I think it

01:20:07   kind of shows that the company's focus is very largely diverted

01:20:12   to Apple intelligence this year, that the non-Apple

01:20:16   intelligence features are thinner than most years. But

01:20:22   again, I don't know that that's a complaint. I feel like maybe

01:20:24   some years they sort of pile on tent pole features just for the

01:20:31   sake of having something to put in the keynote. I don't know.

01:20:34   Yeah. Some of them, I mean, I don't know how numerous they

01:20:38   are, but it feels like at least a few of them have been things

01:20:41   that people have wanted forever. So that way it was kind

01:20:45   of a big year, like home screen customization. Yeah. Which is

01:20:48   what they led with, right? They led with that. Yeah. Control

01:20:52   center. I think that's another big change. Control center has

01:20:55   been a pile of garbage to put it nicely for quite some time and

01:21:00   this new one promises to be much better. So. My theory with

01:21:06   control center is that I feel like the iPhone evolved in a

01:21:12   kind of curious way that I that I at least didn't notice until

01:21:20   recently, but in hindsight, it sort of feels like, huh. It was

01:21:24   sort of a slow boil there. I can't say that there was one

01:21:27   particular year where it switched, but the iPhone evolved

01:21:32   from when it debuted even Apple thinking of it like pre iPhone

01:21:39   phones where a phone is more like it's not part of your

01:21:42   personal computer life and you don't think about locking it

01:21:47   down. I mean, famously Steve Jobs spent tons of time and

01:21:52   everybody loved it. Just the slide to unlock right. It's

01:21:56   like this beautiful. It was a beautiful metal looking 3D

01:22:01   thing and it tracked the little thing that you slid to unlock

01:22:04   tracked your finger perfectly, but there was no passcode

01:22:09   right. II think even in the original iPhone, you could set a

01:22:14   passcode just because even yeah, even dumb phones used to

01:22:17   let you do that too right. You could just buy like a flip

01:22:20   phone and have it so that you have to type a code or

01:22:22   something to be able to use it, but it's clearly some default

01:22:26   behavior right and the assumption was that most people

01:22:30   would just not put a code on their phone and just slide to

01:22:33   unlock. That's how I carried my phone for a few years at least

01:22:36   and then it sort of dawned on everybody like hey all my

01:22:40   emails in here. I'm logged into my bank and I guess most

01:22:44   banking apps would make you put a pin in before touch ID and

01:22:47   face ID, but there was a lot of stuff in your phone that you

01:22:50   wouldn't want a stranger to have access to just by sliding

01:22:54   a thing, but we kind of lost something by locking everything

01:23:01   into face ID and everything and that there are things you just

01:23:04   kind of want to tap into your phone and just quick do and

01:23:07   that's sort of what control center is now. It's sort of like

01:23:10   the thin private version of the iPhone experience that oh, you

01:23:16   don't have to worry about locking people out of this. It's

01:23:17   just right there. Just swipe down from the top and you can

01:23:20   control your TV and change the volume, change the brightness,

01:23:24   stuff like that. It's you know as opposed to, I don't know,

01:23:30   it's just like conceptually heavier to go into your phone

01:23:35   and start launching apps to do things. It's hey, let's just do

01:23:38   this. The homestream customization thing. I think this

01:23:42   is great. I don't plan to make heavy use of it. I don't plan to

01:23:48   tint all of my icons in a certain color, but I totally

01:23:53   understand why people do and it's I mentioned it on stage

01:23:56   during my show like when I first got a Mac in college in the

01:24:00   90s, I spent an inordinate amount of time in ResEdit just

01:24:03   customizing the icons for my apps and folders just for the

01:24:06   same reason. I just I don't know. I wanted to make my own

01:24:10   icons and even though I'm not a very good icon designer, it was

01:24:13   kind of cool to know that I did it myself and but I see people

01:24:19   that it's always Steve Jobs would never right. Steve Jobs

01:24:24   would never have allowed people to make their home screen ugly

01:24:27   by tinting every single icon magenta. What I don't know that

01:24:34   he wouldn't have because that it doesn't happen to anybody by

01:24:38   accident, right? It's not oh if you shake your phone, it's like

01:24:42   a slot machine and you know maybe you'll get a magenta

01:24:47   home screen. You have to do that on purpose and if that's

01:24:50   what you want to do, okay, so be it. So I think the complaints

01:24:54   that that it allows you to make your home screen ugly to me.

01:25:00   That's well then don't do it right. If you don't think

01:25:03   magenta all magenta icons are ugly, then now you can do it.

01:25:06   Yeah, one one parallel that made me think of it around the

01:25:11   same time as WWDC. I had gone to the unveil event of the new

01:25:17   generation of Rivian R1 vehicles and as part of the new

01:25:21   generation of cars and they're bringing this retroactively.

01:25:25   There's just ambient lighting that's been RGB controllable for

01:25:28   years, but it's never been user controllable and in this new

01:25:32   version, unlike Tesla where you can make it any garish color

01:25:35   you they only have. I think it's nine colors that are

01:25:38   acceptable to Rivian that you can select and so it's a very

01:25:41   specific shade of green and a very muted hue of blue and and

01:25:47   the designer said it's because we don't want people making

01:25:50   their cars pink and on the one hand there was a laugh but on

01:25:53   the other hand I was like if you're gonna let people

01:25:56   customize it, let them do whatever they want and so that's

01:25:58   kind of where I'm thinking here. If Apple were to say, oh,

01:26:01   here's the colors or here's the formats in which you can

01:26:03   customize your phone. Well, then it's not really customization

01:26:06   is it so I think this is the only way they could have done

01:26:10   it that I think this is a minimum basically of letting

01:26:14   people make it however they want. It's still fairly limited

01:26:18   in the grand scheme of things. You can't change the whole icon

01:26:20   you know shape in the way that you could with ResEdit.

01:26:25   Technically, there are still loopholes you can use through

01:26:27   Widget Smith and theme packs and using shortcuts and all

01:26:31   that stuff, but that's not a system level thing. That's still

01:26:33   a weird kind of hack around way to do it, but as a base level

01:26:37   customization is about as minimal as it gets and yeah,

01:26:41   people are gonna make it ugly, but you gotta let them do it or

01:26:44   let them not do it.

01:26:45   And I the demand is clearly there the success of Widget

01:26:50   Smith, the viral sensation of those TikTok videos from two

01:26:54   years ago or so when it first became possible to make

01:26:57   shortcuts with a custom icon that would launch the app so

01:27:02   that you more or less like creating an alias to the app so

01:27:05   that you could give it a custom icon. But what a huge advantage

01:27:09   because with the old way of creating a shortcut with a

01:27:12   custom icon for every app you use on your even if you only

01:27:15   did your first home screen, you've got to set up 16 of them,

01:27:19   put them in the right order. And if you get tired of the

01:27:22   black and white monochromatic look and you really want to go

01:27:26   magenta now you've got to do it all over again. Whereas this

01:27:29   way you just bring up the slider and say switch from black

01:27:32   and white to cyan or whatever color suits your mood for the

01:27:36   day. I think that's pretty cool. One of the demos and I

01:27:39   again I often even if the demo isn't particularly interesting,

01:27:44   I always think it's an interesting insight into what

01:27:48   Apple thinks is important or most used and the press hands

01:27:54   on thing for watch OS 10. What number are we up to? Is it 10

01:27:58   or 11?

01:27:59   11 I think. 11. Yeah it'll be the 10th watch and 11th OS or

01:28:03   something. But it was pretty much, well maybe not entirely.

01:28:07   They did a lot of stuff with the vitals app too, but they've

01:28:10   really were focusing on the photos watch face, which is the

01:28:14   one where it picks one of your or you get to pick your photos

01:28:18   and make a watch face based on your kids or your spouse or

01:28:23   whatever pictures from your own library you want to use as a

01:28:27   watch face. Clearly this is a super popular watch face. I

01:28:30   mean they said so. They said this is like one of the most

01:28:33   popular watch faces. And it does, I mean again we didn't get

01:28:39   to do it hands on, but with the, and I don't think they're

01:28:43   calling it Apple intelligence, but they're definitely using

01:28:45   machine learning. So it's coming to all of them. Right. And I

01:28:48   guess the watch doesn't count for Apple intelligence because

01:28:50   it doesn't have a A series or M series chip. But using machine

01:28:56   learning to pick photos that are particularly well suited to

01:29:00   the aspect ratio of the watch, and then using the machine

01:29:05   learning to pick where in the frame would it be good to put

01:29:08   the digital numerals like behind the head because the

01:29:11   person has space on the right, or if with another picture,

01:29:14   it's on the left. You get all of this assistance and it was

01:29:17   Apple's demo, so it did work well, but it really was sort of

01:29:21   like having a graphic designer helping you. Oh, this is great.

01:29:24   And we could even put like the part of the minutes behind your

01:29:28   head because there's space for that and we have the depth in

01:29:31   the photo to do it. But if you want to customize it your own

01:29:35   way and make it ugly, you're still free to. It's a good mix

01:29:39   where automatically it tries to use machine learning to give

01:29:42   you an attractive Apple style version of your photos. But if

01:29:46   you want to make it garish, it's your watch. And that's the

01:29:51   way it should be. Try to encourage you to make it look

01:29:55   good. Allow you to make it look ugly. Yep. 100%. Let me see. I

01:30:01   thought that the feature and again, it's one of those things

01:30:05   where I would love to be in the meeting where they decided how

01:30:09   to demo it. The hide an app from your phone feature and what

01:30:15   was their example? It was Craig saying like a hairdryer app or

01:30:19   something like that. Like a USB. It's like an AR hair app or

01:30:23   something. I think. Right. Or hair kit. I don't remember.

01:30:27   Because everybody's mind goes to apps that that somebody,

01:30:31   whether you yourself might have apps you'd want to hide for

01:30:34   whatever reason. And you know, I'm laughing but it's you know,

01:30:38   could be there's all sorts of things. It could be something

01:30:43   that's not funny at all like a help managing an addiction or

01:30:48   something like that right. It's not just about like married

01:30:52   people with the dating apps or something like that that you'd

01:30:56   want to hide from your spouse, but it could be something like

01:30:58   a medical type thing. I thought that was pretty clever and it.

01:31:05   They showed it. I feel like in the keynote people were like

01:31:08   well, but if they put it in a folder called hidden apps,

01:31:12   isn't that where everybody will go? But the key is that the

01:31:16   hidden apps folder has a generic icon that doesn't show what's

01:31:22   inside it. So yes, you'll be able to go to somebody's phone

01:31:25   and and in their app library, there is going to be a hidden

01:31:29   apps folder, but you won't be able to get into it because

01:31:32   your face won't do the face ID unlock and the icon for hidden

01:31:37   apps isn't going to give you a hint as to which apps are

01:31:39   inside hidden. It's you know. Moreover, it's it's there by

01:31:42   default. So right. And if you don't have anything in there,

01:31:45   it's still there. Yeah, exactly. So you you you don't know

01:31:48   what's in there. It's sort of like I think it's generally

01:31:51   considered best practice now, but if if like you go to your

01:31:55   account page on a site and it's it says email, here's your

01:31:59   email address password in it. It might show bullets, but the

01:32:04   bullets don't correspond to the length of your password like it

01:32:08   always shows eight bullets whether your passwords 20

01:32:10   characters long or thirteen characters long. It just shows

01:32:14   eight bullets so that it doesn't give you a hint what's in

01:32:16   there right. Yeah. Seems like that sort of thing, but a cool

01:32:20   feature. The mail categorization seems a little late. Frankly

01:32:27   given the big big time big time if they can use Apple

01:32:31   intelligence to do it more intelligently, then that's

01:32:34   that's cool. I think the surfacing of kind of like

01:32:38   contextually relevant stuff is kind of neat. So like yeah when

01:32:41   it knows that you're on you're headed to the airport or it's

01:32:44   the day of your flight. It resurfaces those emails. That's

01:32:47   cool. But yeah, the basic categorization thing. I mean

01:32:50   welcome to mailbox from 2013 or whenever that was. Alright.

01:32:55   Wasn't that that was the app Google bought everybody loved

01:32:58   it and then it just disappeared. I think so right that was yeah

01:33:01   inbox and then yeah the first one was bought by Dropbox. I

01:33:05   think that was the one that popularized it was that one.

01:33:08   Oh, I forget email. I don't remember mailbox. Yeah. So

01:33:14   Dropbox purchased it back in 2014. Never to be heard of

01:33:20   again. Yep killed it a year later. Yeah. I'd never thought

01:33:24   of it, but I do have to say it's one of those ideas where

01:33:27   once I heard Fedor or whoever said it on the in the keynote,

01:33:31   I was like, oh, that's brilliant. I wish I've had that

01:33:34   for 15 years is in the list of emails. Don't just show me the

01:33:38   first two lines of the email. Show me a two line summary of

01:33:42   the email because so many emails the first two lines

01:33:46   aren't helpful at all. It's so true. I've been irritated by

01:33:51   that, but it never occurred to me that it's a problem that

01:33:53   could be solved. I'm very much looking forward to that. I'm

01:33:58   always a little surprised when Apple Mail is late to get

01:34:01   certain features simply because I know how much Apple

01:34:05   institutionally runs on mail. They are a very very heavily

01:34:09   email driven company internally. So when and they all

01:34:14   do use Apple Mail, it's always surprising to me when things

01:34:19   like this are late. I love using my main mail is still on

01:34:23   Gmail and the main reason why is that using Mime stream on

01:34:28   my Mac, I get product or categorization of the email and

01:34:32   it is such a game changer to have all your sales receipts

01:34:38   and all the **** you get every time you buy something and

01:34:41   even if you uncheck the box to send you promotional emails,

01:34:44   they still send you some. It's like like send me no

01:34:48   promotional emails means in the in modern ecommerce. Okay,

01:34:53   maybe we'll send you fewer promotional emails, but having

01:34:56   them out of my main inbox so that my main inbox is really

01:34:59   just email from people emailing me is like it's like time

01:35:04   travel to when I first got email in the 90s and email was

01:35:08   nothing but joy like oh, I got new email. What is it? What I

01:35:12   get? It's like getting a present every day. I love it,

01:35:15   but it's hard hard not to say finally to that one. Let me see

01:35:20   here from my notes. What else anything else stood out to you?

01:35:23   I guess we should do photos on that's a big deal. Yeah. Have

01:35:27   you looked at it in beta? I have of course it's the

01:35:31   developer beta under NDA so I haven't used it, but if I had

01:35:33   used it, I might say that you do. It's a it's an adjustment,

01:35:39   but I think it's pretty well laid out. There will be some

01:35:42   people that will be irritated if they're heavy albums users

01:35:45   cuz they've moved albums that are user created to basically

01:35:48   the very bottom of the page and there's there's some weird

01:35:52   stuff like for example when you open up the app half of the

01:35:55   display shows the typical grid you're used to and half of it

01:35:58   shows the smart albums. The idea being that they're more in

01:36:01   your face. You're more privy to actually look at them, but

01:36:05   there's weird kind of UX stuff. It's a little strange. So if

01:36:08   you go to select multiple photos, you can't do that when

01:36:12   half of the grid is on the screen. You have to actually

01:36:14   pull the grid down so that it's all photos. Then the select

01:36:17   tool comes up and that might change over the beta. There's

01:36:20   kind of weird stuff like that, but in general, I think it's a

01:36:24   lot better. Those smart albums have been there for years and

01:36:28   they're pretty cool. I mean that was one thing that you

01:36:30   know, Craig. I think it maybe it was even on your show was

01:36:32   like well. we added the pets and people one last year. Oh

01:36:35   yeah, but a lot of people a lot of people think these are all

01:36:37   new. They're not. I mean they've been there for a long

01:36:39   time. We're just adding more and more of them and getting

01:36:41   better and then I'm really interested in the optics of

01:36:47   improved search because that's something that when you have

01:36:50   large libraries, it's just miserable. Apple's always done

01:36:55   that well not always, but over the last several years, they've

01:36:57   been better at OCR stuff works really well just because of

01:37:00   that being so capable on device, but object recognition.

01:37:04   It's not bad, but you open something like Google photos

01:37:07   who trains probably by stealing all of your photos and and and

01:37:11   it's really good. So when I need to actually find something

01:37:14   I'll open it in Google photos rather than Apple photos and

01:37:17   hopefully that'll change to the point where if the Apple

01:37:21   intelligence stuff is to be believed, you can say Megan

01:37:24   when she was wearing the spotted dress on Tuesday back

01:37:27   in Rome get that hyper specific and that will be great. We're

01:37:31   not just a green sweater, but a green fuzzy sweater. Oh my god

01:37:34   that it's actually the picture I was thinking of. I'll say that

01:37:37   that's the sort of feature that I continue even before WWDC

01:37:41   just in regular life and talking to family members and

01:37:45   stuff. People have no idea that you can search Apple photos for

01:37:50   stuff like that. They really don't that you can and and that

01:37:53   it works pretty well for a lot of things certainly works well

01:37:56   for people once you've tagged some people, but that you can

01:37:59   do things like say a cheeseburger and if you have got

01:38:04   some pictures in your library of like cheeseburgers, it'll

01:38:09   find them and now whether it finds every single one. I think

01:38:12   Apple is not at the state of the art of that, but it's pretty

01:38:15   useful and when I do it, it's really it blows people's

01:38:19   minds if they're not aware that you could search for stuff like

01:38:22   that in photos and I definitely think they're doing the right

01:38:25   thing because I probably organize my photos a lot more

01:38:31   than the typical user does, but my photos in my library are

01:38:35   largely uncategorized, unalbumed and it's really just a

01:38:40   big pile. So doing more stuff automatically is going to be a

01:38:45   help and I and I agree though that by pushing your manually

01:38:49   organized folders or albums, whatever you call them down, it

01:38:53   is a small price of convenience cost to the diligent users who

01:38:59   really do dutifully put all of their photos into albums

01:39:03   manually as they shoot them. Yeah. Well, I think it's a

01:39:06   really good way to surface features like you just said one

01:39:10   of the problems that I think iOS and iPad OS in particular

01:39:13   have had over the last few years is that they have become

01:39:15   so hyper capable, but on the surface, they still wanna be

01:39:19   approachable and simplistic to everyone that however comes at

01:39:22   the expense of people that aren't watching the keynotes

01:39:25   and reading your blog and other people not knowing that things

01:39:28   exist. So one example is when you open up the photos app and

01:39:32   hit the search button now before it was just a random field

01:39:35   that you could enter text into and like you just said people

01:39:38   didn't know you could search for a cheeseburger or for a dog

01:39:41   and now it actually shows you suggestions, which is great. So

01:39:45   it says like solely who's my cat in 2023 last month with

01:39:50   Campbell who's my brother. That is a it's just a really simple

01:39:53   visual cue that lets people know. Hey, you can type

01:39:57   whatever you want in here and we'll try to find it and I think

01:40:00   iOS this year in particular got a lot more of those subtle kind

01:40:05   of pokes of hey, this is a thing that you might not be able

01:40:08   to know that you can do, but actually you can. Yeah, it's

01:40:11   cool. Exposing and I think that's the sort of that's the

01:40:14   whole thinking behind this photos redesign is you know,

01:40:17   let's surface these smart albums and and make a

01:40:22   collection from your nephew's birthday party last weekend.

01:40:26   Let's just assume that that that's an event you'd

01:40:28   categorized into an album and push it to you. So you see oh

01:40:32   here here are all the photos I have from that, including the

01:40:35   ones shot by other family members that I downloaded from

01:40:39   messages when they shared them to me and messages and put them

01:40:41   in there very very cool. I love I have to mention this as one

01:40:46   of my absolute top features of the year is the whole I guess

01:40:51   they're calling math notes, but it's more than just in the

01:40:55   notes app and then the calculator app that's finally on

01:40:59   on the iPad, but you can type math anywhere in a text field

01:41:04   and get the answer. So in messages you could type 96

01:41:09   divided by twelve equals and it'll auto complete eight as

01:41:15   the answer and it's just like auto completing a word. You just

01:41:19   hit the space bar to get it and the whole math that simple

01:41:23   whether it's simple arithmetic or more complex notation like

01:41:27   they showed in the notes app. You can do it and that's super

01:41:31   cool cuz that's something the computers. It's like the the

01:41:34   first thing in the 60s when people heard about. Hey, you

01:41:38   ever hear of these things called computers. They're really

01:41:40   they're good at math. It's kinda awesome and it kinda

01:41:43   makes you wonder why you haven't been able to again.

01:41:45   Again. It's one of these features where it's like why

01:41:47   haven't we been able to do this forever right computers are

01:41:50   really good at math. Yeah. No, it's super handy and it's

01:41:54   available on all platforms too. That's not just an iPad thing,

01:41:56   which is great and the calculator app in general has

01:42:00   gotten huge improvements. Now you can see the actual the the

01:42:04   math that you typed out before you hit the equals button. It'll

01:42:08   show you the solution. There's now a log of all of the stuff

01:42:11   you've entered. It's just oh, it's a way better. So that's

01:42:14   good long and I'm. Longtime friends with James Thompson,

01:42:19   author of Peacock and I think Peacock will be fine as a truly

01:42:25   excellent calculator. That's always going to go way further

01:42:29   than Apple's ever willing to go in terms of customization.

01:42:32   It's always mixed feelings, though when you're friends with

01:42:35   somebody who's been Sherlock even to a minor degree right.

01:42:39   We can all complain that it seems very strange that the

01:42:44   iPad went 13 years without a built-in calculator app. It's

01:42:49   really does seem curious, but if you make a calculator app

01:42:53   that kind of was good news for a long time, but I'll just say

01:42:58   I've I ran into this issue two mother's days ago. I don't

01:43:04   know it wasn't this year. So it's either last year or the

01:43:07   year before, but we got my wife's mom a new iPad and. She

01:43:14   didn't need a lot of help setting up. She had an iPad

01:43:16   before she just upgraded from the old one, but she was when

01:43:20   you get a new device is when you ask questions of your son

01:43:24   in law of things. Hey she she not knowing that this is like a

01:43:30   meme in our circles. Why doesn't the iPad have a

01:43:32   calculator? She asked me why my phone my iPhone has a

01:43:35   calculator and she uses it all the time. She does tips and

01:43:39   people use calculators a lot. I asked about an iPad calculator

01:43:44   app and I looked at the app store for calculator apps and I

01:43:49   eventually send her to the free version of but there are I

01:43:55   don't know 150 million calculator apps in the app

01:43:59   store. It's there's a lot of them and none of them make me

01:44:04   feel good. There are so bad. There are some of them are

01:44:09   really bad. They all have ads because it's this race to the

01:44:13   bottom and because there really are a zillion that are free

01:44:17   with ads. There's not a lot of room for a paid version and my

01:44:22   mother-in-law isn't thinking of a calculator app is the thing

01:44:25   she should pay three or $4 for and I know again. I'm friends

01:44:29   with James. I pay for P Calc happily so, but it's people

01:44:34   don't think of a calculator is a thing they should pay for so

01:44:37   it's it is good, but it's also really interesting that they

01:44:42   did what Fedor said for years. We don't wanna bring a

01:44:45   calculator to the iPad unless we can bring something new to

01:44:48   it. They brought a lot new to it. It's really interesting.

01:44:51   They did if you will permit me to complain about one thing

01:44:54   that I noticed in trying to search for a good scanning app.

01:44:58   I like I know you can scan documents in iOS and it's fine.

01:45:02   What I don't love about it is that you've gotta do it a weird

01:45:07   way, not a way that goes with my flow. I would much prefer to

01:45:10   open a camera effectively that scans you know a photo into my

01:45:14   image library. Normally you have to go through notes and

01:45:17   create a note and then once you get into notes, you can say hit

01:45:20   the little share sheet and then you can send it to files. You

01:45:22   can't save it to your photo library. I get it. I understand

01:45:25   the flow. They don't wanna jam up your photo library with but

01:45:28   sometimes I just want an image of a of a scan. One thing that

01:45:33   is been very frustrating and I think it's maybe why you

01:45:36   weren't able to find a good free calculator app because I

01:45:38   have no doubt that there is one that's out there and doesn't

01:45:40   have ads but the app store search is just so abominable

01:45:44   because now it's all ranked searching based on who pays for

01:45:47   surfacing all that stuff and keyword jamming and so it's

01:45:52   it's a really unfortunate that the discoverability on the app

01:45:55   store is so poor that it you would be hard pressed to find a

01:45:59   free calculator app without in-app purchases and without

01:46:02   ads but I guarantee there's one out there. You just never find

01:46:05   it because the app store search sucks. And you can't you can't

01:46:09   search with the natural language. No calculator app

01:46:13   without ads and without in-app purchases. That's what we need

01:46:15   Apple intelligence for the app store. Alright. Alright. Let me

01:46:19   blow your mind, Quinn. You can go to the files app. On your

01:46:24   phone and go up to the little dot dot dot menu and the second

01:46:28   item down or third item down scan documents. Hey, there you

01:46:33   go. That does. I just learned that about 2 weeks ago on

01:46:36   threads where somebody was complaining. Somebody was

01:46:39   complaining about the exact same thing you just mentioned. I

01:46:41   only ever did it in the notes app and it seems very hard to

01:46:44   get them out of the notes app. It does. But you can actually

01:46:48   do it right in the files app and so here I am telling you

01:46:51   how good the files app is. There you go. You still can't

01:46:55   save it as an image without saving it too as a PDF in your

01:46:59   file browser and then so but that is better. I will start

01:47:02   using that. That is much improved. Thank you. Yeah.

01:47:05   There you go. Tip of the day. We might as well. We gotta talk

01:47:08   about the iPad though cuz you've you've it's been a topic

01:47:11   of conversation. You've reviewed the what how would you

01:47:16   summarize your your M4 iPad Pro review? Well, I sang praises of

01:47:21   the hardware. It's the best. I think it's it's seriously the

01:47:25   best device they've ever made full stop. It's incredible.

01:47:29   True as well. Yeah. And it's it's unbelievably good. And I

01:47:33   think that's exact and I think people recognize that even

01:47:37   especially when they see it like when you it's thin. They

01:47:40   keep telling you it's thin. They had that that controversial

01:47:43   ad that the whole point of it was to emphasize that it's so

01:47:47   thin. It's you need a two-ton hydraulic press to make

01:47:50   something that then then you hold it in your hand and look

01:47:53   at the screen and it's wow. This is amazing, but I think

01:47:56   it's the it's this weird. It's the fact that the hardware is

01:48:02   probably the best computing device they've ever made full

01:48:04   stop that makes the people who want iPad OS to be

01:48:07   fundamentally different enraged. Yeah and II am I

01:48:13   wouldn't put myself in that group in the sense that there's

01:48:16   a lot of people that are pushing for the iPad to run

01:48:20   virtualized Mac OS or to dual boot or for there to be a Mac

01:48:24   OS icon. You can open on the iPad and have some sort of file

01:48:28   syncing between the iPad environment and the Mac

01:48:30   environment. I don't want that. I don't even want a kind of

01:48:34   parallel style environment where you can open Mac apps in

01:48:38   the kind of iPad environment without you seeing the Mac

01:48:42   underpinnings. I don't want that either. I am truly a fan of

01:48:45   iPad OS and that's what makes it so frustrating is year after

01:48:50   year. It's like death by a thousand paper cuts. There's

01:48:53   not there's not much that iPad users are asking for, but what

01:48:59   we're asking for just never seems to come and it's just

01:49:03   rudimentary basic stuff with the files app being fairly

01:49:07   hyperlimited tagging or being able to change the extension,

01:49:10   which you can't do of any file name. Just basic stuff that any

01:49:15   file browser since forever has been able to do and then you

01:49:19   augment that with limitations in multitasking because of

01:49:23   stage manager and they're just artificial limitations. I mean

01:49:27   this thing's running an M4. So one of my favorite examples is

01:49:30   and this even counts when you're using a huge external

01:49:33   display. So I've run my iPad on this 27 inch Apple Studio

01:49:39   display. You can only have four windows open at any given time.

01:49:43   That's the maximum limit and there's tons of room for more,

01:49:47   but you can only have four and if you choose to open up

01:49:50   another window either in a new app or maybe in an app that you

01:49:54   already have open, you just want to open another window. It

01:49:56   will at random just close one of those windows and the only

01:50:00   way to get it back is to resummon it through the app

01:50:03   environment like it's just bananas and so it's just small

01:50:08   stuff like this that individually it's not that big

01:50:12   of a deal, but in aggregate it really makes for true

01:50:16   limitation and this is not limitation in the sense of I

01:50:19   want the iPad to be a Mac. It's limitation in the sense of I

01:50:22   want an iPad to do a thing that I know the iPad should be able

01:50:25   to do, but it can't and they don't need to change. I I iPad

01:50:29   OS. They don't need to make it feel like a Mac. They just need

01:50:32   to add some of this basic stuff that frankly should have been

01:50:36   there for a decade now. Yeah, you posted last month. I guess

01:50:41   it was when you were first testing the new iPad Pro. It

01:50:46   was like a 3 minute video. You posted the threads. I don't know

01:50:49   if you posted it elsewhere, but just sort of explaining your

01:50:52   frustration as to what ought to be a very simple thing. How do I

01:50:58   get I already have an app a window open for I forget which

01:51:01   app you were using was it notes. I think it may have been

01:51:03   craft, which is a third party notes app. Yeah. Okay. But it's

01:51:07   it's you've got a window. You've got your iPad connected

01:51:09   to a display a studio display and you just want to open

01:51:14   another window for the same app. Yeah. Shouldn't be hard and

01:51:18   it shouldn't be hard and you said you've been using Apple

01:51:21   platforms for 25 years. You're pretty familiar with them. This

01:51:25   really is your job explaining things like this and again. It

01:51:30   comes back to intuit as a verb. You could not intuit how to do

01:51:36   it and I'm watching the video and I've I'm in the same

01:51:39   racket, including the sub racket of complaining about

01:51:43   limitations of iPad OS and I'm sitting here thinking I'm

01:51:47   stumped too and you're like what if I hit these the three

01:51:50   little dots up here and I'm like yeah, that's probably it.

01:51:53   No that wasn't it. I was like. Oh, I thought he had it and

01:51:56   then I remember the first time I watched your video like three

01:51:59   times cuz I was like this is so succinct at explaining the

01:52:05   frustrations that people are talking about and but it's like

01:52:08   the first time I watched it. I remember looking down at the

01:52:12   little bar telling me how far for the video. I'm like. Oh,

01:52:15   he's not even close to finding the answer to this yet right and

01:52:18   I followed it up on threads with my comment was this is

01:52:21   just command end on a Mac right and it is and it is and and one

01:52:26   of the the leading kind of dissenting arguments on that

01:52:30   post was well, you're trying to make this behave like a Mac.

01:52:34   It's not a Mac. Stop trying to treat it like one and I'm like

01:52:36   I'm just trying to open a window man. I'm not trying to

01:52:39   make it like a Mac. I understand that they're

01:52:41   inherently different platforms, but it's not intuitive. I

01:52:44   couldn't give my grandmother an iPad and she would figure out

01:52:48   how to open up a window on a Mac sooner cuz she'd be like

01:52:51   hmm. Let's look at crab. No, it's not there. What about

01:52:54   file? Oh yeah. File new window. There you go and on on iPad OS.

01:52:59   It's some arbitrary. You've gotta drag the icon out of the

01:53:02   dock which yeah, but you've gotta do it only in stage

01:53:05   manager cuz if you do it in the standard mode, then it enters

01:53:07   jiggle mode right. Just all these behaviors that make no

01:53:11   sense and aren't consistent with between themselves. All of

01:53:14   the things convenient or they're not intuitive. It's like a

01:53:17   double whammy right and it it's it's like they've somehow

01:53:23   painted themselves into a corner with some of these

01:53:27   features with the way iPad OS works that I and I feel I don't

01:53:32   know I'm mixing metaphors here, but it's like they've painted

01:53:34   themselves into a corner and blindfolded themselves so that

01:53:38   they can't see that they're painted in a corner and it's.

01:53:43   I had that 3 minute video just exemplifies that there is no

01:53:47   easy solution. It's not oh, I see how the the whole thing

01:53:53   Apple should do is they should just do blank and then Quinn's

01:53:56   problem that he describes goes away. It's it's deeper than

01:54:00   that, but because it's deeper, I feel like collectively Apple's

01:54:04   just pretending it's not a problem right and and then you

01:54:07   have different modal kind of environments that conflict with

01:54:11   one another. so one of the problems that I found out and I

01:54:14   didn't even realize it when I made this video was that the

01:54:17   studio display to which the iPad was outputting runs stage

01:54:21   manager all external displays now on when connected to an

01:54:25   iPad run stage manager, but you can still run the windowing and

01:54:29   multitasking environment on the iPad in either stage manager

01:54:34   or the old split view. So I was running stage manager on the

01:54:38   external display and split view on the iPad. Unbeknownst to me,

01:54:42   they're completely separate from one another and you can go

01:54:45   into control center on the iPad and turn stage manager on and

01:54:49   if you do, then you do get window interactions more

01:54:52   similar to what you'd have on a Mac, but the fact that it

01:54:55   wasn't turned on by default and I didn't even know there was no

01:54:58   obvious indicator that hey you're actually running two

01:55:01   completely different windowing environments on each display.

01:55:05   Right. It was almost like they were incompatible with one

01:55:08   another because they are because split view is this old

01:55:11   archaic windowing interface of yesteryear. That's pretty good

01:55:14   for touch, but not really very good for a keyboard and mouse

01:55:17   and so we bring stage manager to help with that, but let's

01:55:20   not get rid of the old way cuz some people still like

01:55:22   multitasking that way and so it honestly feels kind of like

01:55:25   Windows product where you've got seven or eight old kind of

01:55:31   somewhat deprecated, but not really cuz some people still

01:55:34   use them environments that don't really work with one

01:55:36   another, but kinda sometimes they do, but only in this

01:55:39   specific way. That's not obvious to you and then you

01:55:42   enter you end up with a product that is neither intuitive or

01:55:46   simple to use and also not powerful because they're trying

01:55:51   to keep it simple while actually making it very complex

01:55:55   and so I think there needs to be a fundamental rethinking of

01:55:59   iPad OS. I'm not saying turn it into the Mac at all. I'm just

01:56:02   basically saying let's go back to the core. Go back to the

01:56:04   springboard. We've got a home screen. How can we make this

01:56:08   powerful and capable for the people that want that while

01:56:11   making sure that that's out of the way and simple for grandma

01:56:14   that doesn't intend to use that at all and maybe the answer is

01:56:17   you put some switch and control center. That's like a pro mode

01:56:20   that enters more of a desktop class UI that still looks like

01:56:24   an iPad, but has more windowing features and capability that

01:56:28   you might wanna use if you were interested in that kind of

01:56:31   environment, but they've gotta do something cuz every year

01:56:35   they keep adding a thing of oh and they did it again this

01:56:38   year. They added this new little menu bar thing. I forget

01:56:40   what it's even called, but yeah, the the concept of the

01:56:45   hamburger menu. They're like we know a hamburger menu is not a

01:56:47   really good solution. It's designed for touch and so now

01:56:50   you can put some of your most common menu items that you used

01:56:53   to have to go into the hamburger menu. Now it's in

01:56:55   this kind of floating menu bar, but it's not a menu bar in the

01:56:59   traditional sense and nobody that's using it thinks it's a

01:57:02   menu bar and and and there's not sub options in that. So it's

01:57:06   just another thing that convulates the OS without

01:57:09   really addressing any of the real problems. It's quite

01:57:12   frustrating. Yeah. I I I was just discussing this with the

01:57:19   just by coincidence, but I think it's cuz everybody because

01:57:21   the new iPad pros just came out. Everybody just sort of

01:57:24   resurfaced their frustrations and it's just made it a topic

01:57:28   of conversation right now and until last week, there was hope

01:57:32   I did not think that iPad OS was going to get a major

01:57:36   rethink this year, but the people who really wanted to

01:57:39   were like maybe they've one month after unveiling the new

01:57:43   hardware. They're going to unveil a radical rethinking of

01:57:46   the operating system. I was like don't hold your breath.

01:57:49   But there there was a guy who was at Xerox for a long time in

01:57:53   the 70s and then he was at Apple during like their first

01:57:57   heyday of the original Macintosh years into the 90s

01:57:59   named Larry Tesla and he died just a few years ago, but a

01:58:04   real real Titan in the field of sort of defining the modern

01:58:11   graphical user interface and sort of humanizing computer

01:58:15   interfaces and his big thing was not having modes that modes

01:58:19   were bad. In fact, famously, he even had a custom license plate

01:58:23   called no modes was his license plate in California and. It's

01:58:29   like you can say no modes, but I think if Larry Tesla were

01:58:32   still around even he would admit that there were always

01:58:35   edge cases where okay, here's an edge case where yes, you kind

01:58:38   of have to have a mode for blank, but that is a general

01:58:42   rule of thumb. You should eliminate modes because modes

01:58:46   just aren't intuitive and they're not the way people

01:58:48   think, but the iPad is all about modes like you said that

01:58:53   you can have the iPad itself in the non stage manager mode

01:58:58   while it's connected to a display that isn't stage

01:59:01   manager and you get two entirely different interaction

01:59:06   models between windows on the two at the same time and it's

01:59:10   very difficult to explain how you switch between the two

01:59:13   like why can I make the iPad use either, but I can't on the

01:59:17   studio display the studio display only support stage

01:59:21   manager. Why I don't know. It's just the way it is. Yeah. Yep.

01:59:26   So it's one of those things that again. I'm I'm not asking

01:59:30   for Apple to completely redo iPad OS. I think it's so much

01:59:34   of what makes the iPad great is the way that it's been designed

01:59:39   it like the iPad has been a success for a reason. It's not

01:59:42   a mistake. It's not the issue is that they are trying to

01:59:48   suggest that this is a legitimate pro product. There

01:59:51   are pro apps available for it. They're pushing Final Cut Pro.

01:59:54   This new version of Final Cut Pro is amazing. They have

01:59:56   logic. They're trying to really argue that this is a pro

02:00:00   platform and yet they're hesitant to bring features that

02:00:04   would make it so much more convenient and likely to be

02:00:08   used by pros and so it's one of those you can't really have it

02:00:11   both ways. Either it's a simple limited really good product.

02:00:15   It's great for reading news and watching content and browsing

02:00:18   the web and doing short emails or it's a powerful machine and

02:00:24   it can do all this stuff and they kind of want it to be both

02:00:27   which I think it can be but not without creating and I don't

02:00:31   know if this is a different mode or a different environment

02:00:34   or at least some sort of kind of like bringing some of the

02:00:38   features that you know you can't just say that oh it's

02:00:42   powerful and it can do anything you want but then when you hit

02:00:45   a limitation and you go well it's just an iPad. I mean, what

02:00:47   do you want it to be a Mac? You can't have it both ways. You

02:00:50   gotta pick your lane, right? So, yeah. Well, till next year.

02:00:58   As we say every year, maybe, maybe next year. I can't let

02:01:04   this show just good that we didn't record Friday but I do

02:01:09   want to talk about the Apple's sort of bummer. The other that

02:01:15   shoe that dropped on Friday was the announcement that Apple

02:01:20   Intelligence and iPhone mirroring and the new share

02:01:24   place screen sharing features are not coming to the EU when

02:01:30   they come to the rest of the world. Now, none of those, none

02:01:33   of those features are available to anybody yet but apparently,

02:01:37   it may and literally while you and I are talking here on

02:01:40   Monday afternoon, the betas might be out. Supposedly, the

02:01:43   iPhone mirroring and screen the share play screen sharing are

02:01:47   coming in developer beta twos which are supposed to come out

02:01:50   today should be out by the time everybody's listening to this

02:01:54   but will not be out even to users of the beta operating

02:01:57   systems in the EU and Apple Intelligence is postponed

02:02:02   indefinitely there even though it's also postponed for

02:02:06   everybody else too and why? Well, because of uncertainty

02:02:10   around the DMA. Needless to say, people have thoughts. I'm

02:02:17   curious what your take is. II have a take that I'll I'll

02:02:20   share but I'd like to hear yours first. Yeah, I think it's

02:02:25   difficult because on the one hand, you can argue that

02:02:29   regulation does often push companies to do things that

02:02:34   they wouldn't do otherwise. Apple had shown for a pretty

02:02:38   long time that they were willing to make the experience

02:02:42   of messaging people with Android phones worse such that

02:02:47   they well, I shouldn't say worse. They didn't they didn't

02:02:50   try to make it better. Right. And part of that you could say

02:02:54   was malicious because they wanted more people to buy

02:02:57   iPhones cuz they were guilted into it. The more charitable

02:02:59   thing would be that they just didn't care to spend resources

02:03:02   building out RCS support and there were problems with the way

02:03:05   that Google had implemented their version of RCS that

02:03:08   didn't. It's it it goes on ad nauseum but the reality is is

02:03:12   that regulation does get companies to do things whether

02:03:15   they want to or not within a time frame. The problem is and

02:03:19   this is where I've been not super I've been empathetic I

02:03:25   guess towards Apple in regards to the DMA is and this is the

02:03:30   way that Europe has just chosen to do legislation in some ways.

02:03:33   I think it's better than the way we do it here in the US.

02:03:36   There's not really any strict their guidelines right. So

02:03:40   there's not this is what you can and can do. It's the onus is

02:03:43   on Apple that kind of come to the DMA and say hey or come to

02:03:46   the EU and say hey this is what we think as per the DMA

02:03:50   guidelines fit within the framework of what's expected and

02:03:53   then the EU either goes okay or nope try again you're still not

02:03:57   there and so it's been this kind of teetering back and

02:04:00   forth where it gets problematic is that the more regions that

02:04:06   try to do this the less I don't want to say accommodating but

02:04:13   the more fragmented the iOS and iPad OS and Mac OS experience

02:04:18   is likely to become for people all around the world and I

02:04:22   think that is problematic because once you've shoehorned

02:04:26   yourself into a region that's pretty small that's worth

02:04:29   investing money into they're not Apple's not gonna stop sales

02:04:31   altogether but they're also not gonna continue to upgrade this

02:04:36   weird legislative thing that your government said was

02:04:39   required five years ago and there's just you're gonna have

02:04:42   more bugs you're gonna have less consistency as there's more

02:04:44   binaries that need to be installed when different

02:04:47   devices around the world at different times I empathize

02:04:51   with Apple because it seems incredibly unrealistic to do on

02:04:54   a global scale and the EU I think they're playing ball

02:04:58   because it's such a huge market and they have to comply if

02:05:00   they're gonna sell stuff there and it's a big enough market

02:05:03   that they're willing to do it and some of these things that

02:05:06   they do to comply I think will come by extension to the US and

02:05:10   other parts of the world just because it's too difficult to

02:05:13   not if they have to do it you might as well just bring it

02:05:16   everywhere but it does kind of in some ways seem to be

02:05:20   innovation by consensus which as history would dictate is not

02:05:23   always the best way to do stuff so I think what I've been

02:05:28   thinking about over the weekend is that I think there's a lot of

02:05:32   people who think they are in favor of the DMA who if they

02:05:37   really read it and understood the full implications of it

02:05:41   would realize that it's it's not what they think it is and I

02:05:46   think there's a lot of people nerds people in our collective

02:05:49   audience you remind whose main issues are a couple of things

02:05:54   where they see Apple being dicks that Apple is a dick about

02:05:59   the 30% 15% mandatory cut of the app store and that they're a

02:06:06   dick on the same vector about the whole idea that the only

02:06:10   way to distribute software for iOS is through the app store and

02:06:15   the DMA addresses both of those issues that the both it both

02:06:20   says that you can't mandate that all transactions go through

02:06:26   the gatekeepers payment processing which is where Apple

02:06:29   that's the that is the actual gate where Apple takes their cut

02:06:33   because all transactions have to use their payment processing

02:06:37   here's this gateway where they can take their 30% 15% cut and

02:06:41   the DMA mandates that you can't be the exclusive distribution

02:06:46   channel for software on your gatekeeping platform and in fact

02:06:50   they further clarified since Apple's initial attempt at

02:06:54   compliance where they said okay we'll have alternate

02:06:57   marketplaces and then they went back and forth and said no

02:06:59   you've also got to allow direct sideloading from the web and

02:07:03   there's a lot of people out there listening to the show

02:07:05   listening watching your videos people who are yelling at me on

02:07:08   social networks who see those as the their main complaints

02:07:12   about Apple and what they see a lot of these people see and I'm

02:07:16   not saying they're wrong. I think I disagree to some extent,

02:07:21   but I at least see the argument that they're saying that Apple

02:07:27   itself that this isn't even an Apple anti-Apple sentiment.

02:07:30   This would be good for Apple. Apple's too obstinate for its

02:07:34   own good and opening up would be good for Apple, not just good

02:07:38   for the people who want to distribute apps that aren't

02:07:41   allowed through the app store etcetera that this would be

02:07:43   competition is good for Apple and being insular and

02:07:48   protective in the long run is always detrimental. I get it, but

02:07:52   the DMA is so much more than that. So what I saw on Friday

02:07:56   when Apple announced this is this stuff isn't even about the

02:08:00   DMA. Why in the world is Apple withholding iPhone mirroring or

02:08:04   the screen sharing this is just spite. It's Apple being a dick

02:08:09   again to make people in the EU mad about the DMA so that

02:08:15   dot dot dot what they're going to get the DMA revoked. That's

02:08:18   not Apple's goal at all. Apple's not telling people in the

02:08:21   EU to call their Parliament representative and try to get

02:08:25   the DMA revoked the DMA passed by a vote of 588 to 11 in the

02:08:31   European Parliament. This was not a controversial law and it

02:08:35   were like 31 abstentions. So 31 no votes 11 against 588 for the

02:08:41   DMA is not going anywhere, but the DMA is so much more wide

02:08:46   reaching than just breaking up the app store. It the whole

02:08:50   point of it is more or less in my opinion to attack the whole

02:08:56   idea of a vertically integrated company, which is exactly what

02:09:00   Apple is and it says pretty much anything that Apple does or

02:09:05   adds to the system that's exclusive to their own products

02:09:09   and services is contrary to the DMA. So I do think I don't

02:09:14   think they're being spiteful. I think they're they are

02:09:16   legitimately 100% uncertain whether the iPhone mirroring

02:09:20   feature would be deemed compliant or not because it

02:09:23   only the iPhone is a gatekeeping platform and iPhone mirroring

02:09:28   only works with Apple's own desktop computers. Max it does

02:09:31   not work with any other computer and I think the DMA

02:09:35   could reasonably be held again. It's the ambiguity is built in

02:09:39   it's up to the European Commission to interpret it, but

02:09:43   they could definitely interpret the DMA as saying that this is

02:09:47   illegal under the DMA. Apple needs to open this up as an API

02:09:50   so that any computer could do the iPhone mirroring in the

02:09:53   same way and what Apple is saying is this is a security

02:09:56   concern because this is actually a very complicated. Doing

02:10:00   this in a very secure way is very complicated. I sort of

02:10:03   brought it up on my show last week that Apple's way of doing

02:10:07   things is to sort of do it themselves privately and if they

02:10:13   years two years in think, hey, this is good. Our private APIs

02:10:18   that we've been using are stable. Let's make it a public

02:10:21   API and now everybody can do it right and this year's example

02:10:24   of that is the way that you get a really cool pairing

02:10:28   interface for AirPods on your phone or iPad. You just get

02:10:32   your brand new AirPods, open the case, hold it near your

02:10:35   phone and you get a little three-dimensional animated

02:10:40   thing of the AirPods. It's so cool that if you get your AirPod

02:10:44   case engraved, the engraving is actually on the 3D

02:10:47   representation on the screen and you say pair and now your

02:10:51   AirPods are connected and every other company in the world

02:10:54   that makes wireless Bluetooth headphones has to go through

02:10:58   settings, Bluetooth, pair, weight. I think if AirPods were

02:11:03   brand new this year, if this was the year that they announced

02:11:05   the first version of AirPods, I think they would be delayed in

02:11:08   the EU because they'd have to wait and see whether the

02:11:12   pairing process would be deemed legal under the DMA or you

02:11:18   know, is it only legal if third parties can get the same thing

02:11:21   and if that's the case, what is Apple going to do? Are they

02:11:25   going to delay AirPods for two years while they work through

02:11:30   EU regulatory approval to see if it's legal and the two years

02:11:35   it might take to get public APIs to do it or are they going

02:11:39   to roll it out everywhere else, the 93% rest of the world and

02:11:43   make the EU wait two years? It's the latter. That's what's

02:11:47   going to happen. Yeah, I think there's really kind of like

02:11:53   three things that the DMA kind of ignores and doesn't really

02:11:56   take into consideration and then one area where Apple really

02:11:59   kind of fumbled. So you talked about the technical issue.

02:12:04   There's tons of things that from a technical standpoint are

02:12:08   just quite literally not feasible. So it's not that

02:12:11   Apple's intending to be malicious. It's just the way

02:12:13   that different security protocols work and different

02:12:16   authentication methods and the shared code bases. It's not

02:12:20   practical or sometimes even possible to let that be

02:12:23   accessible to everyone. The second argument is one that

02:12:26   Craig made on the show that I actually am glad he said,

02:12:29   because I've been suggesting this for a long time and a lot

02:12:32   of people seem to disagree with me was once you create a

02:12:36   publicly accessible API, the onus is on you as the company

02:12:39   to support that. Apple can't just come out and say, hey, we

02:12:42   have this new feature and here's a public API for it.

02:12:44   Yeah, we're doing it, but also you can too and then find out

02:12:48   that the feature doesn't really work that well or consumers

02:12:50   don't really care about it or and you're still on the hook to

02:12:53   support that API for years and years and years to come because

02:12:57   you have hardware or software partners that have built their

02:12:59   own products and services and platforms upon that. So that's

02:13:02   number two and then I think the third one is really the

02:13:06   compelling selling point of the product. I think they're and

02:13:11   you can make the argument that it's not true for everyone, but

02:13:14   it is absolutely the case that Apple has sold iPads because it

02:13:18   is fairly locked down. I buy my grandparents iPads because of

02:13:22   this very reason. I know they cannot screw it up. Right. And

02:13:25   is it does that mean that iPad OS has to be locked down for it

02:13:28   to be successful or that there couldn't be a way to side load

02:13:32   stuff in a way that's that's pretty easy for, but at the

02:13:35   same time, the social engineering aspects of it of oh

02:13:38   well, I can just already envision the boondoggle of of

02:13:41   Apple and their customers that are slighted and or financially

02:13:45   hurt because they get a call from a scam center saying, okay,

02:13:48   your iPad has a as a bug. You have to go into settings and

02:13:51   then go into alternate app stores and click the enable side

02:13:54   loading button because we've gotta help you get this. I mean,

02:13:57   you know what I mean? It's just so those are the three aspects

02:14:00   that I think there's technical issues surrounding it. There are

02:14:02   practical considerations from a financial support standpoint

02:14:06   and then last, there are product reasons why it just

02:14:09   legitimately is not of benefit. Now, if we go to the other side

02:14:13   where I think Apple really screwed up here is that there

02:14:16   has been mounting pressure for them to have done stuff for

02:14:22   years and I'm not confident that the DMA, despite it being

02:14:26   super wide sweeping and going beyond the app store is not a

02:14:30   consequence of Apple and other companies cuz they're not the

02:14:34   only one beholden to the DMA, but they're certainly a player

02:14:38   in the sense that if they had not compromised a little bit in

02:14:42   the past to allow third party payment processing or whatever

02:14:45   that the extent of this legislation might not have been

02:14:48   as wide sweeping as it has been. They've fought this for

02:14:51   years to the point where it just looks bad and part of the

02:14:56   problem is because Apple's a massive organization. It's not

02:14:59   the engineers that are deciding this. This is legal and

02:15:03   compliance official. There's all these people that are

02:15:05   involved in this massive process and that's why it's so

02:15:08   slow moving and complicated, but one has to wonder if the

02:15:12   extent to which they had to compromise on product is not

02:15:17   kind of a consequence of them being unwilling to do it on

02:15:20   their own. But in the same breath, it sucks. Yeah, right.

02:15:26   Cuz you know. Money is different, right? Everybody

02:15:30   knows money is occupies a special part of the human brain

02:15:36   and what people see is a lot of some of that not a lot some of

02:15:42   Apple's decisions that frustrate people are specifically

02:15:46   related to protecting their money they make from the App

02:15:49   Store. I mean there's no denying it and I don't think

02:15:52   anything better exemplifies it than the anti steering to the

02:15:56   web provisions that they've been enforcing all along and I

02:16:01   bring it up all the time. Ben Thompson brings it up all the

02:16:03   time, but it's the single to me. It's the one it's maybe not

02:16:08   the biggest issue, but it's the one that can really only be

02:16:12   justified from Apple's own own institutional perspective of

02:16:17   the only people. The only institution that benefits from

02:16:20   the anti steering to the web provisions is Apple itself to

02:16:24   protect the amount of money that goes through the App Store

02:16:27   instead of going through the web. It certainly is of zero

02:16:30   benefit. It's anti benefit to users in particular with apps

02:16:35   like Kindle right that you can use Kindle on your iPad and you

02:16:38   cannot buy Kindle books in the Kindle app because Amazon not

02:16:43   just won't but really couldn't sell through the App Store

02:16:47   without losing money on every book because of the way ebooks

02:16:50   are accounted for it. So it's not even a choice like say

02:16:56   Spotify or Netflix makes the choice that they years ago

02:17:01   stopped selling subscriptions through the App Store because

02:17:05   they didn't wanna share any of their subscription revenue with

02:17:09   Apple. That was a choice that Netflix made. They certainly

02:17:12   could have done what Disney does and just sell through the

02:17:15   App Store and give Apple their 15% cut of the subscription

02:17:18   over time, but ebooks can't without losing money because the

02:17:23   publisher gets 30% of the price and therefore that's the Apple

02:17:27   wants 30% and now the the ebook seller is left with less than

02:17:32   they paid the publisher wholesale. How in the world is

02:17:35   that a benefit to users? It's A, they can't buy the books in

02:17:38   the app they're using and B, they're confused as as hell,

02:17:41   right? Yeah. It only benefits Apple. There's no instructions.

02:17:44   Nobody tells you. You can't talk about. Right. So there's we

02:17:47   can prove that there's some degree of greed involved in

02:17:50   Apple's decision making here and therefore. Absolutely. But

02:17:54   therefore the fact that everybody knows there's some

02:17:57   greed involved in the decision making process. It just opens

02:18:01   the door to presuming that all of the decisions they make are

02:18:06   about greed and I really don't think that's true and we can

02:18:09   litigate this for hours on podcasts and columns and stuff

02:18:13   like that. But of course there are many people who are going

02:18:17   to just say nope. It's all about the money and it's not

02:18:20   that unreasonable because some of it clearly is about the

02:18:22   money and that's where Apple shot itself in the in its foot

02:18:25   because however much money they have saved them or made for

02:18:29   themselves by steering X percent of transactions through

02:18:35   the app store that would have gone to the web if they had

02:18:39   just allowed apps to link out to the web. They've already

02:18:43   wasted way more than that money and on their time complying

02:18:46   with all this and sending executives to lawsuits and that

02:18:50   might never have happened in the first place. They've already

02:18:52   I guarantee you they've lost money on this. It's just not

02:18:55   worth it and even if you can't prove it in a spreadsheet, it

02:18:59   has to be what they feel in their gut in terms of why are we

02:19:03   bothering with this? Why are we wasting our time on this? It's

02:19:06   no fun. It's not helping. It's not making. It's not why

02:19:10   anybody went to work at Apple in the first place. It's yeah,

02:19:14   but II really can't help but think though that there's just

02:19:17   a lot of DMA proponents who really think it's just about

02:19:21   busting up the app store when it's way more far reaching than

02:19:25   that and really does open up Apple to non compliance finds

02:19:31   for anything that they do that is vertically integrated within

02:19:35   their own products and services and I really do think that this

02:19:40   it'll become it's fresh right now because these WWDC

02:19:44   announcements are the first new announcements Apple's made

02:19:47   under the DMA. So it's the first time they're saying this

02:19:49   but people in the EU should get used to it because I don't

02:19:52   think the DMA is going the DMA isn't going anywhere. Like I

02:19:55   said, it's it's on the books. It was passed overwhelmingly,

02:19:59   but I think henceforth every future thing Apple does is

02:20:03   going to come late and possibly limited to the EU. Apple Watch

02:20:08   if Apple Watch were brand new, it would definitely be delayed

02:20:12   because it only works with iPhones and II think that's I

02:20:16   think that's illegal under the DMA or it could be they would

02:20:19   need them to to rule and they don't rule in advance and Apple

02:20:23   doesn't want to tell them in advance. It's not the Apple way

02:20:26   of doing things. Hey, we're thinking about unveiling these

02:20:28   new smart glasses next year. What do you guys think in the

02:20:31   EU? They're not gonna do that. No. Yeah, it's a mess and it's

02:20:36   a real bummer and I agree that it's it's the first of many

02:20:39   things that are going to start to be problematic, but it's

02:20:44   also somewhat hard to be fully empathetic because it is

02:20:47   somewhat of their own doing. It's just unfortunate. Really

02:20:51   the bottom line is it's unfortunate for users. Yeah and

02:20:54   consumers are the ones that lose really at the end of the

02:20:57   day. Yeah. Yeah. And and I really don't think I just know

02:21:01   people that sort of people who work at Apple. They've II don't

02:21:04   think they're like yeah stick it to the citizens of the EU.

02:21:08   They I think they feel terrible about it. II think along the

02:21:11   lines of what I'm arguing though that they they realize

02:21:15   that the the Apple using citizens of the EU really don't

02:21:19   understand the full ramifications of the DMA and

02:21:22   again, it's the European commissioners themselves.

02:21:25   Margaret Vestager who's they're they're continually

02:21:28   emphasizing. They're the ones making the threats emphasizing

02:21:31   that they can find Apple massive amounts of money and

02:21:35   even talking in my opinion almost laughably, but she's the

02:21:40   one bringing it up breaking up the company, which I guess

02:21:43   would not really breaking up company nation or

02:21:49   internationally, but forcing them to what make a European

02:21:53   subsidiary of the app store that I don't even know what

02:21:56   that would mean, but she's the one making the threats and when

02:21:59   government regulators are threatening a company publicly

02:22:02   they take it seriously. I guess that's the other point I would

02:22:05   make is that so many people are saying Apple isn't taking the

02:22:08   DMA seriously. I think Apple's taking the DMA as seriously as

02:22:12   a heart attack. I think they're seeing they see it as an

02:22:14   existential threat not because the EU itself is so big, but

02:22:19   because if DMA like legislation passes in more countries like

02:22:24   you said, it won't be an exact clone of the DMA. So the

02:22:28   compliance would need to be entirely different. Yeah

02:22:31   federated around the world and it really does attack the very

02:22:36   nature of what Apple does best, which is the vertical

02:22:39   integration of if you buy our phone and our smartwatch, they

02:22:42   will work better together in an easier, more reliable way than

02:22:48   a watch from company A and a phone from company S to pick a

02:22:53   letter. Anyway, there's my rant on the DMA. I guess that's

02:22:57   about it. Anything else you wanted to touch on or you're

02:23:00   ready to wrap it up? Yeah, I think that's about it. I mean

02:23:05   Vision OS 2 was kind of there were some nice updates there.

02:23:08   Yeah, I will say my experience echoes everybody else's where

02:23:12   it's almost a shame that they have to go through a beta

02:23:15   because I actually I actually think it might be it might be

02:23:19   better. I think it's a couple of obvious bugs, but I suspect by

02:23:23   Beta two or three, it might be better than Vision OS 1.1 and I

02:23:28   kind of feel like the nature of Vision OS being a 35 four

02:23:33   3500 $4,000 product that the the people who have it are sort

02:23:39   of the sort of people who aren't afraid to run betas. I

02:23:42   don't know. It's I would I would love to see the percentage

02:23:45   of Vision pros that are running the developer Beta. I bet it is

02:23:48   shockingly high. It's got definitely double digits. No

02:23:50   doubt about it. I'll try to find it from a friend. I maybe

02:23:53   sandwich will know he might have maybe yeah. A lot of the

02:23:57   features are just much improved. I think the new

02:23:59   control center interaction and gestures are way better. It's

02:24:03   brilliant how well they work. I mean this is another example of

02:24:06   so the the hold out your palm and touch your your index and

02:24:10   your thumb together to to summon the home screen right.

02:24:13   That's the identical interaction that meta uses so

02:24:16   Apple just straight up copied Facebook. However, it's so much

02:24:20   better. I mean because you can do it without and I don't know

02:24:25   you should try this. It's very very interesting. You have to

02:24:29   look at your hand with your eyes using eye tracking to get

02:24:32   it to actually perform the action, which makes sense. You

02:24:34   don't want it. Misinterpreting some sort of function without

02:24:38   you looking at your hand, but what you can do is for the

02:24:41   control center gesture, you have to open your hand and then

02:24:44   you rotate your palm over. You basically twist your hand to

02:24:48   get the control center little bubble to show up. What you can

02:24:51   do is you can begin that gesture before you've looked at

02:24:55   your hand and then after completing the gesture, you can

02:24:58   look over at your hand and the little indicator will show up.

02:25:01   It's just super well thought out and so seamless and smooth

02:25:04   and that in addition to the enhanced polling rate for hand

02:25:08   tracking just frankly makes the device feel way different like

02:25:14   if you had if you had given me this vision pro and if you put

02:25:19   me in a box and said you're in a time machine, it's now 2025

02:25:22   and you gave me the the you had me put this on and I played

02:25:26   around the OS. I'd say. Oh yeah. This is definitely the

02:25:28   vision pro two. This is not the same hardware. It it truly feels

02:25:33   substantially better. It's wild. Yeah and it's just sort of

02:25:37   it's they're still like I described it in my piece like

02:25:40   vision OS two doesn't really get seemingly doesn't get any of

02:25:44   the Apple intelligence features. They didn't mention it

02:25:46   and it doesn't even get stuff like math notes, which isn't

02:25:48   Apple intelligence because they've still got so much low

02:25:51   hanging fruit because it's such a dripping. I mean you

02:25:54   couldn't even rearrange the icons on the home view till

02:25:57   this beta right. I mean I mean and they know right. I mean and

02:26:00   again the the iPhone didn't have copy and paste until

02:26:03   version three right. I mean never and Apple knew people

02:26:06   wanted copy and paste. It's stuff has to get prioritized. I

02:26:09   guess though the thing that really hit me once I started

02:26:11   using vision OS two beta is just how much I never realized

02:26:18   how how every time I hit the crown to go to the home view.

02:26:22   It took me out of the immersion right because and I know it's

02:26:28   everybody loves to emphasize that the first generation vision

02:26:32   pro is too big too heavy sticks out too far. I mean of course

02:26:36   it's the biggest heaviest vision pro they're ever gonna

02:26:39   make probably but once you once you're yeah. Hopefully once

02:26:43   you're in it, it is immersive and you ideally until you get

02:26:50   fatigued or something you're you're like in it it and your

02:26:55   mind is in it and going out of the world that you see around

02:27:01   you to tap that button totally takes you out of it every time.

02:27:04   It just is like all of a sudden instead of being in this

02:27:08   virtual world. All of a sudden you're back to a computer is

02:27:11   strapped on my eyes and I've gotta hit a button. No I'm

02:27:15   because touching the button almost moves the headset on

02:27:17   your face and you're like. Oh yeah. This is hot and large and

02:27:19   uncomfortable right and it's like double tapping the other

02:27:22   button to confirm a purchase from the store. I feel like

02:27:25   that's good right. You want to be taken out. Oh you're making

02:27:29   a purchase take me out of the moment. Let me confirm but for

02:27:33   just switching between apps, it was in now that I've tried the

02:27:37   new way. I'm like. Oh, this is yeah. This is way better way

02:27:41   better than the whole look up to get the control center was

02:27:44   always never good. Yeah. No because even after I got used

02:27:47   to it, I'm like how far up do I need to look. I feel like I'm

02:27:50   like looking all the way up at the ceiling. Well, it's weird.

02:27:52   It's it's it's counterintuitive because you you've technically

02:27:55   only have to move your eyeballs up cuz you're right. I try but

02:27:59   people as we move our eyes up. We also tilt our head up. So

02:28:03   it's that same thing where you keep tilting your head and

02:28:04   you're like when is this bubble gonna show up and you're just

02:28:07   craning your neck until you're in pain and then you're like

02:28:09   well, I've hit the end stop of my neck. I guess I gotta move

02:28:12   my eyes further and then it would finally work. Yeah. It's

02:28:16   it's just much more refined and much more polished and even

02:28:18   though it is a small update from a feature standpoint, I

02:28:21   think that's okay because the platform is so young and some

02:28:25   of the stuff that I didn't like has already been fixed, which

02:28:28   is good and then one of the things that I never really

02:28:33   didn't speak to me that much was the spatial photos and

02:28:35   videos. So II saw all these reviews from reviewers that had

02:28:39   had time with the headset and I think you as well had

02:28:42   mentioned it. I know other reviewers had basically said

02:28:45   this is a huge deal being able to see your photos with spatial

02:28:49   depth and you kinda relive the moment and I remember being

02:28:52   super excited for that. I had been shooting spatial videos

02:28:54   for quite some time leading up to the vision Pro launching and

02:28:58   I remember opening them and going. Yeah, it's kinda cool,

02:29:02   but it didn't really speak to me now with vision OS two. It

02:29:07   creates artificial depth on any photo at all. It can be photos

02:29:12   that you took 20 years ago on a DSLR. It could be photos from

02:29:15   the web. It does it to any photo and because this is all

02:29:18   done from an ML standpoint, it doesn't deal with the reality

02:29:22   that the iPhone 15 Pro Max and 15 Pro have two camera lenses

02:29:27   that are still quite close together. so they don't really

02:29:29   create that convincing of a depth effect. It's very shallow

02:29:34   once you can do that all artificially. It's actually

02:29:37   pretty amazing and so I've spent hours looking at photos

02:29:43   inside of the vision Pro from when I was a kid and it truly

02:29:47   was emotional in a way that I never anticipated because I was

02:29:52   able to see my grandparents who have since lost in this almost

02:29:56   hyper realistic 3D depth and the amount of work they do to

02:30:00   make it convincing is really impressive. There is one of me

02:30:02   in front of a tree, maybe a few feet and not only did it cut

02:30:08   me out as the subject and separate me from the tree, but

02:30:10   it even added a drop shadow in between me and the tree and so

02:30:15   it did this really convincing depth effect and I really only

02:30:18   had I've probably looked at. I don't know 150 200 photos.

02:30:21   There was one photo where I was like. Oh yeah, this this got

02:30:25   messed up, but it's unbelievable how reliable it is

02:30:29   and and if you look at the try to pull the thread, you can

02:30:33   start to see there's weirdness like one example is I was on a

02:30:36   pier next to a cruise ship as a kid and the the camera was

02:30:42   shot at a fairly high aperture. So there's no depth of field.

02:30:45   Everything is in focus. It's infinity focus and inside of

02:30:49   the vision Pro you look at me and there's perceivable depth,

02:30:53   but then you look at one of the light posts six light post

02:30:56   backs. The is pretty far away and there's faux depth added to

02:31:00   that too. So it's it's kinda weird cuz everything is in

02:31:02   focus, but it's also kind of not, but it still was

02:31:05   convincing and then when you have a photo that does actually

02:31:09   have natural bokeh like a portrait photographer or

02:31:12   portraiture photos. It's unbelievable. It looks like it

02:31:16   was legitimately taken with a 3D camera and it's it's really

02:31:20   really really cool and it's one of the things that I've been

02:31:23   really excited to show people that try it cuz it it is it's

02:31:27   surreal to see stuff that was taken 2030 4050 years ago and

02:31:31   make it feel like it's happening again, which I just

02:31:34   didn't get the same effect with when it was a video I had taken

02:31:37   three months ago and I'm like, Oh yeah, I remember that it was

02:31:40   a little while back and it feels like one of those things

02:31:43   that they probably were hoping to have a year ago so that they

02:31:46   could talk about it when it first came out because the

02:31:48   other thing I've been thinking ever since the end of the year

02:31:51   November when they first started letting people in the

02:31:54   press play with the spatial video on the iPhone 15 Pro.

02:31:59   Well, okay. This is cool, but why can't I take still photos

02:32:02   in three days and I get the answer. It seems is oh you

02:32:06   don't need to go to a special you do need the special mode

02:32:09   for video because video shooting is obviously more

02:32:13   complex than still but totally just wait just shoot photos.

02:32:18   I wish they would have told me that it was it would have been

02:32:20   a much better answer than here. They are cuz I asked about it

02:32:24   and I just got one of those non answer Apple answers. That's

02:32:28   come on. That's not an answer. This is weird right. You can't

02:32:30   you can shoot 3D video, but you can't shoot 3D stills, but now

02:32:33   you don't need to and it is definitely compelling and it's

02:32:37   I know it's people aren't gonna spend most people aren't gonna

02:32:39   spend $4000 just to see their photos big and in 3D, but it is

02:32:44   really something to see your personal photo library big. I

02:32:48   mean there's a lot of photos that I have that I've never

02:32:51   really seen most of my photos. I've never really seen even as

02:32:55   big as my studio display. I don't look at all my photos

02:32:57   full screen, but even then it's only 27 inches as opposed to

02:33:01   movie theater size and the depth is definitely. I don't

02:33:04   know it's emotionally relevant resonant to experience depth

02:33:08   like that. So good good for them. Good update sort of

02:33:12   finally again. That's right everybody. Where's the best

02:33:17   place your home on YouTube is what youtube.com/snazzyq

02:33:21   slash snazzy. Yep. Yep. Yep. Snazzy Labs everywhere else.

02:33:28   Snazzy Labs everywhere else. You're you're on the threads.

02:33:31   You're on the mastodon. Yeah. I'm on Mastodon. I'm still on

02:33:36   Twitter. Unfortunately, the old one can't get away from that

02:33:38   cesspool. Something about it. Something they put in the water

02:33:42   that makes me come back. I don't know why. Well, I enjoy

02:33:45   it, but I'm there. I'm not making apologies. I mean, but I

02:33:48   have we we and I know you do too. I have readers and

02:33:51   followers who are there who aren't elsewhere and I'm not

02:33:55   I'm not delighted about it, but I'm happy to go to where my

02:33:58   readers and listeners are so still there. Quinn always a

02:34:01   pleasure to have you on the show. Thanks for having me

02:34:04   again. John keep up the good work. Let me thank our

02:34:05   sponsors. We had trade coffee where you can get a tailored

02:34:08   subscription plan just for yourself and Squarespace where

02:34:11   you can build a website and have your own presence on the

02:34:13   web thanks to them. Thank you Quinn. Alright thanks man appreciate it.