0:00

I have found a great use

0:02

for the Vision Pro. This

0:05

is going to be good. I don't think I

0:07

even knew you had yours back, or do you

0:09

not? I have. I've had it back. It has

0:11

been in its little marshmallow pod for, you know,

0:13

a month or two, at least, since the

0:15

last time I used it. I thought you had loaned it to

0:17

somebody. That's why I got it. I did. Yeah, I loaned it

0:19

to a friend for like two or three weeks earlier in the

0:21

spring. I got you. But, you know,

0:23

I've had it back now for a while. It took me

0:25

a while to, like, you know, actually motivate myself to, you

0:27

know, actually reset it back up as

0:29

me and, you know, charge it, update it, and all that stuff.

0:33

Because, like, you know, every time that you put it on

0:36

and it's either discharged or it needs a software update, that's

0:38

an excuse for you or that's kind of a requirement for you to,

0:40

like, all right, let me take it off. I'll let it do its

0:42

thing and I'll come back to it. And

0:45

I never come back to it. This

0:47

last few days, I have

0:49

been strictly constrained to

0:52

lying on the couch because

0:55

I got a vasectomy. It doesn't really

0:57

matter for everyone to know this, but

0:59

it's minor surgery and more men in

1:02

this country should do our part for

1:04

birth control, since especially women's options are

1:06

being needlessly and horribly

1:08

limited. So anyway, it's no big

1:10

deal. I've been laid up on the

1:12

couch for a few days and I was told, you

1:14

know, don't do anything. Like, don't sit at a desk,

1:17

don't walk around, like, don't lift heavy things, don't do

1:19

anything. And honestly,

1:21

really, men out there, it really isn't that bad. If

1:24

you're on the fence, just do it. I was scared

1:26

and it was fine and it's totally fine. Anyway,

1:29

I had to lay on the couch for, you know, a couple days and I

1:32

am really bad at doing nothing. Like, I just want to get up and do,

1:34

I want to work, I want to do stuff for in the house, I want

1:36

to walk the dog, like, there's a million

1:38

things I want to be doing. But

1:40

I had to lie on the couch. And when you're

1:42

lying on the couch, it's really not that

1:44

comfortable to use a laptop, you know, when you're,

1:46

like, really lying down. It's also, you know, in

1:48

a situation like this, you might not necessarily want

1:50

to be putting things in your lap. Except

1:54

frozen peas. Right, yes. And by

1:56

the way, those work great. Noted.

1:59

way better than the special Amazon underwear

2:01

with the ice packs. I

2:04

can say a bag of frozen peas works better than

2:06

that. Noted. As you save you 40 bucks. Anyway.

2:10

That is true. Because as opposed to say I

2:12

will be falling in your footsteps at the end

2:14

of the summer. So this is good advice. Jeez,

2:16

all right. Well, after the apocalypse, I guess it'll

2:18

be up to me to repopulate the planet. Yep.

2:20

It'll be all little Syracuse's. It'll be great. Oh

2:23

my gosh. Anyway, this

2:25

was the perfect use case for the vision

2:28

pro. I got to like lie

2:30

down on the couch and like, you

2:32

know, I didn't want to, you know, if you're lying

2:34

down on the couch in most living room arrangements, including

2:36

mine, you're not, you can't look directly at the TV.

2:38

The TV is like off to your side. Oh,

2:41

contrary. Yeah, I think I

2:43

have that moved down. But anyway, go on. Oh yeah. Because

2:45

you're like more diagonal. Anyway, we

2:47

have a more, you know, a more like

2:49

standard like, you know, linear couch to TV

2:51

layout. Easy places to put your

2:53

speakers. Yeah, I know it is. Anyway,

2:57

so I've

2:59

been watching the vision, watching stuff in the vision

3:01

pro and kind of messing around with it as

3:04

basically lying on the couch, being able to look straight

3:06

up and have a virtual screen

3:09

projected straight up above my head. I've

3:11

been like watching WBC session videos mostly and stuff

3:13

like that. But like a long time ago, when

3:16

Federico Viticchi had to

3:18

spend a lot of time in hospital beds for

3:20

a while, that's when he fell

3:22

in love with the iPad because the iPad was easy

3:24

to use in a hospital bed. And

3:27

that's that was like a huge computing win

3:29

for him. I

3:32

can say with confidence that the vision

3:34

pro is not necessarily

3:36

as big of a win for, you

3:39

know, general purpose computing as the iPad is

3:41

in that context. But it is

3:43

it does serve an interesting role in that, like, you

3:45

can compute while you are lying down and

3:48

bear and you can do so while holding

3:50

nothing in your hands and barely even moving

3:52

your hands. And so there are decent

3:54

numbers of cases where that can be very useful to

3:57

people. It is a lot harder to get

3:59

a lot of things done. done in the Vision Pro, to

4:01

be honest. And there's a lot less software and a lot

4:03

of totally missing apps that you just have no way to

4:05

run in there. Like the iPad. But

4:07

man, it actually was like, I

4:11

actually really came to appreciate it during these

4:13

two days for that purpose. I thought you

4:15

were going to say that the person who

4:17

did the procedure used the Vision Pro to

4:19

do it. Oh, wow. No, I sure would

4:21

not have appreciated that. No. Maybe they did.

4:24

You don't know. That's true. I have no

4:26

idea. Well, I'm glad that it

4:28

took a little minor surgery to get you

4:30

on the Vision Pro bandwagon. Welcome. We're happy

4:32

to have you. So

4:35

all kidding aside, you said you were watching

4:38

some developer videos. Is there anything else that

4:40

or any moments that you've had that you're

4:42

like, hey, this is actually pretty great. I

4:45

also did watch the talk show live in

4:47

there. It was actually really cool to

4:49

see it that way. It worked very well. In fact, Casey,

4:52

you can even hear you talking at the very,

4:54

very end, like when the lights come up and

4:57

the lights come up because you were apparently sitting right next

4:59

to the camera rig. Right at the end, you hear Casey

5:01

say something right before the audio cuts out. I

5:04

was sitting right by the camera rig because it blocked my view

5:06

of the stage. Well. It's very upsetting.

5:08

But I survived. I was like, if you wanted

5:10

to see the show from the perspective of where

5:12

I sit, that's the 3D thing that they put

5:14

up. And I'll tell you what,

5:17

so the talk show live, so our friend

5:19

Adam Lissigore at Sandwich and this

5:21

other company, I forget the name of it, Spatial,

5:23

something? Something Gen maybe. Yeah, I think

5:25

that's it. They did a live, so

5:27

Sandwich has a new app called Theater.

5:30

It's kind of like the big screen

5:32

equivalent of their television app where you

5:34

can watch arbitrary videos inside like cool

5:37

settings or in the case of television, inside retro TV

5:39

sets in the Vision Pro. So they

5:41

made this Theater app. And they did a live

5:43

broadcast of John Gruber's The Talk

5:46

Show live at MVDC event in

5:48

stereo video, so in 3D video.

5:51

It's really cool. And

5:53

first of all, it's an interesting and remarkable

5:55

technical achievement that they were able to do

5:57

this live event this way with. with

6:00

what appeared to be two very small companies, relatively

6:05

speaking, while Apple has currently broadcast zero live

6:07

events to the Vision Pro. So as far

6:09

as I can tell, there have been no

6:11

other live events broadcast to the Vision Pro.

6:15

And it was really compelling. And I don't want

6:17

to speak for them, but from what I heard, it

6:20

sounded like the engagement numbers were pretty good

6:22

to my ears. There

6:25

was a cool potential for the

6:27

Vision Pro for spatial

6:30

video broadcast of live cool events.

6:34

And I just hope someone else ever does it. That

6:39

really is a really cool idea.

6:41

It worked very, very well. It

6:44

turned out great. And I would

6:46

love to watch other concerts or

6:48

productions or other live events this

6:50

way. Obviously, sports people are dying

6:52

to watch sports this way too.

6:55

There is so much potential there. And so

6:57

I hope that potential is

6:59

realized. Because obviously,

7:02

if this one or these

7:04

two small companies could get together and do this, obviously

7:07

there's nothing stopping Apple or major sports leagues

7:09

or major content providers from

7:12

doing this themselves. There's still a long way to go in a

7:14

lot of the technical angles of it. But

7:16

this was literally just like this was a fixed

7:18

camera at a fixed location. It was like you

7:20

were sitting in the front row and you just

7:22

saw a fixed viewport and it was 3D. And

7:25

it was a little bit low resolution and a little bit

7:27

low frame rate, but it looked great. It looked like you

7:29

were there. It was a really cool thing to see.

7:31

And so I really hope we see more of this coming to the

7:33

Vision Pro. That's really awesome. I

7:35

do plan to at least take a quick watch

7:38

of it. John and

7:40

a few of our friends and I were all

7:43

sitting front row because Gruber was kind enough to

7:45

leave some reserved seats and whatnot. And so the

7:47

view of the camera is basically what Syracuse and

7:49

I had seen. We were on either side of

7:52

it, but basically what we had seen for the

7:54

show. And it was it was a good show. It

7:57

ran long. I was surprised that that the

7:59

Apple Ex were willing to give Gruber

8:01

two hours. Not that he's undeserving, just that

8:03

I feel like typically they're getting antsy at

8:05

like 90 minutes. But it was a full

8:08

two hours and it was good. So yeah,

8:10

you should definitely check it out. And they

8:12

have a rendered or mastered in 4K YouTube

8:14

version, like a standard 2D YouTube version, which

8:17

I also have not watched yet, or I mean, since

8:19

I was there. But nevertheless, if you don't have a

8:21

Vision Pro, don't feel like you're entirely missing out. You

8:23

can just watch it on YouTube. It's also a podcast.

8:25

Like, there was a podcast version today as well. And

8:28

content wise, I got to say that I really enjoyed

8:30

this live talk show. So they

8:32

were interviewing John Jean Andrea, Jaws,

8:35

and Federighi. And I think

8:37

it went great for all three of them. And

8:39

I think John Gruber had a really good balance

8:41

of like, questions to kind of

8:43

let them flex and show off and tell us

8:45

a few more cool details, but also some hard

8:48

questions that kind of had them have

8:50

to answer for certain things or have to explain certain things.

8:52

It was a really good balance of that. And

8:54

so I really enjoyed it. Highly recommended for anybody

8:56

who listened to this show. You

8:58

know what else is highly recommended? Our

9:00

interview, which we were lucky enough to do, we're

9:04

recording this on Friday. It was

9:06

Tuesday that we sat down with Holly

9:08

Borla and Ben Cohen, both Swift

9:11

compiler engineers in that

9:13

vicinity. Court team. Yes, Swift

9:15

Court team, thank you. I

9:18

wasn't nervous going into it, but

9:20

you don't know how it's going to go.

9:23

And having five people on one show can

9:25

be challenging. And I

9:27

thought that Holly and Ben did a

9:29

phenomenal job. I thought it was really

9:31

great. They are very willing to

9:34

get in the weeds, but they don't jump immediately

9:36

there. You know, they're extraordinarily good communicators, both of

9:38

them. It was really, really great. And I had

9:40

an absolute blast. I don't want to speak for

9:42

you too. Well, maybe I do a little bit,

9:44

but I could have gone easily another

9:46

hour, probably another two, if we had the space

9:48

and the time and the, I

9:51

don't know if Holly and Ben would have like that, but I

9:53

would have. It was a lot of fun, and it was really,

9:55

really great. And that is not behind any

9:57

sort of paywall or anything like that. It's just a little

9:59

bonus episode. that we released a couple of

10:01

days ago now. Yeah, and it's

10:03

primarily for programmers. Like if you're a programmer,

10:06

we do go heavy into programery stuff. It

10:08

helps if you know about Swift, Apple's ecosystem.

10:10

We went very deep because to

10:13

take advantage of the people we got to speak to,

10:15

that was the best use of our time. Because if

10:17

you're talking to experts, you don't ask them just the

10:19

basic stuff. But we have heard from

10:21

some people who are like, look, I don't develop for

10:23

Apple's platforms, or maybe I'm not even a programmer, and

10:25

they still found it interesting to get a feel for

10:27

things, because we did cover stuff at a higher level,

10:29

as well as getting way down from the nitty gritty.

10:31

So please check it out. Yep, yep,

10:34

it was really, really fun. And I think you would enjoy

10:36

it. All right, this

10:38

is the customary WWDC or

10:40

post WWDC all follow up

10:42

all the time episode. So

10:44

unless we somehow absolutely fly through this, I'm

10:47

just setting the stage now, this is all

10:49

follow up, but that's okay. We got to

10:51

clear the decks so we can get back

10:53

to regularly scheduled programming in the

10:55

next episode. So Microsoft has had

10:57

a bit of a roller coaster over the last few days

10:59

or last couple of weeks, really. So

11:01

on June 7th, there's a Verge

11:04

article, Microsoft Changes Recall, which is

11:06

their thing which records your screen

11:08

and lets you ask questions of

11:10

what you saw where. Anyways,

11:12

Microsoft changes recall to be opt-in

11:15

and improves the security from the Verge. As

11:18

we predicted, by the way, in the pre-WWC episode,

11:20

that they would have to change it to opt-in.

11:22

And we weren't just kidding. It was obvious next

11:24

move and they did it almost immediately after we

11:26

released the show. Yeah, and that's the right move.

11:29

Microsoft says it's making its new

11:31

recall an opt-in feature and addressing

11:33

various security concerns. Windows

11:36

and devices VP, Pavan

11:38

Davolri, Davolri, hopefully

11:40

that's somewhere close, said if

11:43

you don't proactively choose turn it on,

11:45

it will be off by default. They

11:47

also said, we are adding additional layers

11:49

of data protection, including just in time

11:51

decryption protected by Windows Hello, enhanced sign-in

11:53

security. So recall snapshots will only be

11:55

decrypted and accessible when the user authenticates.

11:57

In addition, we encrypted the search index

11:59

database. June 7. Fast

12:02

forward just barely under a week. It's

12:04

now June 13th. Microsoft delays recall again.

12:06

Won't debut it with the new co-pilot

12:09

plus PCs after all. Reading

12:11

this time from Ars Technica, Microsoft will

12:13

be delaying its controversial recall feature again,

12:15

according to an updated blog post by

12:17

Windows and Devices VP, Pavin Davowuri. And

12:20

when the feature does return quote in the coming weeks

12:22

quote, Pavin writes, it will

12:24

be as a preview available to PCs

12:26

in the Windows Insider program, the same

12:28

public testing and validation pipeline that all

12:30

the other windows features usually go through

12:32

before being released to the general public.

12:34

That was the thirteenth. So, so it

12:36

went from a flagship feature to a

12:38

big controversy to not being opt in

12:40

to not shipping at all except as

12:42

a Windows Insider sort of beta preview.

12:46

Really this has soured the whole home

12:48

plus PC launch for like the,

12:51

you know, Snapdragon ARM processor and

12:53

everything. It's just what

12:55

should have been such a clean win for them. Hey,

12:57

we have good laptops now, a software

12:59

feature. And I think actually a software feature

13:01

with a potential to be a good software

13:03

feature is just done so poorly

13:05

with, you know, this is what it's so

13:08

important to like pick the right defaults to

13:10

know your audience, to know how to frame

13:12

things like the same basic feature

13:14

could have been released without all of this.

13:16

If it had been implemented better, if it

13:19

had been off by default, if it had

13:21

been like not present on the enterprise version

13:23

of windows, like you have to really know

13:25

who, who is able, who wants this, who's

13:27

willing to give it a chance and who

13:29

absolutely does not want this on

13:31

their computers. And Microsoft really blew

13:34

this one. Sure

13:36

seems like it. All right. So

13:38

Intel and AMD's co-pilot plus PCs

13:40

won't have the co-pilot AI features

13:42

at launch. Whoopsie doopsies. It's reading

13:44

from the verge Microsoft's new windows

13:46

AI features like auto super resolution

13:49

for smoother gaming arts exclusive to

13:51

Qualcomm Intel's lunar lake and AMD's

13:53

Strix point chips will have enough

13:55

AI co-processing performance too. But

13:57

when in Dell and AMD's new co-pilot plus

13:59

PCs around this fall, no one is promising

14:01

they'll ship with any or with all or

14:03

even any of the new AI features. Each

14:06

of those laptops will require free software updates

14:08

before they get Microsoft's Copilot plus AI features

14:10

and those updates won't necessarily even arrive before

14:12

the end of 2024. Microsoft said Intel, Lunar

14:14

Lake and AMD Strix PCs are Windows 11

14:17

AI PCs that meet our Copilot plus

14:19

PC hardware requirements. We are partnering closely

14:21

with Intel and AMD to deliver Copilot

14:24

plus PC experiences through free updates when

14:26

available. I mean this

14:28

is like also Qualcomm's having problems, the ARM,

14:30

the big coming out party for Windows on ARM

14:32

with good laptops is not going well and Intel

14:35

and AMD as we noted on the pre-WDC episode,

14:37

they have their own processes that also qualify for

14:39

this and they're not here

14:41

to rescue anybody because Microsoft doesn't have the software features

14:43

ready for them yet. It seems like Microsoft

14:45

was like this is gonna be, I mean

14:47

I don't know, this is our chance to make ARM

14:50

a thing, you know we won't

14:52

even support the Intel AMD things or maybe

14:54

it was just poor planning but this whole

14:56

launch, I guess it's just a 2025 thing

15:00

and they're hoping we'll just ignore it until

15:02

then. I'm very disappointed. Obviously I

15:04

want Windows to go entirely to ARM, that seems

15:06

like it's not happening but they can't even get

15:08

the Intel and AMD versions of these features out.

15:11

Huge wonder for the games. I do, 100% I

15:13

do. I mean I think this you know this shows though like

15:15

a massive architectural

15:17

change and all new processors and all

15:20

new hardware, it's a big ordeal.

15:22

Like you know it shows yeah Apple was able

15:24

to do it but it took a ton of

15:26

work and that's just one

15:28

company who you know could get aligned

15:30

behind it. The PC ecosystem

15:33

does not work that way at all. You

15:35

have, I mean you know think about the

15:37

uphill battle they have here. They have

15:39

an architecture transition that most of their

15:41

customers don't want, that totally

15:43

screws some of their biggest partners

15:45

in Intel and AMD. Dealing

15:48

with a company Qualcomm that is certainly not

15:50

like super easy and friendly to deal with

15:52

by most metrics that

15:55

we hear about. But

15:57

you know mainly selling into companies

15:59

that that kind of don't want

16:01

this, selling to users that kind of

16:04

don't want this, with a

16:06

bunch of software that's not ready for it. So

16:09

yeah, they kind of have a tough

16:11

environment to get this through. They've

16:14

tried this before. It didn't work. Obviously, things are

16:16

a little bit different now. Technology's

16:18

better. The translation slash

16:20

emulation story is better. So

16:23

I think they have a better chance now, but

16:25

it's still far from an easy thing. And what

16:27

they're doing is trying to convince a whole bunch

16:29

of companies and a whole bunch of customers to

16:31

take a move that many of them don't want

16:33

to take. Yeah, so the recall thing doesn't

16:35

help with that, because that's like just an additional thing to deter

16:37

you. You didn't even know this was on the table. It's a

16:39

thing that could make you not want it. But guess what? We

16:41

added a thing that's scary and you don't want. It's on top

16:44

of that. And the thing is, I don't

16:46

think they're framing it as an architecture transition. It's more

16:48

like, here's another way that you

16:50

can run Windows. And it's complicated by the fact

16:52

that they keep bringing up AMD and Intel. And

16:54

AMD and Intel do have hardware that competes, which

16:57

is why it's not a transition. It says, look, these

16:59

are new software features in Windows. And by the way,

17:01

they're also supported on a new architecture. And you shouldn't

17:03

have to know which is which. But

17:05

the reason for you to buy an ARM PC is

17:09

undercut by Intel and AMD

17:11

having competitive SOCs, which they

17:13

will soon-ish. And then

17:15

those SOCs won't have all the features that you're rolling out

17:17

in software. Again, it seems like their

17:19

role is to, in 2027, if

17:22

Microsoft paint a picture of what the PC

17:25

market would look like, I guess it's like

17:27

you can buy a Windows PC with AI

17:29

features. And it will either have an x86

17:31

processor or an ARM processor. And what

17:34

is their ideal percentage of the market? Is it

17:36

50-50? Is it 60-40? Is it 90-10? It

17:40

doesn't seem like they're even trying to make a

17:42

transition. Because as you noted, a transition would totally

17:44

screw over into AMD. And

17:47

they don't want to do that. But I

17:49

don't honestly know what they're trying to do here.

17:51

But whatever it is, they're doing it poorly. UTM,

17:54

which is a general purpose emulator, won't

17:58

be in the App Store. Excuse me. So this

18:00

is a post from one of

18:02

the authors, I guess, Thomas, after

18:04

an almost two-month-long review process, Apple

18:06

has rejected UTMSE from the iOS

18:08

App Store as well as from

18:11

notarization for third-party app stores. Yikes.

18:14

Their reasoning is that rule 4.7, which

18:16

Apple recently introduced that allows for Delta,

18:18

PPSSPP, and other emulators to be allowed,

18:20

does not apply to UTMSE. The App

18:22

Store review board determined that, quote, PC

18:24

is not a console, quote, regardless of

18:26

the fact that there are retro Windows

18:28

and DOS games for the PC, that

18:31

UTMSE can be useful in running. Additionally,

18:33

Apple stances that UTMSE is not allowed

18:35

on third-party marketplaces either because rule 4.7

18:37

also applies to the notarization review

18:41

guidelines. So

18:43

rule 4.7 covers mini apps, mini

18:45

games, streaming games, chatbot, plugins,

18:48

and game emulators. Then there was

18:50

an update later on from UTM. Apple's

18:52

reached out and clarified that the notarization was rejected under

18:55

rule 2.5.2, and that 4.7

18:57

is an exception that only applies to App

18:59

Store apps but which UTMSE does not qualify

19:01

for. And then you can see more

19:03

on Michael Tsai's blog. Finally, UTM

19:05

writes, we will adhere by Apple's

19:07

content and policy decision because we

19:09

believe UTMSE, which does not have

19:11

just-in-time compilation, is a sub-parce experience

19:13

and isn't worth fighting for. See

19:16

a blog post that I think we're pretty sure we

19:18

talked about this in the past but about why Dolphin

19:20

the, what is that, a Wii emulator? Is that right?

19:23

Yep. Thank you. Why it doesn't

19:25

come into the App Store. And so UTM continues, we do

19:27

not wish to invest any additional time or effort trying to

19:29

get UTMSE in the App Store or third party stores unless

19:31

Apple changes their stance. I'm not

19:33

loving this. Not loving this at all. So this

19:35

is something because, like, so what? Apple rejecting stuff

19:37

in the App Store, whatever. Like, even just, it's

19:39

weird that they're rejecting because they allowed Delta. And

19:41

this is an emulator that doesn't have a JIT,

19:43

so it should fall within the rules. But it's

19:46

a PC emulator, and PCs aren't consoles. Remember, we

19:48

talked about the definition of, like, retro

19:50

console games? What do all those words mean?

19:52

Apple has now said after two months, PC

19:54

is not a console. Fine, whatever. But they

19:56

also rejected it from the notarization process, which

19:59

is an over- overloaded term in the Apple

20:01

world that they do for things going to third

20:03

party app stores. So they can't

20:05

even get this into a third party app

20:07

store. And Apple's only supposed to reject things

20:10

from third party app stores for using

20:12

private API's and for security reasons. They

20:14

could maybe make an argument that there's

20:16

some sort of security issue with this

20:18

thing of like, well, it's an emulator

20:20

and like it can download

20:22

arbitrary but like I just I don't understand how they're

20:24

going to defend this to say, Oh, remember

20:27

when we said we're just going to check for private

20:29

API's and security also, we're just going to make decisions like

20:31

we don't want this emulator in any stores, not even ours. I

20:34

don't know. I don't know why they would do this. Like

20:36

why do they care if a thing that emulates windows

20:38

and DOS games is on a third party app store?

20:41

This doesn't make any sense to me. And maybe there

20:43

is a completely fair and logical reason, but I can't

20:45

put my finger on it. If so, if there is,

20:47

I feel like they would have communicated it to the

20:49

UTM authors. You know what I mean? They would have

20:51

said, here's why. Cause like, you know, the virtualization framework

20:53

you're using has a security flaw and it would allow

20:55

people to root phones or something like just say that

20:58

if it's the case, but they haven't.

21:00

Not a great look. Not not a great look at all. All right.

21:03

Apple and meta could face charges for violating

21:05

EU tech rules. Apple and meta could soon

21:07

face charge. I'm sorry. This is from the

21:09

verge. Could soon face charges from the European

21:11

commission for violating digital markets, actor DMA rules.

21:14

The commission is reported to be targeting

21:16

apple over its steering rules that charge

21:18

developers were pointing to third party purchase

21:20

options. Metas charges will reportedly

21:22

revolve around its ad free subscription for

21:24

Facebook and Instagram and the EU. The

21:26

commission will be using preliminary findings according

21:28

to Reuters, meaning that the companies can

21:31

make changes to try to correct things

21:33

before the commission makes final decision. Apple is set

21:35

to be charged first Reuters reports and the financial

21:37

time says we could see the charges in the

21:39

coming weeks and they can charge something

21:41

like 5% of annual revenue or something like that.

21:43

So this is like pretty

21:46

serious money. If, if, if they choose

21:48

to go that deep in it. Yeah,

21:50

this is a leak. I mean, we talked about this right

21:52

after apple rolled out its DMA compliance

21:54

and we said the EU is,

21:57

is investigating to see if what apple

21:59

did is actually compliant. And

22:01

we're coming to the part where we're going to get that

22:03

answer. And it seems like this is a leak from the

22:06

EU to say our answer is going to be no. What

22:08

Apple did is not compliant. This doesn't even

22:11

mention things like rejecting UTM from third party

22:13

app stores. This is just talking

22:15

about the steering rules

22:17

and taking money from developers who are

22:19

going to third party purchase pages and stuff like that.

22:21

We'll see what the ruling says. They're always kind of

22:23

slow to move and a little bit back in time.

22:26

Like they can't keep up with all the violations that

22:28

Apple is doing. Yeah,

22:31

that's always the risk with these things is they

22:33

make a rule. Apple says they comply. And then

22:35

the EU takes its time to say, but have

22:37

you really complied? And their ruling is coming. And

22:40

it doesn't look good for Apple. Say it again

22:42

like you mean it. Yeah. Yeah.

22:44

It's not looking good for Apple. And

22:46

I mean, I

22:48

don't know. I have such mixed feelings

22:50

about this. And depending on when you catch

22:52

me, sometimes I think the EU is being

22:54

a bit heavy-handed. And then I'll tell you

22:56

10 minutes later that Apple deserves everything it's

22:58

getting. Right now I'm leaning

23:00

more towards, well, you kind of deserve it. But

23:02

ask me again in 10 minutes, like I said.

23:05

Well, but I mean, that right there, that is

23:07

the risk of failing to

23:09

self-regulate and creating. Yes. Yes. Creating a

23:11

need for governments to step in. Because

23:13

when governments do step in, they're not

23:15

going to get everything right. They're going

23:17

to do things because these are largely

23:19

not technology people. Certainly, whatever

23:22

technology people who talk to the government and influence

23:24

them are going to be only from a certain

23:26

side of it. So when governments

23:28

are forced to regulate tech, they

23:31

don't always do what's good for everybody or what's good

23:33

for us in the industry or what's good for our

23:35

customers. And that's the risk. But

23:38

for Apple failing to self-regulate to an acceptable

23:40

degree for all these years, and

23:42

I think getting worse over time in a lot of these

23:44

areas, they

23:46

have invited the government regulation

23:49

risk by

23:51

their own failure to self-regulate. And that's

23:54

the risk of doing that. Like, yes, they've made

23:56

some extra money here. On their various app stores.

24:00

cuts and anti-competitive behavior they've done there,

24:02

but they also provoked governments to regulate

24:04

them. Now they have to accept the

24:06

consequences of that. I think it would

24:08

have been a better long-term strategy to

24:11

hold back a little on the anti-competitive

24:13

behavior and maybe avoid some of this

24:15

regulation. But I mean, hey, I'm not

24:17

the CEO of Apple. So they

24:19

didn't take my advice, obviously, and we'll see how it

24:22

turns out. But that's, you know, they rolled the dice

24:24

themselves and this is what they got. And

24:26

they're still kind of betting that their compliance,

24:28

people call it malicious compliance. It's not quite

24:30

that bad, but it's like, can we plausibly

24:33

comply with this in a way that makes

24:36

all of the alternatives they're trying to introduce

24:38

unattractive? And as we've discussed in the episode

24:40

about the DMI, DMA compliance, even if they

24:42

were complying with the letter of the law

24:44

here, they are not complying with the spirit.

24:46

They have worked very hard

24:48

to arrange things to make the alternatives

24:52

basically impossible for them to be more attractive

24:54

than what Apple offers because of the rules

24:56

that Apple itself makes. They made rules to

24:58

make all the other options at

25:00

best on equal footing with Apple's,

25:02

but most of the time, you know, worse. And that

25:04

is not the spirit of the law. The spirit of

25:06

the law is supposed to allow competition. It's not supposed

25:08

to allow Apple to make a set

25:11

of rules that doesn't allow anything better

25:13

to ever exist. And, you know, and

25:15

rejecting apps like UTM from third party

25:17

app stores is just icing

25:19

on the cake. So, so far they've been betting

25:21

they can get away with this. The penalties are

25:23

supposedly huge, but like all government things, this everything

25:26

happens slowly. We've waited how many months for the EU

25:28

to say whether they're

25:30

compliant. They're probably going to say that they're not. And

25:32

who knows how much longer we'll have to wait after

25:34

that for Apple to say, okay, well, what about now?

25:37

Now are we compliant? And this could just go on

25:39

for ages. So, you know, the

25:41

wheels of government move slowly. We

25:44

are brought to you this episode by

25:46

Squarespace, the all on one website platform

25:48

for entrepreneurs to stand out and succeed

25:51

online. Whether you're just starting out or

25:53

managing a growing brand, Squarespace makes it

25:55

easy to create a beautiful website, engage

25:57

with your audience and sell anything from

25:59

your products to your. content to your

26:01

time, all in one place and all

26:03

on your terms. Squarespace makes it super

26:05

easy to make websites and they are

26:07

especially good at making storefronts and business

26:09

websites. So they had this brand new

26:12

guided design system, Squarespace Blueprint. You start

26:14

with a professionally curated layout and styling

26:16

options and you build a unique online

26:18

presence from the ground up tailored to

26:20

your brand or business and optimized for

26:22

every device. You can easily launch a

26:24

website and get discovered quickly with integrated

26:26

optimized SEO tools so you show up

26:28

more often to more people and grow

26:30

the way you want. When your customers

26:33

are ready to buy from you, you

26:35

also have amazing checkout and flexible payments

26:37

options. Simple and powerful tools that

26:39

you accept credit cards, PayPal, Apple Pay and even

26:41

in some countries offer the customers the option to

26:43

buy now and pay later with after pay and

26:45

clear pay. There is so much

26:47

more to look at at Squarespace. It's

26:49

all powered by their new next generation

26:52

Fluid Engine. It's never been easier with

26:54

the Fluid Engine to unlock your unbreakable

26:56

creativity. You choose your website starting point

26:58

and you customize every design detail. It's

27:00

reimagined drag and drop that works for

27:03

desktop and mobile so you can really

27:05

stretch your imagination online with the Fluid

27:07

Engine. Go to squarespace.com to

27:09

try all this out with a free

27:11

trial. I am so strongly recommending Squarespace.

27:13

It is great. It will almost certainly

27:15

fit all of your needs plus more

27:18

and it's so easy. Anyone can use

27:20

it. Go there, start a free trial.

27:22

You can build the whole site in trial mode.

27:24

You can see for yourself how great it is.

27:26

When you're ready to launch, go to squarespace.com/ATP to get

27:28

10% off your first purchase of

27:30

a website or domain. Once again,

27:32

squarespace.com for that free trial. When you're

27:35

ready to launch, squarespace.com/ATP to save 10%

27:37

off your first purchase of a website

27:39

or domain. Thank you so much

27:41

to Squarespace for sponsoring our show. Chiefment

27:48

unlocked. I was a little,

27:50

little itty bitty speck in the keynote,

27:52

in the actual WWDC keynote. We

27:55

were seated fairly far away from the screen, which

27:57

is fine. I'm not complaining. My

28:00

eyes are not good enough to have been able to discern,

28:02

uh, you know, the, the wall

28:04

in the wall of icons of vision pro

28:06

apps. Does mine exist there? And I mean,

28:09

given that there's not too many, you would think so. And

28:11

so that was all of them on screen. Uh,

28:14

and eagle eyed, uh, viewers have pointed out to

28:16

me and I think John, uh, I think you

28:18

made this a very helpful image. No sarcasm. I

28:20

did make it for you. Thank you.

28:23

Uh, but there, uh, at the top of the

28:25

screen is call sheet. So I was actually with

28:27

my friend Steve that did that icon. Actually did

28:29

the, the kind of default one for iOS. Steve

28:31

did the, uh, default one for vision

28:33

OS. And I was with him

28:35

just an hour ago and we were both sharing a happy

28:38

moment about how we had made it into the keynote. So

28:40

that's very, very exciting. And I was very pleased to see

28:42

that. That's awesome. Congrats. Thanks.

28:45

That was the good news. That was the good news.

28:48

You Casey now. Yes. Now

28:50

some not so great news. So Apple TVs insight feature,

28:52

which is like the Amazon X-ray thing, uh, which for

28:54

a brief window of time, I thought completely Sherlock to

28:56

me. And then the more I've learned, the more I

28:58

think that's not true. I say, is I knock

29:00

on wood. Um, anyways, that

29:03

feature apparently will also use

29:05

your iPhone, which admittedly like taking off my

29:07

selfish hat for a second. That's the sounds

29:09

really slick. So reading from nine to five

29:11

Mac when using the existing remote

29:13

app on iOS, Apple

29:15

will populate your iPhone's display with the info

29:17

provided by insight. This means you won't need

29:20

to obstruct your view on the TV with

29:22

insight panel, but instead you can view and

29:24

interact with insight entirely on your iPhone. Again,

29:26

taking off my selfish hat. That is very

29:28

freaking cool. Yeah. I thought I

29:30

thought you were totally safe because no one wants to junk

29:32

up their screen with stuff that blocks the view. When someone

29:34

wants to ask something, they should just look it up on

29:36

their phone. Obviously call she is still way

29:38

more full featured than the inside feature and has tons of tons

29:41

of information, you know, anyway, but

29:43

still Apple did apparently provide a way

29:45

for you to look up this information

29:47

without jumping up your screen. This

29:49

next one, uh, I did not see when I

29:51

was going through the show notes literally three hours

29:54

ago. There's going to be another Britley breaking news,

29:56

uh, and it makes me miserable. So apparently insight

29:58

isn't using like. metadata on Apple

30:01

TV or Apple, yeah, Apple TV

30:03

plus provided media. It's

30:05

using ML to identify actors and songs. Please tell

30:07

me Danza. This

30:09

is very late breaking news from a

30:11

digital trends YouTube channel that I watch

30:13

for TV reviews and they're

30:15

doing like a news segment and they're always talking

30:17

about Apple TV. I think this is info straight

30:19

from Apple. Apparently that instead of doing what Amazon

30:21

does, which is of course higher armies of people

30:24

to watch every single show on their streaming service

30:26

and manually annotate when every single actor is on

30:28

the screen so that the little pop like that's

30:31

what they do. It's like who's visible on the

30:33

screen now? Who's in this scene? X-ray

30:35

shows that and the Amazon video

30:37

thing. I thought that's Apple was doing that.

30:40

They'll just do it for their own shows because I believe

30:42

this is limited to Apple TV plus. But

30:44

according to this video that we will link in

30:46

the show notes, apparently no, they're using machine learning

30:48

to identify both the actors faces and the Shazam

30:50

thing to identify the songs because that's all they

30:52

show. And this is very limited feature. They show

30:55

the actors who are on screen and if there's

30:57

music, they show the song that's playing. And,

31:00

I mean, what they said in the video that

31:02

we'll link is that that means insight

31:05

could potentially be sort of an OS

31:07

wide thing on TV OS and not

31:09

just an Apple TV plus because if

31:11

they're just looking at the video and

31:14

identifying faces, you know, and it

31:16

also means, well, okay, is it going

31:18

to be able to identify pieces faces of people when

31:20

they have, you know, makeup on

31:22

or they're dressed as like a, you know, a fantasy

31:25

creature or something or their backs

31:28

to the camera or they're

31:30

in shadow. I'm really curious

31:32

to see a if this

31:34

is actually true and be how well this

31:36

feature works as compared to the let's just

31:38

brute force it method of Amazon of having

31:40

people enter all that information. Yeah,

31:43

this makes you very sad. I mean, well, again,

31:46

as a user, that sounds incredible and really,

31:48

really useful. But as a developer of a

31:50

competing product that that's making me hear a

31:52

very, very sad trombone and it makes Apple

31:54

happy because they don't want to spend money.

31:57

And hey, we don't have to pay hundreds of people to

31:59

watch. thousands of hours of content and manually annotate it,

32:01

why don't we just let computers do it for us? Indeed.

32:06

We'll see what happens. But selfishly,

32:08

I hope not. Unselfishly, yeah, let's go. All

32:11

right, so you had cast about John. I think it was you, John.

32:13

Maybe it was Marco, but I presume it was John. It was all

32:15

of us. It was all of us. There you go. How

32:17

do we disable sports? Well, it wasn't me, because I actually

32:20

kind of like this. Anyway, how did disable? Do you like

32:22

the sports? You didn't chime in. Do you like them? You're

32:24

like, ooh, I didn't know that my favorite college

32:26

football team, something happened to them. Does

32:29

that happen? I like it on

32:32

the occasion. It's a team that I care about.

32:34

Now it will, if memory serves, not even I

32:36

don't other than F1. I'm not really

32:38

paying attention to sports at the moment. So

32:41

anyways, if it's a team that I don't care about

32:43

or especially a sport, I don't care about. Then I

32:45

find it frustrating and annoying like I think you guys

32:47

do. But if it's at least

32:49

a sport that I care about, like let's

32:51

say it's a close game between two teams.

32:53

I don't really care that much about that.

32:56

I don't mind that so much at all. But so like

32:58

if it's two college football teams, for example, that are

33:00

having a close game, maybe I'll be interested in that.

33:03

But if it's two, I don't know,

33:05

baseball teams that are having a close

33:07

game could not care less. See, here's

33:09

why this thing has irritated me so

33:11

much. People were going to say,

33:13

oh, just unfollow all the sports teams. You

33:15

must have launched Apple Sports and followed sports

33:17

teams. And so I went in there and, nope,

33:20

sure didn't. Then some other

33:22

people said, oh, you have to go unfollow your

33:24

local teams in the Apple News app under their

33:26

sports section. So I went there and, nope,

33:29

I didn't have anything there either. This

33:32

is literally just an opted in

33:34

feature that they added to TVOS

33:36

for, I think, everyone. Because nothing

33:39

about any of my Apple activity would suggest that

33:41

I ever want to watch or ever have watched

33:44

sports using an Apple device. So

33:46

I guess this is

33:48

just opted in for everyone. And what

33:51

dressed me nuts is like Apple is, or

33:53

at least used to be, so careful

33:56

about respecting the user

33:58

experience when doing something. something really

34:00

immersive, like watching a drama on

34:03

TV. Like, I've been

34:05

sitting there watching these like, you know, really serious

34:07

TV shows, often by the way, not even using

34:09

the Apple's TV app, often in other apps like

34:11

the Macs app. And then

34:13

this sports pop-up pops up in the corner, like

34:16

I never would have enabled that. I never did enable

34:18

that. I've never watched sports on an Apple device. I

34:20

don't follow any sports teams on any device, let alone

34:22

an Apple device, because I don't follow any sports. They

34:25

are just intruding upon the sanctity

34:27

of a full-screen TV episode, a

34:30

drama that I'm watching on my

34:32

TV with their premium

34:34

experience, allegedly, premium experience Apple TV

34:36

platform and box. That

34:38

is so against their ethos. That

34:41

is so gross. Apple

34:43

used to never do stuff like that,

34:45

and there are so many paper cuts

34:47

like this creeping into their products in

34:49

the effort of ever-increasing services, engagement, and

34:51

revenue. It is really irritating me. And

34:54

it just seems like the standards of the company

34:56

around things like this are just going

34:59

down and down and down over time, and it makes me

35:01

sad. Yeah, you can see why they would want to

35:04

throw this in people's face, because no one's ever gonna

35:06

find it, like

35:08

to turn it on manually. But the normal

35:10

way to do that, which is still irritating,

35:12

but is way better than what they did,

35:14

is on first startup after

35:16

the OS update, or on the first time you

35:19

go back to the home screen after the OS

35:21

update, pop up a one-time thing, you hope it's

35:23

one time. Pop up a one-time thing, this is,

35:25

hey, by the way, it looks like you just

35:27

upgraded to CVOS123. There's

35:29

a new feature that will show you sports scores. When

35:32

something exciting happens, do you want to enable that? Yes, no. But

35:35

they didn't do that. They just literally, apparently,

35:37

he turns it on, and as Marco mentioned,

35:40

those places that he mentioned to look, you

35:42

should look there, because I did add favorite

35:44

teams to the sports app, and I'm like,

35:46

that must be it, so I deleted them.

35:49

And I looked at Apple News, and I didn't have anything there. But

35:52

the actual location of this feature, thanks

35:54

to Jason Snow, is in TV OS,

35:56

go to Settings, Apps,

35:59

TV. And they can find

36:01

the item that says exciting games in the

36:03

notification section and turn that off That

36:05

is the thing that they turned on for you No one

36:07

would ever find that on their own to turn it on

36:10

manually Which is why you'd have to prompt them on first

36:12

boot after the new OS or whatever but they just turned

36:14

it on from everybody as Marco mentioned like of all the

36:16

things to Intrude on a

36:19

full screen watching television

36:21

experience that they

36:23

need to become like sort of company-wide guidelines that

36:25

like look unless someone has Manually

36:27

opted into it. You cannot ever pop up anything on the

36:30

screen Let's like their houses on fire like the home kit

36:32

can pop up things if there's a smoke alarm going off

36:34

I'll allow that or like a security camera, but other than

36:36

that Sports scores are

36:38

not the same as your house on fire Yeah,

36:41

I mean I can totally understand why this

36:43

would be very frustrating and I do not

36:46

like how buried it is Let me just

36:48

repeat what you said settings apps TV exciting

36:50

games That's not where I would

36:52

look to find this at all. But what are you

36:54

gonna do? All right, we

36:57

got a phenomenal flex from Matthew

36:59

Willoughby who is extremely excited about

37:01

finally being able to get a

37:03

rest day and they sent us

37:05

a Picture of an Apple

37:07

Watch ultra where it says you've received

37:09

this award This is your longest move

37:11

streak. You've received this award for your

37:13

longest move streak, which lasted 3,000

37:16

three hundred and thirty eight days. This

37:18

was sent on such a day that I guess

37:21

John you computed that 3,338

37:24

days before when this was sent was April

37:26

23rd of 2015. What does that have any

37:29

significance? That was the day the

37:31

original Apple Watch first launched on the previous episode

37:33

I said hey this move streak thing where you

37:35

are allowed to have rest days I bet there's

37:37

someone out there who got a series zero watch

37:39

and has had a move streak going since that

37:41

day and somehow didn't lose

37:43

their streak every time they upgraded watches and

37:46

it's just Thankful that finally they

37:48

can get a rest day. This person truly does

37:50

exist. It's Matthew Willoughby Yep,

37:53

this he's wearing an Apple Watch ultra So

37:55

he has upgraded from the series zero through

37:57

a series of Apple Watch us to the

37:59

ultra He is his moves has lasted the

38:01

longest that any Apple watch move Street could

38:03

possibly last for someone who bought this watch

38:05

at retail because that was literally Day one

38:07

of the release of the Apple watch. It's

38:09

impressive. It's very impressive Yeah, he says and

38:11

I quote all caps. I can finally have

38:14

a rest day. Yes, Matthew you can you've

38:16

earned it I'm not sure that's what he said. I

38:19

think what he said is I can finally have

38:21

a rest day Something

38:23

more along those lines. I'm Apple should fly someone

38:25

to his house Like

38:28

like you like the the gamification

38:30

of fitness, I think he won the game.

38:32

Yeah Well

38:36

done Alright Apple ID

38:38

has been renamed to Apple account

38:40

as was foretold as the prophecy

38:42

foretold So Apple in one

38:44

of their newsroom post says with the releases

38:46

of iOS 18 iPad OS 18 Mac OS

38:49

Sequoia and watch OS 11 Apple ID

38:51

is renamed to Apple account for a

38:54

consistent sign-in experience across Apple services and

38:56

devices and relies on a user users

38:58

existing credentials. They did it is

39:00

gonna now so they did this and now Make

39:04

a marker listener here. How many years will it take us

39:06

to not say Apple ID on this? Infinite

39:09

infinite years it's gonna be a while Alright,

39:12

let's start doing some OS

39:14

based follow-up and let's start

39:16

because John wrote pretty much all

39:18

these show notes Where do you think we're gonna start? We're

39:20

gonna start with Mac OS So tell

39:22

me about Mac OS 15 Sequoia with with

39:24

the Apple pencil and iPad inside car mode

39:26

So what's the ask here? What are we

39:28

talking about? Technically we start

39:31

with TV OS because we wanted to get your

39:33

your glory slash sadness in but anyway Yes, Mac

39:35

OS is the next one up. So we're talking

39:37

before like we're talking about math notes Like

39:39

do you think you could use math notes with handwriting? In

39:43

Sequoia if you use an iPad inside karma, we're

39:45

like they should do that. That would be really

39:47

cool And apparently

39:49

that already works You know

39:51

not in Sequoia, but if you take your

39:54

iPad and use it inside car mode as a

39:56

second monitor, which I did with my iPad and

39:58

have an Apple pencil You can

40:01

go into for example the notes app and scribble

40:03

yourself a little sketch on your iPad

40:05

Even though you're using the Mac OS version of

40:07

notes because then you're just using your iPad as

40:09

a secondary screen But it's basically a touchscreen on

40:12

your Mac It even does the hover effect with

40:14

the cursor and everything that already works now. I

40:16

didn't try it with math notes on Sequoia Mostly

40:20

because math notes on Sequoia and the very first

40:22

developer beta is super duper buggy But

40:25

I'm hoping this will mean that if you really

40:27

want to hand write math notes on your Mac

40:29

And you have an iPad you can do it

40:32

that is very cool. That's super neat Chris

40:34

Chellberg writes with regard to iPhone

40:36

mirroring in Sequoia is

40:38

iPhone mirroring a way to rearrange your iOS

40:40

home screen more easily I don't know possibly

40:42

right yes I hadn't thought of it But

40:45

like I would much rather use a mouse

40:47

pointer than my finger because a it doesn't

40:49

obstruct everything and be I have pixel perfect

40:51

Precision as I try to drag because it

40:53

will still be that weird game of bumping

40:55

around icons and stuff But this

40:57

actually is a big upgrade In

41:00

my ability to rearrange my home screen without pulling

41:02

my hair out We'll see how it goes but

41:04

and also the thing with where the being able

41:06

to leave space I hope it doesn't make the

41:08

icons as squirmy and as you know Collapsible

41:11

as they were before so I

41:13

am actually looking forward to trying this John

41:16

I'm not trying to be funny remind me why

41:18

people like what is what is this network locations

41:20

thing like I've heard of this and I remember

41:22

talking about it having left and everyone was upset,

41:24

but I don't think I've ever used it So

41:26

can you give me like a two-second tour of

41:28

what network locations is please yeah? I don't use

41:30

it either, but I believe the idea is that

41:33

when you are in different locations You

41:35

might have different network setups when I'm at work.

41:38

I use works DNS servers. I have a VPN

41:41

Wi-Fi is ahead of Ethernet in my

41:43

network order like all sorts of stuff

41:45

like that And you don't want to have to

41:47

manually switch all this networking stuff network

41:49

locations I don't know if it's actually location aware,

41:51

but it does let you have a pop-up menu

41:53

that says I'm on this network now I'm

41:55

on my home network. I'm on my work network. I'm on

41:58

my traveling network where I use a VPN or whatever We're

42:00

like, I believe that's what

42:02

the feature is for. The reason it's a story

42:04

is because it disappeared when they redid the settings

42:06

app. When the system preferences became system settings, network

42:08

locations disappeared. The functionality was still in the OS,

42:11

but the GUI for it was gone. And

42:13

the story is now, thanks to Raycat, let us

42:15

know network locations are back in the GUI in

42:18

Sequoia. So if you miss them, they are there.

42:20

They're still kind of buried in the network pane,

42:22

but it's better than trying to do it from the

42:24

command line. We do have some

42:26

sad news though from Rob. Bad

42:28

news, Sarakusa. It looks like the password field is

42:30

still right aligned in the passwords app, at least

42:33

on iOS. And it's the same on

42:35

Mac OS. I launched the, I installed Sequoia. I was using

42:37

the beta, launched the password app. Yeah, it's

42:39

just so weird. Like if

42:41

you stick the insertion point at

42:43

like, the

42:46

beginning of the word, yeah, it is the beginning. It's the

42:48

first letter of the word. And you

42:50

type a character, the character appears to the

42:52

left of your insertion point. That's very weird.

42:54

Which makes sense if you think about it,

42:56

but like when you're typing, it's like, this

42:58

is not how typing should be. Like again,

43:00

I think this is possible on

43:02

the web using modern web technology, but no one

43:05

would ever do this. Like

43:07

why? Just, it's

43:10

literally a form. There are labels and there

43:12

are text fields, username, colon,

43:14

field, password, colon, field. It doesn't

43:16

have to be this hard Apple.

43:18

Just use regular text fields. You're

43:21

breaking my brain. I

43:23

should file a bug on it now. I guess I, you know, you gotta get

43:25

those bugs in early. I mean, I don't know why

43:27

I didn't file it for the year and a half since this,

43:29

or two years, whatever it's been since this. I was like, I

43:31

gotta file it now. Like you made a whole new app. It's

43:33

all new. Can we fix the text fields, please? Our

43:36

friend, M.B. Bischoff writes, with regard

43:38

to Sequoia window tiling, window

43:40

tiling in Sequoia can be disabled or set to happen

43:43

only when option is held down. Also, it can optionally

43:45

leave margins between windows. That's pretty cool. I didn't know

43:47

any of that. Yeah, I'm shocked if there are any

43:49

settings related to this. I was afraid I would just

43:51

have to like find some hidden P list key to

43:54

turn it off. But if you

43:56

can find it in system settings, it's a little

43:58

bit hidden. There are three whole options. to do

44:00

it. And I was playing with the tiling. I

44:03

don't know if I know all the options,

44:05

but it seems not, it

44:07

seems kind of limited. Like I didn't, it didn't seem

44:09

like a lot of flexibility for the tiling where I

44:11

can do eighths and thirds and grids and stuff like

44:14

that. It's more like just left half of the screen,

44:16

right half of the screen, top bottom limited

44:18

top bottom stuff. Because when you do the top, you go,

44:20

you end up going to the spaces thing. But anyway, margins

44:22

between windows. I am, I am a big fan of, but

44:25

I know some people don't like it. So Hey, make a

44:27

toggle. If you don't like margins, no margins. If you do

44:29

like them, there they are. It doesn't let you adjust the

44:31

margins, which they should. And

44:33

my dream of having full windows server access to

44:35

make a real window manager continues to be a

44:37

dream. But for now, all those people who like

44:40

a window style windows tiling, they will

44:42

get thrown a bone in Sequoia. We

44:45

have a miracle. A miracle has

44:48

happened in Sequoia. Steve Trout and Smith

44:50

noticed that there's an

44:52

update to the chess app.

44:54

Yes. And MacOS there's been a chess

44:56

app, what forever, I think, or at least in, in my,

44:58

I believe it was there next as well, but don't quote

45:00

me up. It was in next step. It's a very old

45:03

app. Right. So apparently the renderer

45:05

has been updated. And according to us,

45:07

to Steve Trout and Smith, it's

45:10

the first time since MacOS 10.3 Panther, which is

45:12

21 years ago. My goodness. Wow.

45:18

We'll let we'll link to a, a post from

45:20

a cable sash that shows the old renderer and

45:22

the new one, the new one I'm assuming is

45:24

using a reality kit and it looks nice. And

45:26

the old one looks super dated. And remember the

45:28

old one was new and

45:30

MacOS 10 10.3 21 years ago.

45:34

So the chess app, not only will the chess app

45:36

not die, it's seems like it's getting more update than

45:38

a lot of other apps and

45:41

MacOS these days. Indeed.

45:43

Uh, with regard to iCloud key chain

45:45

and browser integration in Sequoia, Jonathan Freese

45:47

noticed that a clean install of MacOS

45:49

Sequoia has what appears to be pre-installed

45:52

extensions for passwords. And, uh, John, you've

45:54

put in a bunch of Jason in

45:56

our show documents. So can you talk

45:58

about this, please? Yeah, I

46:00

he just sent like the the path

46:03

to like You know

46:05

a JSON file and it was like

46:07

slash library slash Google slash Chrome slash

46:09

native message hosts slash com dot Apple

46:12

Dot password manager dot JSON like okay But

46:15

does that say that they're the extension

46:17

is pre-installed or is that just information about where they

46:19

might get it? And I looked at it and it's

46:21

if you look at the JSON It's got a name

46:23

and a description and it's got a path and the

46:25

path is to the Cryptex thing I think we talked

46:28

about on the past shows the system has these Cryptexes

46:30

Which are like sub images that

46:32

they're allowed to be overlaid on top of the

46:34

cryptographic We secure OS image so that the combination

46:36

of them is also cryptographically secure and there the

46:38

Cryptexes are so that Apple can update those Separately

46:41

from the whole OS so they don't have to

46:43

do a full OS update when one little thing

46:45

changes And it looks

46:47

like it's called password

46:49

manager browser extension helper.

46:52

Is that the full extension? It

46:55

looks like it. I mean it's a dot

46:57

app. It's you know dot app contents Mac OS

46:59

But I think that Apple is literally bundling

47:01

the iCloud You

47:03

know like Jonathan says the the iCloud

47:05

keychain extensions for Chrome

47:08

and for Firefox I

47:10

guess I should I mean I had limited time with

47:12

the developer beta It seems pretty solid, but like a

47:15

lot of the new features are buggy or entirely missing

47:17

like the AI stuff So

47:19

I'll have to you to look into this, but that

47:21

would be a bold move Pre

47:23

shipping like you know really just saying hey

47:26

You know third parties have to make you install

47:28

their browser extensions to use it for their password

47:31

manager, but we can just ship them Well everyone

47:33

who gets this yeah, and honestly I think well

47:35

in one respect I think it's a good idea

47:37

because you get splaining to anybody that like you

47:39

know You can

47:42

use like lucky chain. It does two-factor does this does

47:44

that hey? They don't know what I call a keychain

47:46

is and they say oh But

47:48

I use Chrome no problem

47:51

just Install the

47:53

iCloud keychain Chrome extension. I

47:55

would never trust someone to

47:58

just some you know someone who's not a tech nerd

48:00

to be able to find the correct non scammy real

48:02

live Apple iCloud key chain. I don't even know what

48:04

it's called. I have to check a hundred times before

48:06

I install it to make sure I'm, this is actually

48:09

from Apple or is this just going to steal all

48:11

my passwords? Pre-installing in the US is the

48:13

right way to go. Uh, must

48:15

be nice to be the platform owner. Indeed.

48:18

You can, you can just make these things so

48:20

much easier. There is

48:22

incredibly great news for a lot of

48:24

developers, particularly Mac OS developers, uh, for

48:26

the longest time, if you wanted to

48:28

have a virtual machine on your computer,

48:30

maybe of an old version of Mac

48:32

OS or something like that, uh,

48:35

you couldn't sign into iCloud. So that means

48:37

on this VM, you couldn't do anything that

48:39

relates to iCloud, you know, because it wouldn't

48:41

let you sign in. And that's still true

48:43

of, um, of

48:45

all the existing or the like already

48:47

released versions of Mac OS, but

48:50

Sequoia virtual machines will allow logging into iCloud,

48:52

which is super duper exciting for those that

48:54

have that need. That's not me personally, but

48:56

that is really great. No, no joke. Uh,

48:58

so reading from our Stecnica, as long as

49:00

your host operating system is Mac OS 15

49:03

or newer and your guest operating system is

49:05

Mac OS 15 or newer VMS will now

49:07

be able to sign into, sign into and

49:09

use iCloud and other Apple ID related services

49:11

just as they would when running directly on

49:13

the hardware. That's very cool. I think they mean

49:15

Apple account related service. No, that's true. They say already did

49:17

it. Uh, using it

49:20

begins. Uh, then there's a, uh, there's a

49:22

doc on Apple's developer site using iCloud with

49:25

Mac OS virtual machines and reading from that

49:27

doc. When you create a VM in Mac

49:29

iOS 15 from a Mac iOS 15 software

49:31

image and blah, blah, blah, virtualization configures an

49:33

identity for the VM that it derives from

49:36

the security information in the host secure enclave.

49:38

Just as individual physical devices have distinct

49:41

identities based on their secure enclaves. This

49:43

identity is distinct from other VMS. Yeah. So

49:45

this is a thing that pretty much only Apple could

49:47

do it because they are the ones who are the

49:49

keeper of all the software that interacts with this secure

49:51

enclave. The reason it didn't work before is

49:53

because the VMS didn't have access to that. And you need

49:55

that to sign into iCloud. You may be thinking, who cares

49:57

if you can sign into iCloud. So

49:59

many things. so many apps require an

50:01

iCloud login, not just like in-app purchase or

50:03

stuff like that, but anything that uses CloudKit

50:05

or any of Apple's cloud services. It

50:08

was very difficult to do

50:10

what iOS users take for granted of

50:12

being able to have a simulator or a VM where you

50:15

can test your software, especially for a

50:17

Mac OS when you want to support two versions

50:19

back, you had to keep around old Macs running old

50:21

versions. And that's a little bit more cumbersome than keeping

50:23

around old phones, because the Macs are just bigger,

50:25

and you've got to have a keyboard and mouse

50:27

touch them and everything. It would be

50:29

great if we could just do this all in

50:31

virtualization. Oh, but we can't, because even though you

50:34

can and have been able to virtualize Mac OS

50:36

for years, the inability to sign into iCloud put

50:38

a big damper on that. Now, it's not great

50:40

that this requires you have to be running Sequoia

50:42

to run the VM, and the only OS you

50:45

can run in the VM is Sequoia or later.

50:47

So it doesn't help people now, but

50:49

two years from now, you'll

50:51

be able to run the two-year-old version of Mac OS

50:53

in a VM. So I mean, if you quibble with

50:56

how they did it or whatever, but obviously, it kind

50:58

of requires support in both the OS and the VM

51:00

thing, so I can see why it requires 15 on

51:02

both. And in a few years, this

51:04

will fix itself. I'm just glad this finally came. They

51:06

heard the cries of their developers. Next

51:09

stop, Mac OS simulators

51:11

in Xcode? Dare to dream. That

51:14

would be cool. And I mean, there's no reason not

51:16

to, right? The reason is it takes work,

51:18

but hypothetically, there's no reason

51:21

that Apple couldn't do it. All right,

51:23

let's move on to Apple Intelligence. Rick

51:25

Noller writes with regard to Apple Intelligence

51:27

and third party mail apps. Could you

51:29

add your Gmail account to the Apple

51:31

Mail app and then hide it? That

51:33

might allow Apple Intelligence to access that

51:35

info, but you wouldn't have to use

51:37

it or see it in mail.app. Yeah,

51:40

I do this on my Mac. I think I mentioned this before. I

51:43

have a weird arrangement. Apple Mail on

51:45

my Mac, which is an app that I

51:47

do not use, I occasionally launch it, and

51:49

I have my Gmail account configured with this

51:52

ancient Google thing that has been

51:54

in there forever where you can pop

51:56

from your Gmail account. It's

51:59

very strange. still supported, I keep waiting for

52:01

it to break. But the old pop protocol

52:03

where you just say, what's new since the

52:05

last time I checked and it downloads the

52:07

messages individually, right? That exists for Gmail. And

52:09

I use it to basically make a network

52:11

backup of my Gmail, right? An incremental network

52:13

backup. So I launched Apple Mail every once

52:16

in a while, it downloads all the messages

52:18

into a pop local mailbox and I just

52:20

stick them into a folder called Gmail Archive.

52:22

It's the tertiary backup

52:25

of my Gmail. And

52:27

what that means is Apple

52:30

Intelligence on, not my Mac, but on my future

52:32

Mac will be able to have access to my

52:34

mail to use that to give it the context

52:36

that it needs to do smart things. But on

52:38

my phone, I don't do that. And

52:41

so the suggestion from Rick is something I'm

52:43

gonna try. I configured my Gmail account,

52:46

not with pop, but just the default way you can configure

52:48

it on your phone. I haven't figured out

52:50

how to hide the account. I just ignore it and

52:52

I'm into my other inbox. I'm not in like the

52:54

all inboxes view or whatever. And

52:56

I hope what that means is that my phone will

52:58

also see my mail. Now, unfortunately,

53:00

I don't ever wanna use

53:02

Apple Mail and I'll set my, you know, my default

53:05

email client is set to the Gmail app. But I

53:07

hope this means that Apple Intelligence, again, not on my

53:09

current phone, because it's only a 14 pro, will

53:12

be able to be smart about

53:14

my email, even though I don't use Apple Mail.

53:17

All right, Xcode AI code

53:19

completions in the documentation for

53:22

Xcode 16. It reads

53:24

Xcode 16 includes predictive code completion powered by

53:26

a machine learning model specifically trained for Swift

53:28

and Apple SDKs. See objective C,

53:30

predictive code completion requires a Mac with Apple

53:32

Silicon and 16 gigs

53:35

of unified memory running Mac OS 15. So

53:38

another double whammy. So A, Xcode

53:40

16, if you want the cool code completion, you

53:42

have to also be running Mac OS 15, which

53:44

is not always been the case. Usually

53:46

when there's new versions of Xcode, you can run them on

53:48

the current stable OS and

53:50

they work and have all the features, but not

53:53

this year. And the second thing is,

53:55

hey, you want code completion? Hope you didn't buy a

53:57

base model MacBook or MacBook Pro with eight gigs of

53:59

RAM. because the cool smart

54:01

code completion requires 16 gigs

54:03

of memory. In other words,

54:06

it requires a Copilot plus PC. Ooh. All

54:08

right. Yeah, that's tough. But I

54:11

mean, here we are. With

54:13

regard to more with regards to the

54:15

hardware requirements, David Steer writes, I'm

54:17

curious as to why Apple Intelligence works on

54:20

M1 chips, but you need an A18 Pro

54:22

to use it on iPhone. If I recall

54:24

correctly, the M1 is roughly equivalent to the

54:26

A15 Bionic, which means anything after iPhone 13,

54:28

including iPhone SE third generation, could possibly

54:30

have the necessary power, but crucially, not

54:33

the required amount of RAM. Do you

54:35

think it's possible that Apple's notoriously stingy

54:37

RAM provision could be coming back to bite

54:39

them in the era of AI? It's true that

54:41

lack of backwards compatibility could help them manage server

54:43

capacity or drive customers to upgrade their devices, but

54:45

they'll need to balance that with the difficult messaging

54:47

that their new flagship features are not available to

54:49

the vast majority of their user base. Yep, 100%.

54:52

It's the RAM. I mean, Gruber asked that

54:54

on the talk show, and they were typically cagey

54:56

about it, but they did confirm. It's lots of

54:58

factors, including the RAM. It's 100% the RAM. That's

55:01

why the 15 Pro can do it and the other

55:04

ones don't. Yes, the 15 Pro does have a better

55:06

neural engine than the 14 Pro, but not that

55:08

much better. But what the 15 Pro

55:10

has is more RAM. And it

55:12

seems to me that getting any of

55:14

this stuff to work on a phone with the limited

55:16

amount of RAM that's on a phone, because I believe

55:18

the 15 Pro has, what, 8 gigs? I

55:21

believe that's right. So we just got done saying

55:23

that on the Mac, to get a

55:26

somewhat pedestrian AI feature, like

55:29

code, a completion of where I write some

55:31

code for you or whatever, you need 16 gigs

55:33

of RAM. And they're trying to get this to work on a

55:35

phone with 8. They cannot get

55:37

it to work on a phone with 6, apparently. RAM

55:41

is the thing here. If you bought a base model

55:43

MacBook Air, after hearing people rave that

55:45

you can do everything on this. You can buy an 8

55:47

gig MacBook Air. You can do Xcode. You can do all

55:49

your development. Well, you don't get the AI features that you

55:51

want from the new version of Xcode, because you need 16.

55:54

And we discussed on a past show how the rumor was

55:56

that all of the phones for this year in September are

55:58

going to have 8 gigs of RAM. of RAM, this

56:00

is why, right? And even the eight gigs is

56:02

probably pushing it because that's the same amount as the 15 pro.

56:04

Why didn't they go to nine or 10 or 11 or 12

56:06

or 16, whatever. AI

56:10

eats Ram. It needs a lot of it.

56:12

Ram takes battery. Ran takes space. Ram

56:15

produces heat like Apple

56:18

has stingy because they're cheap, but also stingy

56:20

because especially in a portable device, Ram has

56:22

a cost to it. Well, now they're rolling

56:24

out tons of AI features. And I don't

56:27

think Apple wants the only

56:30

phone that can run this stuff to be the

56:32

very tippy top flagship. Because even the iPhone 15

56:34

can't run it. Only the 15 pro can run

56:36

it. That is not ideal. It's nice on the

56:38

Mac that it can run all, you know, all

56:41

the AI features can run all the way back

56:43

down to the M1. So kudos to that those

56:45

teams, but like the Mac, you know, Macs have

56:47

more stuff. They have more battery, they have more

56:50

memory, they have more CPU. And even though the

56:52

X code code completion requires 16 gigs, Apple intelligence,

56:54

broadly speaking, as far as I know, does not

56:56

require 16 gigs. So a lot of the Apple

56:58

intelligence features will work with eight gigs, just not the

57:00

X code thing. It seems, but yeah,

57:02

this is, this is definitely chickens coming home to

57:05

roost. Apple being stingy with Ram seems like it's

57:07

fine. Every time someone probably made an internal argument,

57:09

we should put more Ram that say, that's fine.

57:11

I'll show you. Like we even got some inside

57:13

information ages ago. They said we did

57:15

test with, it was just about SSDs and it

57:18

turns out one SSD chip isn't actually that bad.

57:20

It's not noticeably worse than having the two SSD

57:22

chips. So we didn't do it. They

57:25

reversed that decision because of, you know, presumably public outcry,

57:27

or did they just want to hear people whine about

57:29

it anymore? But all the previous cases

57:31

where they said actually eight gigs of Ram is fine.

57:34

They need to keep up with

57:36

the pace of the industry. Even if it doesn't seem

57:38

like it's strictly necessary, they can lag behind a little

57:40

bit, but you can't just ignore it forever and saying

57:42

there will never be another thing that we need to

57:44

do that requires more Ram than we have now. Here's

57:47

AI saying, guess what? We found a

57:49

use for all that Ram that always happens.

57:51

There's always something over the horizon that

57:53

requires more resources. Usually it's just games.

57:56

Honestly, they're always games will always eat everything you

57:58

give them. Right? But sometimes there's

58:00

applications that everybody uses, although at this point

58:03

everybody games to some degree or another. Something

58:05

is going to want those resources. The

58:08

computers are never fast enough and never have enough

58:10

RAM. So you have to keep up with the

58:12

industry. You can't say we finally plateaued. Computers will

58:14

never need more than eight gigs of RAM. It's

58:17

AI now, who knows what it'll be in 20 years. Apple,

58:20

please give us RAM. I think

58:23

this also should inform your

58:25

purchasing decisions of Macs over

58:28

time. If you buy a

58:30

Mac today and you want it to still have

58:32

cutting edge features for as long as possible, this

58:35

is a pretty big reason to not just leave it at

58:37

the eight gig default. Even 16, because LLMs,

58:39

as we know them, are

58:43

giant RAM hogs, and because we don't

58:46

really know what the future will hold with

58:48

features, there might be some

58:50

really killer feature that comes out in

58:52

two or three years that requires 16

58:54

gigs of RAM or more on

58:56

a Mac to actually be usable. And

58:59

even if Apple doesn't do it, someone else might. So

59:01

this should inform your purchases

59:04

even today. Sometimes

59:07

you can kind of peer into the future and be

59:09

like, well, I think we're on the cusp of something

59:11

that's about to need a lot of resource X. In

59:14

this case, we are there right now for

59:16

RAM. We are in

59:18

the early days of something that needs a lot

59:20

of RAM, and so maybe for your next Mac

59:22

purchase, get a little more than you otherwise would

59:25

have. Apple's AI training

59:27

data at machinelearning.apple.com, and we'll put the

59:29

full URL in the show notes, of

59:31

course, it reads, we train our foundation

59:34

models on licensed data, including data selected

59:36

to enhance specific features, as well as

59:38

publicly available data collected by our web

59:40

crawler, Applebot. Web publishers have

59:42

the option to opt out of the

59:45

use of their web content for Apple

59:47

Intelligence Training with a data usage control.

59:49

We never use our users' private personal

59:51

data or user interactions when training our

59:53

foundation models, and we apply filters to

59:55

remove personally identical information, like social security

59:57

and credit card numbers that are publicly

59:59

available. on the internet. We also filter

1:00:01

profanity and other low quality content to

1:00:03

prevent its inclusion in the training corpus.

1:00:05

In addition to filtering, we perform data

1:00:08

extraction, deduplication, and the application of a

1:00:10

model-based classifier to identify high quality documents.

1:00:13

So this is kind of the same answer that they

1:00:15

gave on stage. We saw them twice talk about this,

1:00:17

I think once in the iJustine interview, and then again

1:00:19

on the talk show. And

1:00:22

as I said before, and I'll say again, their answer

1:00:24

to how do you train your AI models is not

1:00:26

great. It could be worse. I like the idea they're

1:00:28

saying we are not using your private or personal data.

1:00:30

We're not training anything that you do. And

1:00:33

they do use license data and so on

1:00:35

and so forth, but they also always have

1:00:38

this item. They say we use publicly available

1:00:40

data. Oh, you can

1:00:42

opt out. Well, it's kind of hard for us to opt

1:00:44

out when this is the first we're hearing of the fact

1:00:46

that you're training AIs on our data. So you already got

1:00:48

it. You already did it. You already trained. I think Applebot

1:00:51

is a pre-existing thing, and

1:00:53

we could have been blocking things with robots. That

1:00:55

text or whatever. But if you care about this,

1:00:57

if you know that Apple was training its AI,

1:01:00

you wouldn't have maybe been blocking Applebot. Because you're like, oh, Apple's

1:01:02

not doing anything like that. I don't have to worry about it.

1:01:05

Now, setting aside the legality

1:01:07

and ethics of training on

1:01:09

this type of data, I'm

1:01:12

still kind of surprised that Apple didn't

1:01:15

use one of its greatest resources, money, to

1:01:17

pull an Adobe and say, we

1:01:19

do not train on publicly available information. We

1:01:21

train only on information that we licensed. Cut

1:01:24

deals with the people that have the information.

1:01:26

Make a licensing deal with the New York

1:01:28

Times. If you have to do with

1:01:30

Wikipedia, with Encyclopedia Britannica, with

1:01:32

just having sort

1:01:36

of consensual data sharing relationships. But apparently,

1:01:38

that is insufficient to train something as

1:01:40

complicated as what they're attempting to do.

1:01:42

So they've been training on publicly available

1:01:44

information. When they're crawling the web, looking

1:01:47

for stuff and throwing it through this engine.

1:01:49

And how that's going to shake out in

1:01:51

the US anyway has yet to be determined.

1:01:53

Because those cases are winding in their way

1:01:56

through the court system. I don't think Apple's

1:01:58

really pinning itself into a corner here. because

1:02:00

there are other companies that are much worse

1:02:02

off. If it turns

1:02:04

out that training on

1:02:07

publicly available information requires some

1:02:09

kind of legal arrangement or whatever, Apple will

1:02:11

just make those legal arrangements. They're not doomed

1:02:13

or anything like that. It just kind of

1:02:15

surprises me that they didn't take an even

1:02:17

more conservative approach than they have. And so

1:02:19

it forces Apple executives to be on stage.

1:02:21

And they have to say, we

1:02:23

use this, we use that, and we also use publicly

1:02:25

available information. If you press them and say, what is

1:02:27

publicly available information, web pages,

1:02:29

web pages that are on the web, your web pages.

1:02:31

Do you have a blog? Is it on the web?

1:02:34

Does it have stuff in it? We probably trained our

1:02:36

AI on it, unless it contains profanity or credit card

1:02:38

numbers or whatever, right? That

1:02:40

is not an easy thing for a company

1:02:42

like Apple to talk about or say. And

1:02:44

in this case, they can't say, oh, but

1:02:46

open AI. We push all the problems off

1:02:48

onto them, because this is Apple training its

1:02:51

models, its foundation models, and presumably

1:02:53

its future chat GPT competitor on

1:02:56

publicly available information. All

1:02:58

right, so what if you don't want your

1:03:00

website to be included in Apple's AI models?

1:03:02

Well, there's a knowledge-based document that we will

1:03:04

link in the show notes that describes how

1:03:06

you can modify your robots.txt in order to

1:03:08

tell it to kindly bugger off. Yeah, just

1:03:11

go back in time and do that like

1:03:13

two years ago. Yeah,

1:03:15

exactly. But we'll put a link in the show notes

1:03:17

if you're interested. Tim Cook did

1:03:19

a interview on The Washington Post, and

1:03:22

The Washington Post asked a lot of questions.

1:03:24

John, you've extracted a few. Do

1:03:27

you want me to play the role of Tim and you'll play the

1:03:29

role of The Washington Post? You can't do both

1:03:31

of them? I can. I thought we would play it. Play the space.

1:03:34

Can you do a Kim Cook's accent? I can't do it.

1:03:36

Good morning. Exactly.

1:03:38

All right, The Washington Post asked, did you take

1:03:40

any special delight in calling it Apple intelligence as

1:03:43

opposed to artificial intelligence, to which Tim replied, it

1:03:45

seems sort of a logical conclusion after looking at

1:03:47

so many names. At least for me, I can

1:03:49

tell you it wasn't a riff off of artificial.

1:03:51

Artificial, it tell it intelligence. Wow, I'm struggling. It

1:03:53

wasn't? Yeah, right. It

1:03:55

was sort of calling it what it is. I'm

1:03:57

sure a lot will be said about it. But

1:03:59

it's probably not. as it appears. Yeah. As far

1:04:02

as Tim is concerned, it's just like, Oh, it's

1:04:04

an Apple feature. That's intelligent. It has nothing to

1:04:06

do with AI. Okay. Oh yeah. Sure. Okay. I

1:04:08

mean, at least for him,

1:04:10

he's saying his perspective. Uh-huh. Uh-huh.

1:04:13

Uh, Washington Post asked, what's your confidence

1:04:15

that Apple intelligence will not hallucinate? To

1:04:18

which Tim replied, it's not a hundred percent, but

1:04:20

I think we've done everything that we know to

1:04:22

do, including thinking very deeply about the readiness of

1:04:24

the technology in the areas that we're using it

1:04:26

in. So I'm confident that it will be very

1:04:28

high quality, but I'd say in all honesty, that's

1:04:30

short of a hundred percent. I would never claim

1:04:32

it's a hundred percent. So this is interesting because,

1:04:34

you know, as people have noted, like when they,

1:04:36

you opt into open AI, it, you know, you

1:04:38

have to explicitly say you're sending your data to

1:04:40

them. And when, when it comes back, there's a

1:04:42

little disclaimer underneath it. Like there isn't all these

1:04:44

things that says, you know, check

1:04:47

important information or this information might not be correct

1:04:49

or so on and so forth. But this question

1:04:51

was not about open AI. It was not about

1:04:53

the open AI integration. It was, wasn't about that

1:04:55

screen that has the disclaimer at the bottom. It

1:04:57

was about Apple intelligence. And because Apple

1:04:59

does have models, the chances of Apple's own stuff,

1:05:01

not open AI, the stuff that stays on

1:05:03

device or goes to Apple, super duper private servers.

1:05:06

What about that? What does that have a chance of hallucinating? As

1:05:08

I continue to futilely point out,

1:05:10

hallucinating is a terrible term because there

1:05:13

is no distinction between a quote hallucination

1:05:15

and a quote, correct answer. The

1:05:18

AI model is functioning the

1:05:20

same way in both cases. It's not like, Oh, it

1:05:22

made a mistake or it got like, we know from

1:05:24

the outside, we can judge whether we think this is

1:05:26

a good or bad thing, but the, the

1:05:28

internal machinery is operating as expected

1:05:30

in both cases. Everything from an

1:05:32

LOM is either not a hallucination

1:05:35

or a hallucination. There's no special

1:05:37

like, Oh, it did a bad,

1:05:39

did a good like it's just

1:05:41

a big stew of words and

1:05:43

probabilities. And it's a machine that

1:05:45

is deterministic. That is it

1:05:47

fed various inputs and various parameters that

1:05:49

tweak it and you get output from it. And

1:05:52

every time it does that, it is exactly as

1:05:54

correct or exactly as incorrect as every other instance.

1:05:57

But anyway, what is this thing is can

1:05:59

that machine produce stuff that is

1:06:01

not useful because it doesn't fulfill

1:06:04

the purpose. If you're asking a question, did it

1:06:06

give you an answer that is correct? Did

1:06:09

it actually give you something helpful? Did it understand what you

1:06:11

were asking for and give it to you? Like we

1:06:13

can judge what it does. And

1:06:16

this is Tim Cook saying some of the

1:06:18

things that we ship as Apple Intelligence might

1:06:20

not do the right thing. Now for

1:06:23

Genmoji, what

1:06:25

do you care? You ask for you know you

1:06:28

know a turtle skiing down the Alps and

1:06:30

the turtle doesn't look like a turtle. Who

1:06:32

cares? Right? No harm, no foul. Summarization

1:06:35

I guess it could get wrong if you ask

1:06:37

it to summarize something and its

1:06:39

summary is not

1:06:41

accurate in a way that is significant. Like

1:06:44

if it's a big article about like you know how

1:06:47

much detergent you're supposed to use in the dishwasher and the

1:06:49

article has a whole bunch of paragraphs talking about all the

1:06:51

things that you shouldn't do and then at the end tells

1:06:53

you what you should do and the summary decides

1:06:56

that the summer the best

1:06:58

summary of this article is one of the bad things that

1:07:00

the article is somehow concluding that you should do one of

1:07:02

the things that is actually saying you shouldn't do. I would

1:07:05

call that in Tim's parlance here, a hallucination because

1:07:07

you ask it to summarize the article and it

1:07:10

didn't it didn't summarize the

1:07:12

article the way a human would because it

1:07:14

didn't understand the point of the thing and

1:07:16

if you rely on that summary you'll put

1:07:18

the wrong amount of detergent in your dishwasher

1:07:20

or whatever right and

1:07:22

this is another awkward position for Tim Cook to be

1:07:24

put in. This is a good question from the Washington

1:07:26

Post to say you're rolling out a bunch of these

1:07:28

features are they going to essentially

1:07:30

malfunction and not do what they're supposed to do

1:07:32

in a way that's not like oh we'll file

1:07:34

a radar and we'll fix it like as you

1:07:36

know Appleship software that doesn't do it all the

1:07:39

time but those are bugs and they can fix

1:07:41

them. There's nothing you can do when this

1:07:43

happens you could send this to Apple and say hey

1:07:45

it summarized this thing wrong and they'll try

1:07:47

to make it better next year but it's not as simple

1:07:49

as like this caused a crash or

1:07:51

you know this thing was misaligned or a cosmetic

1:07:53

error or whatever this is a different realm or

1:07:56

Apple is shipping products that might

1:07:59

not function correctly and that there's nothing Apple

1:08:01

can do about it except for like try harder next

1:08:03

year with their next model that is trained on more

1:08:05

Data that they slept from the web So

1:08:08

then the Washington Post asked what makes you

1:08:10

think opening I and Sam Altman specifically are

1:08:12

trustworthy partners who share Apple's values Very well

1:08:14

a very well phrased and a very good

1:08:16

question to which Tim replied They've

1:08:19

done some things on privacy that I like They're

1:08:21

not tracking IP addresses and some of the things

1:08:24

like that that were very keen on not happening

1:08:26

I think they're a pioneer in the area and

1:08:28

today they have the best model and I think

1:08:30

our customers want something with world knowledge Some

1:08:33

of the time so we considered everything and

1:08:35

everyone and obviously we're not stuck on one

1:08:37

person forever for something We're integrating with other

1:08:39

people as well, but they're first and I

1:08:41

think today it's because they're best It's

1:08:44

another uncomfortable situation people keep asking Apple about open AI.

1:08:46

It's uncomfortable that feature exists It's uncomfortable that open AI

1:08:48

is what it is Tim is trying to put a

1:08:50

part of a spin on it said they're doing some

1:08:52

things with privacy like you know why they're doing that

1:08:54

Probably because Apple forced them to like this. I don't

1:08:56

think they were doing that on their own I

1:08:59

think if you use a chat GPT through

1:09:01

the web there They are tracking IP addresses

1:09:03

and other stuff like that, but whatever deal they

1:09:05

did with Apple, you know, it just There's

1:09:08

there's baggage that comes with Sam open and open

1:09:10

AI and Apple I

1:09:13

guess they feel like you know This is this is

1:09:15

on Tim because you know He's the person to give

1:09:17

a go-no-go on this type of thing saying this is

1:09:19

what everyone's talking about We need to

1:09:22

have it. We don't have any of this We can't don't

1:09:24

have anything that matches this internally and in fact, we think

1:09:26

it's a little bit dangerous But either way

1:09:28

we have no choice We have to partner and you

1:09:30

know the Google deal didn't go through and they're still

1:09:32

keeping the door open say it's not just open AI

1:09:35

We'll partner with whoever but yeah They

1:09:37

have to partner with this company and put a

1:09:39

disclaimer on their little thing It's the opt-in and

1:09:41

there's a disclaimer that says it might be wrong

1:09:44

But they feel like they need to have it

1:09:46

because they think that's what people want I

1:09:48

think they're right that that's what people want I think the

1:09:50

whole story would have been that they still don't have an AI

1:09:52

chatbot But I think the

1:09:55

utility of AI chatbots is still Maybe

1:09:58

not it's kind of like when something is good enough to

1:10:00

work in a Disney context. I know Apple and Disney aren't

1:10:02

the same thing, but when is it good enough to work

1:10:05

in an Apple context where we have more

1:10:08

expectations about ... I don't

1:10:11

know. The cultural expectations of an Apple product are

1:10:13

different than a PC type

1:10:16

thing or an Android phone. It's

1:10:21

more cautious about ... The App Store

1:10:23

is not the same

1:10:26

as the open web. Let's put it that way. Apple

1:10:28

runs the App Store in a more cautious

1:10:31

way than the open web runs, which is

1:10:33

bad and good in some ways. But

1:10:35

here they are. They're like, we have to provide

1:10:38

this option and we have to awkwardly answer every

1:10:40

press question about it by saying, people want it.

1:10:42

They're the best one. We're not stuck with them.

1:10:44

Next question, please. The

1:10:47

aforementioned Steve Trouton-Smith asks with regard

1:10:49

to Siri unsupported devices.

1:10:52

Steve writes, I still

1:10:54

have so many questions about Apple intelligence. Does

1:10:56

Siri just not get better on anything below

1:10:58

an iPhone 15 Pro? No improvement to the

1:11:00

cloud-based Siri on older devices, HomePods, Apple TVs?

1:11:02

It's a good point. I

1:11:05

don't know the answer to that. I don't

1:11:07

think anyone asked Apple that and they should have. Obviously

1:11:10

the older devices can't run

1:11:12

Apple intelligence on them. Certainly a HomePod can't.

1:11:15

Talk about RAM limits, right? Which

1:11:17

by the way, like for a device

1:11:19

that is all about Siri and

1:11:21

that is so reliant on Siri

1:11:24

and is so bad

1:11:26

with Siri, that's the

1:11:29

device that needs it the most. It's such a shame

1:11:31

that this is not going to get it. It makes

1:11:33

me wonder is like, is a new HomePod coming with

1:11:35

eight gigs of RAM? Actually, I shouldn't say that. I

1:11:37

don't actually know how much RAM a HomePod has. Do

1:11:39

you know how much RAM a HomePod has? I have

1:11:41

no idea. 128 kilobytes. I don't

1:11:43

know. Wow. Maybe it does have eight gigs.

1:11:45

But anyway, here's the thing though,

1:11:48

for all the devices, like every iPhone except

1:11:50

for the 15 Pro, it's

1:11:53

not like Apple can't make Siri better

1:11:55

for them because they get like the whole,

1:11:57

the whole magic of their. strategy

1:12:00

for Apple intelligence is That

1:12:03

they run it on device if they can and if

1:12:05

they can't they run it on Basically

1:12:07

a logical extension of your device. That's the

1:12:09

whole private cloud computing thing It is like

1:12:12

a bigger iPhone processor that is not in

1:12:14

the room with you And

1:12:16

so that's the the the cloud the transparency

1:12:18

of like we'll run device or run there

1:12:20

But the software doesn't care where it runs

1:12:23

because it's basically running on Apple silicon with

1:12:25

you know of a different size, right? Runs

1:12:28

a big Apple silicon in our data centers runs on

1:12:30

smaller Apple silicon on your phone. It

1:12:32

could be That on devices

1:12:34

like the home pod or every phone except

1:12:36

for the 15 Pro It

1:12:39

could just send everything to private cloud computing which obviously

1:12:41

would be slower and there'd be latency and so on

1:12:43

and so forth But at this point tons of serious

1:12:46

stuff most of the serious stuff goes

1:12:48

across the network. Anyway, that could

1:12:50

make Siri smarter on Devices

1:12:52

that can't run Apple intelligence locally. I would vastly

1:12:55

prefer that I would let my home pod Now

1:12:57

is currently sending things over the network and taking

1:12:59

a long time to do bad things How

1:13:01

about updating it and having it do everything

1:13:03

through private cloud computing to the llm's? I

1:13:05

hope that's what they do, but

1:13:07

I find it not encouraging that

1:13:10

Apple never offered that as A

1:13:13

thing that they're doing everyone was so excited and

1:13:15

jazz about Apple intelligence. No one really said You

1:13:18

know, what about my home pods? Does

1:13:20

you know maybe they just assume we all think we

1:13:22

all understand that's what's going on, but they didn't say

1:13:24

it They didn't say well dynamically choose

1:13:26

if we do it locally or remotely and of course on

1:13:29

devices They can't do it locally. We'll do it all remotely.

1:13:31

They didn't say that at all So we

1:13:33

will all find out when we start installing

1:13:35

the betas that actually have Apple intelligence features

1:13:37

in them on our unsupported devices Apple

1:13:40

silicon in Apple's private cloud computing servers There's

1:13:42

a blog post about this which I did

1:13:44

not have a chance to read but I

1:13:46

presume John you at least glanced at it

1:13:48

You know I need I need some AI

1:13:50

to To summarize this for

1:13:52

me. Yeah. Well, so this was a

1:13:54

question in the talk show live which we talked about before you

1:13:56

can Pull link to the YouTube videos. You can

1:13:59

check it out Or the vision

1:14:01

pro video if that's your if you have one get it

1:14:03

out of its little marshmallow Grier,

1:14:05

it's a pretty big marshmallow. Yeah, Grier

1:14:07

asked them What

1:14:10

what is the Apple silicon that is running

1:14:12

on Apple's private cloud compute servers?

1:14:14

Like what is it and predictably they did not

1:14:16

answer that question, right? It's Apple silicon

1:14:18

There was no way they were gonna answer that yeah,

1:14:21

we talked about the okay I like maybe three or

1:14:23

four shows ago. We talked about the

1:14:25

rumor that said that Apple is Going

1:14:27

to run data centers with m2 ultras in them

1:14:29

or whatever and we thought that doesn't make any

1:14:31

sense And again the private cloud computing thing explains

1:14:33

that we talked about it in the WWDC episode

1:14:35

It now does make sense and using m2 ultras

1:14:37

also makes sense because this is a

1:14:40

chip they already designed The

1:14:42

speculation has been that there hasn't been

1:14:44

enough time For Apple to

1:14:46

make any kind of custom server chip they

1:14:49

like they they sort of decided to do this

1:14:51

too late because the lead times on silicon are

1:14:53

so long and Depending

1:14:55

on how many how successful this effort is and

1:14:58

how many servers they need It

1:15:00

takes a certain number of products to justify

1:15:02

making a custom chip and m2

1:15:04

ultra Despite being very power efficient and

1:15:07

having lots of compute and so on and so forth

1:15:09

is not the ideal chip

1:15:12

for Doing private

1:15:14

cloud computing AI stuff. It doesn't need

1:15:16

the h265 decoder and encoder on there

1:15:19

It probably doesn't need thunderbolt since it doesn't

1:15:21

have persistent storage Like it's

1:15:23

not this is not a a

1:15:26

purpose-built Server chip lots of

1:15:28

other companies do have purpose-built server chips companies that

1:15:30

run huge amounts of servers like Google and Facebook

1:15:32

and Amazon right Apple as Far

1:15:35

as we are aware on the outside even

1:15:37

though they do run servers. They don't

1:15:39

have their own Dedicated

1:15:42

server silicon. I feel like if they did they

1:15:44

would have bragged about it So the rumor about

1:15:46

them using m2 ultras makes sense There's a bunch

1:15:48

of things on that ship that aren't being used

1:15:50

and are just being wasted I think

1:15:52

you know, maybe two or three years ago

1:15:54

They started the project to make a dedicated

1:15:56

server chip and that'll come out in

1:15:59

two or three more years after that. But

1:16:02

that is something to watch for because, you know,

1:16:04

I can tie everything to the Mac Pro. It's a situation

1:16:06

where they have to make a weird custom chip that has

1:16:09

a limited application whose needs are different than

1:16:11

all the needs of their other chips because

1:16:13

servers different than in the same way that

1:16:15

Mac Pro is different. Can they bring themselves

1:16:18

to do that or will it just be

1:16:20

will their server farms just be the dumping

1:16:22

ground for the unsold cheap

1:16:24

to produce inventory of the two

1:16:26

years ago? Pretty good Mac studio

1:16:28

chip. All right. Apple

1:16:30

and open AI aren't paying each other yet,

1:16:33

says Bloomberg. This is

1:16:35

reported on Bloomberg and then covered on the

1:16:38

verge. But Gurman says Apple isn't paying open

1:16:40

AI as part of the partnership. Instead, Apple

1:16:42

believes pushing open AI's brand and technology to

1:16:44

hundreds of millions of its devices is of

1:16:46

equal or greater value than monetary payments. Chat

1:16:49

GPT will be offered for free on Apple's products,

1:16:51

but open AI and Apple could still make money

1:16:53

by converting for users to paid accounts. Today, the

1:16:55

user subscribes to open AI on an Apple device

1:16:57

via the chat GPT app. The process

1:17:00

uses Apple's payment platform, which traditionally gives

1:17:02

the iPhone maker a cut. Then

1:17:05

back to the verge or recapping. The report

1:17:07

also says this deal isn't exclusive to open

1:17:09

AI and that Apple is in talks with

1:17:11

him, Tropic and Google to offer their respective

1:17:13

chatbots as an alternative option with an agreement

1:17:15

for Google's Gemini expected to be in place

1:17:17

later this year. Yeah,

1:17:19

Gruber also asked that on stage, hey, which way is

1:17:22

the money flowing in the situation? They obviously didn't answer.

1:17:25

This makes some sense.

1:17:27

Apple is in the power position

1:17:29

here. Apple essentially owns billions

1:17:31

of customers who have shown a willingness to

1:17:33

spend money. Open AI wants

1:17:35

access to those customers. Apple's

1:17:38

offer to them is we will literally build

1:17:40

you into the operating system. You couldn't ask for

1:17:42

a better customer acquisition tool than this. It's

1:17:45

your job to convert those people by having a good

1:17:47

enough product that people want to pay for it or

1:17:49

whatever. It makes sense that there's no money changing hands

1:17:51

either way, because on the other hand, open AI can

1:17:53

say, well, you don't have anything like what we have.

1:17:55

And we're the market leader. And

1:17:57

in the end, those two things cancel each other out. According to this

1:17:59

rumor and. then nobody pays anybody and open

1:18:02

AI helps to do lots of conversions. And Apple as always

1:18:04

hopes to take a lot of 30 to 15%. Indeed.

1:18:11

All right. And then the same rights

1:18:13

on the impact of Apple intelligence. Do

1:18:15

you think that the arrival of Apple

1:18:17

intelligence is pulling dev time away from

1:18:19

the workflow features and shortcuts, or

1:18:21

might they used to be rolled together in a future

1:18:24

release? I'm sorry for chuckling. It's just, I don't view

1:18:26

those as at all related,

1:18:28

like the sorts of people who work on one.

1:18:30

I mean, I guess in some cases could be

1:18:32

the same people that work on the other, but

1:18:36

from what little I know and from what I know as

1:18:38

a developer, I don't think that people that are really

1:18:40

good at doing the sorts of things that workflow does

1:18:42

are going to be necessarily very good at doing the sorts

1:18:44

of things that Apple intelligence does. I

1:18:46

put this question in here because

1:18:48

a sort of company wide fire

1:18:51

drill effort like Apple intelligence pulls

1:18:53

resources from everything, even if it's

1:18:55

not the same people, because so

1:18:57

many people in so many groups across

1:18:59

the entire company have shifted

1:19:01

their focus because an edict has come

1:19:03

down from on high that, you know,

1:19:05

2024 WWDC is going to

1:19:08

be the coming out of Apple intelligence.

1:19:11

This really does, you know, it

1:19:14

is a company wide tax and it was the right

1:19:16

thing to do. They should have done this. Arguably, they

1:19:18

should have done it earlier, but it

1:19:20

does take away from things. Even, even the problem is

1:19:22

like, even if you're not a person

1:19:24

who was taken off task by doing this, people

1:19:27

you work with were taken off tasks. People have

1:19:29

probably moved around, right? Priorities change. Maybe your thing

1:19:31

that you actually did have time to work on

1:19:33

and finish doesn't even make it into the OS

1:19:35

because it's all hand on deck to debug the

1:19:38

feature that somebody else wrote because that's the important

1:19:40

one to roll out. So, um,

1:19:43

you know, every, and every release priorities

1:19:45

shift around and the thing you might

1:19:47

want to be developed, might not get

1:19:50

the resources that it deserves, but Apple

1:19:52

intelligence is definitely one of those times

1:19:55

every few years where there is

1:19:57

a big movement within the company that has the

1:19:59

potential to impact. every aspect

1:20:01

of the software stack that is

1:20:03

released to WWDC. All

1:20:05

right. Let's talk iOS 18. Um, there

1:20:08

are some indentations on the bezel

1:20:10

in iOS 18 when you engage

1:20:12

Siri, I guess. So

1:20:15

this is, this is really, really difficult to verbally

1:20:17

describe. I'm going to read a little bit from

1:20:19

the verge, but there is a gift link, um,

1:20:22

that you can and should click because no matter how I

1:20:24

describe it, it's not going to make a lot of sense.

1:20:26

So here we go. When you press

1:20:28

the side buttons while running the iOS 18 beta, there's

1:20:31

a clever new animation that makes it look like you're

1:20:33

pushing the bezel into your screen a little bit. At

1:20:35

first glance, there's not much purpose here other than to

1:20:37

add a little whimsy, but it might also be a

1:20:39

practical visual indicator. If Apple eventually releases phones with solid

1:20:42

state side buttons that don't move when you press them.

1:20:44

And again, there's a link in the show notes to

1:20:46

a gift so you can see it in action. All

1:20:49

right. So I kind of get having

1:20:53

a visual indication. You successfully press the button is a

1:20:55

good idea. People should do that on the web. They should

1:20:57

do it in their iOS apps. I am shocked when

1:21:00

this does not happen. I think during WWC, I

1:21:02

was complaining to somebody. It might've been one of

1:21:04

you that the WWC, like the developer app where

1:21:06

you can bookmark, uh,

1:21:09

sessions that you might want to look at later when

1:21:11

you tap the little bookmark icon, it

1:21:13

does not highlight in any way to show

1:21:15

that you've tapped it. And you and I

1:21:17

were talking about this over and over and over

1:21:19

again, and it's not clear whether you're toggling in

1:21:21

on and off or whether you're just saying on,

1:21:23

on, on, or if none of those things are

1:21:25

happening, it's not a good UI. Um,

1:21:28

that said, usually the feedback that

1:21:30

you successfully press the button is that you feel

1:21:32

it go in and out. And

1:21:34

even for like the buttons that don't move

1:21:36

like the iOS, like the iPhone seven home

1:21:38

button, that's what the vibration feedback is to

1:21:40

let you know you have successfully pressed the

1:21:42

button. Now there's

1:21:44

this visual bit of feedback where what it looks

1:21:46

like is you've, you're super strong and you have

1:21:49

dented the side of your phone in momentarily. You've

1:21:51

dented the screen and this little black region from

1:21:53

the side of your screen right next to the

1:21:55

button invades the pixels of the screen and just

1:21:57

goes like, like you're shoving a little black. black

1:22:00

rectangle into the screen region and then letting it go back

1:22:02

out again. I'm not sure I find

1:22:05

it aesthetically pleasing, but I

1:22:07

do like the idea of visual feedback, but

1:22:09

it also makes me worry that the like non-moving

1:22:12

buttons that have been rumored for ages maybe

1:22:14

aren't that great and they need to add this.

1:22:17

That's more clear that you press the button. I don't know.

1:22:19

It is, it looks a little bit weird

1:22:21

to me. I think it looks really cool. Honestly. Yeah. I

1:22:23

agree. I actually thought it looked pretty neat. I mean, obviously

1:22:25

I haven't used it in my own hand, but it looked

1:22:27

pretty slick to me. All right.

1:22:29

Caleb Denman writes that in iOS 18,

1:22:33

you can now change the width of the

1:22:35

beam of the flashlight. It works on the

1:22:37

15 pro 14 pro and maybe any iPhone

1:22:39

with a dynamic Island. And I

1:22:41

read this and I understood the words, but I was

1:22:43

like, I'm sorry, what? How what? And

1:22:46

so a friend of the show, Quinn Nelson has

1:22:48

a helpfully recorded a short little video, which

1:22:51

we will put in the show notes that demonstrates this.

1:22:53

And I, I don't know that this is

1:22:56

incredibly useful, except on

1:22:59

the rare occasions when you're like trying to find

1:23:01

something in the dark room and your partner, whatever

1:23:03

is asleep and you don't want to blind them

1:23:05

with the flashlight going full blast that in the

1:23:07

actually seem, this seems like it could be pretty

1:23:09

useful. The interest of this is

1:23:11

weird. It's like you have like an X and Y

1:23:13

axis for swiping. Like the, the Y axis is brightness

1:23:16

and the X axis is beam with, and since it's

1:23:18

a circle, it's really more like beam diameter. I'm not

1:23:20

actually sure that how they're doing it. Is

1:23:22

there like a matrix of, of LEDs

1:23:24

in there or something? Or I don't know. Apparently

1:23:27

it's on, it's on my phone. So I'll try

1:23:29

it out when I get the beta. And

1:23:31

then finally for iOS, uh, Steve Schouten

1:23:33

Smith, as we've mentioned many times writes

1:23:36

there are new APIs to provide a

1:23:38

locked camera capture extension, which must be

1:23:40

launched via button on the lock screen

1:23:42

control center or the action button. It

1:23:44

cannot be launched by swiping sideways and

1:23:46

lock screen. There is a video about

1:23:48

this, uh, from WWDC this year. I

1:23:50

spent some time with my good friend,

1:23:52

uh, Ben McCarthy, and let me tell

1:23:55

you, they were very excited about this,

1:23:57

uh, for their app Obscura.

1:23:59

So you should. definitely, I know

1:24:01

Ben is definitely going to be playing with this soon.

1:24:04

Yeah, that was my question on the WWC episode. Does

1:24:06

swiping, I know you can put the button for third party

1:24:08

things, but does it also work with swiping? Because I thought

1:24:11

it would be weird that the button launches one camera app

1:24:13

and the swiping launches another. That could also be a feature

1:24:15

if you want to have two camera apps and remember that

1:24:17

the swipey one is the default one and the button one

1:24:19

is the third party one. But it does

1:24:21

seem kind of strange. Like they're giving access to the

1:24:24

lock screen. They have an extension to do this. They

1:24:26

could just as easily have a setting somewhere that says,

1:24:28

hey, do you want swipe to all the buttons? And

1:24:30

then you can also activate the camera that you put

1:24:32

in the little, you know, control center button thing or

1:24:35

whatever, but not in the first

1:24:37

beta anyway. And I didn't watch the video, so maybe they

1:24:39

just explicitly say no in iOS 18. It's not going to

1:24:41

be that way. Let's move

1:24:43

on to Marco's favorite platform, vision OS,

1:24:45

vision OS 2.0, vision OS 2.0 features

1:24:50

that were not mentioned in the keynote.

1:24:52

And there's a vision OS 2.0 preview

1:24:54

that Apple has. You can customize your

1:24:56

home view. Finally. You can

1:24:58

now personalize your home view, simply pinch and

1:25:01

hold to jiggle and arrange apps and bring

1:25:03

them to your home view, including those from

1:25:05

your compatible apps folder. Very nice. Finally, you

1:25:07

can see your keyboard in any environment when

1:25:09

you're immersed in an environment vision OS 2

1:25:12

recognizes and reveals your magic keyboard or MacBook

1:25:14

keyboard. So you can keep typing away. Cool.

1:25:17

Does it only recognize those keyboards or does it recognize?

1:25:19

Sure sounds like it. So

1:25:22

like they should have a keyboard

1:25:24

recognizer and say, is that a keyboard? It's

1:25:26

either probably rectangular. You can probably find the

1:25:28

edges, but okay. Guest

1:25:30

user improvements. All I've read about this so far is

1:25:33

what I'm about to read to you, but oh my

1:25:35

word, I'm here for it. So it

1:25:37

says vision OS 2.0 now lets you save your

1:25:39

most recent guests, I and hand data so they

1:25:41

can easily skip their next setup, which

1:25:43

is pretty cool. I don't know how many recent

1:25:45

guests, just one for now. Just one

1:25:48

Steve Troughton Smith writes the enterprise APIs for

1:25:50

vision OS 2, which there's a,

1:25:53

there was a session about enterprise APIs in

1:25:55

vision OS 2 at WWDC. Anyway,

1:25:57

so it's enterprise enterprise APIs include things like.

1:25:59

to the main camera, pass through capture, and

1:26:02

barcode or QR scanning. But only for in-house

1:26:04

or business-to-business apps. You can't ship this stuff

1:26:06

to the app store. Very

1:26:09

weird. It's interesting that they're making concessions for

1:26:11

what appears to be their one enthusiastic customer

1:26:15

base for the Vision Pro right now, which

1:26:17

is enterprises that don't balk at the price

1:26:19

and have apparently come up with useful

1:26:22

applications for a high quality

1:26:24

VRXR headset thing. They're

1:26:27

giving them much more access. If

1:26:31

you need to get access to this hardware,

1:26:33

more direct access to this hardware, to make

1:26:35

this Vision Pro be useful for use on

1:26:37

your factory line or whatever you're having people

1:26:39

do with this stuff, you

1:26:42

can have it. It's just you can't distribute those

1:26:44

things to the app store. That's just your own

1:26:46

private little world. They've always had the enterprise certificates,

1:26:48

and enterprises can have their own little private per

1:26:50

enterprise app store with their own distribution certificate. And

1:26:53

that is one case where Apple has been a lot

1:26:55

looser, because it's not like the whole world. It's

1:26:57

just a very, very narrow use case.

1:27:00

So it's interesting that they are immediately,

1:27:02

essentially, giving up on a lot of the

1:27:04

restrictions for the Vision Pro for

1:27:07

customers that really, really want it. Finally,

1:27:09

The Verge writes, more new

1:27:11

features coming in Vision OS 2.0 that

1:27:13

either weren't mentioned or were flew right

1:27:16

by. Placing app

1:27:18

windows, you can place them further away than

1:27:20

you could before. That's, I guess,

1:27:22

for people with much better vision than I have.

1:27:25

Volumetric windows, which are ones that let you view an

1:27:27

app's content from all sides, they will tilt to face

1:27:29

you. So you can use them while lying down. I

1:27:31

have Marco. The developers can opt out of

1:27:34

this if they want. You'll also be able

1:27:36

to resize them, which is cool. You

1:27:38

can also offload virtual environments. Their icons will still

1:27:40

be there. But if you're sick of Mount Hood,

1:27:42

it doesn't have to take up space anymore. I

1:27:44

am offended by this The Verge, because that's my

1:27:46

favorite one. And by space, they

1:27:48

mean SSD space, right? I

1:27:51

guess so. How big could those be? I

1:27:53

don't know. Challenge accepted. And then while watching

1:27:55

full-screened videos in a virtual environment, you'll be

1:27:57

able to lie down and recenter them above.

1:28:00

Good stuff. That's especially useful for you, Marco. Yeah. Yeah,

1:28:03

because that was one thing I had to keep sitting up for the

1:28:05

talk show live, because regular windows you can just hold down the Home

1:28:07

button, and it'll just center whatever you're looking

1:28:09

at as the center of the view. But

1:28:12

the fully immersive things seem to not

1:28:14

have any up and down rotational

1:28:16

ability, so I'd just keep sitting up taller

1:28:18

than I probably should have been at that

1:28:20

moment to watch the talk show. Then

1:28:23

with regard to Swift, Swift

1:28:26

has been moved out of Apple's GitHub

1:28:28

account, and it is now in its

1:28:30

own account, Swiftlang. S-W-I-F-T-L-A-N-G, Swiftlang. Yeah, Ben

1:28:33

Cohen actually mentioned this on our interview.

1:28:36

So Apple writes on a blog post. But we didn't dive into

1:28:38

it. Yeah, Swift is migrating

1:28:41

to a dedicated GitHub org

1:28:43

at github.com/Swiftlang. This migration

1:28:45

reflects the growth and maturity of the Swift

1:28:47

community and highlights Swift's versatility beyond Apple's own

1:28:49

ecosystems. The migration to the Swiftlang organization

1:28:51

will be phased over the coming weeks and months. Initially,

1:28:54

the Swiftlang organization will include foundational elements

1:28:56

of the Swift projects, such as compiler

1:28:59

and core tools, standard libraries

1:29:01

and core APIs, samples, the swift.org website,

1:29:03

and official clients, drivers, and other packages.

1:29:07

Since the dawn of Swift that has been presented with

1:29:09

the whole world domination joke I made in the interview

1:29:12

as a language that is good for a wide range of

1:29:14

things, and Apple

1:29:17

didn't want it to just be, oh, this is the language you

1:29:20

need to use to write for Apple's devices. They

1:29:22

wanted it to be a general purpose programming

1:29:24

language that everybody can use. Now,

1:29:26

their number one priority has been making it

1:29:29

work for Apple's devices. So it's a 10-year-old

1:29:31

language, and a lot

1:29:33

of the effort in those 10 years

1:29:35

has been spent to make it good

1:29:37

for programming Apple's platforms, which makes sense.

1:29:39

But during that time, there's been server-side

1:29:41

Swift. There's been Swift on Linux. There's

1:29:43

been attempts to evangelize Swift outside the

1:29:45

Apple ecosystem. And all of

1:29:47

those efforts have run into sort of

1:29:49

stumbling blocks of saying, well, but, you

1:29:52

know, our foundation isn't the same that you

1:29:54

ship with, that Apple ships with its platform,

1:29:57

so we have to have our own alternative

1:29:59

version of it. of it or the server

1:30:01

side stuff is a little bit weird and

1:30:04

it's clear that Apple's heart isn't into it

1:30:06

and the Linux version has a bunch of

1:30:08

gaps in interoperability that don't exist

1:30:11

on Apple's platforms. And if

1:30:13

you keep chasing these things down at the very root,

1:30:15

it's like Swift is

1:30:17

an open source project, but

1:30:20

it's at github.com/Apple. That's

1:30:22

where it is. This is an Apple project.

1:30:24

And yes, you allow people to use it

1:30:26

on Linux and we could fork it because

1:30:29

it's open source or whatever, but like people

1:30:31

just didn't have faith that like Apple was

1:30:33

serious about the idea of this being a

1:30:35

language that is as

1:30:37

journal purpose as C. C has a

1:30:39

standards body and you know, has committees

1:30:41

adding features to it and C plus

1:30:44

whatever, but no one company owns

1:30:46

C right. And again,

1:30:48

even though the switch is open source, say, well, nobody owns it. It's open

1:30:50

source. If Apple ever goes evil, just fork it

1:30:53

or whatever, but no one has the staff or the desire

1:30:55

or the ability to

1:30:57

keep developing Swift outside of Apple. And

1:30:59

so moving the language out of Apple's

1:31:01

area into what I assume will be

1:31:03

a different legal entity and different governing

1:31:05

system or at least a different github

1:31:07

organization. Granted, it's still, you know, the

1:31:09

Swift core team is still staffed by

1:31:11

Apple employees and stuff like that. So

1:31:14

it's not as if Apple is giving

1:31:16

Swift away to someone else who's going to parent

1:31:18

it from now on. It will

1:31:20

still be Apple driving this with their money and

1:31:22

their employees. But this

1:31:24

is an important, both symbolic and

1:31:26

practical step to show Apple's dedication

1:31:29

10 years into finally saying, no, we're actually

1:31:31

serious about Swift. You know,

1:31:33

as they said in one of the

1:31:35

slides, replacing C plus plus not replacing

1:31:38

C plus plus on Apple's platforms, replacing

1:31:40

C plus plus period everywhere someday, maybe.

1:31:45

All right. And then we got

1:31:47

through almost everything we wanted to

1:31:49

in the interview with Holly and

1:31:51

Ben. One of the major

1:31:53

things, perhaps the only major thing that we didn't

1:31:55

have a chance to talk to them about, but

1:31:57

we really wanted to was Swift testing. So

1:32:00

we were talking, um, I guess a couple

1:32:02

episodes ago about how we really would love

1:32:05

to see, you know, XC tests kind of go

1:32:07

the way the dodo and swift

1:32:09

testing is the new hotness baby and it looks

1:32:11

pretty good. At least a glance. Yep.

1:32:13

You can, uh, it's a, it's an open source thing. It's been

1:32:16

out for awhile without before WWC. I

1:32:18

could not remember for the life of me which one of

1:32:20

the several testing frameworks it was. I should have just guessed

1:32:22

the most obvious name, which is swift hyphen testing. Glad

1:32:25

to see that happening. It does use macros and macros

1:32:27

are still a little bit slow and X code, but

1:32:29

you know, maybe next year they'll fix that. All

1:32:32

right. Home kit. You can now

1:32:34

pick your preferred home kit hub

1:32:36

reading from the verge Apple home

1:32:39

users can rejoice over an update discovered in

1:32:42

the first iOS 18 beta, the option to

1:32:44

choose a quote preferred home hub quote. This

1:32:46

fixes the problem of your smart home deciding to

1:32:49

run over wifi through a home pod when there's

1:32:51

a perfectly good Apple TV using ethernet sitting right

1:32:53

there. How long is this taking? Yep.

1:32:56

Like, like home kit has always been like, you

1:32:58

don't have to worry about it. We'll intelligently pick

1:33:00

the right thing, but very often

1:33:03

in your home, you know, uh, if you're

1:33:05

a tech nerd, which one of your devices has the

1:33:07

best network connection and the best hardware. And if it's

1:33:09

Apple stuff, most of the time that's the Apple TV.

1:33:12

If you have a recent Apple TV, it

1:33:14

has the highest chance of being plugged into ethernet because

1:33:17

it has an ethernet port if you bought the expensive one. Uh, and

1:33:19

if you keep buying a new one every year, the

1:33:21

processor does occasionally get better. So no, I don't want

1:33:23

my original home pod

1:33:27

to be my home. Right.

1:33:30

The Apple TV is always plugged in, right? My

1:33:32

Apple TV is always going to sleep or whatever,

1:33:34

but I don't just, yes, please. I'm going to,

1:33:36

I'm going to designate my Apple TV as my

1:33:38

home, get hub. And I hope this,

1:33:41

uh, improves matters. Yep. All

1:33:44

right. Let's talk car play. Uh, another one

1:33:46

of Marco's favorite things. Uh, there

1:33:48

are new car play features. Uh, these are,

1:33:50

uh, detailed on a Mac rumors

1:33:52

post, which we will link in the show notes. Uh,

1:33:55

there are contact photos and messages. I don't think it

1:33:57

ever occurred to me that that's not a thing until

1:34:00

I. And I was like, holy crap, that's not a

1:34:02

thing. He said, so here we are. So

1:34:05

that's very exciting. You'll get silent

1:34:07

mode improvements. You can now choose to have

1:34:10

silent mode on your iPhone immediately

1:34:12

turn and automatically turn on or off when the

1:34:14

device is connected to CarPlay. So that's cool. Speaking

1:34:17

of that, again, I haven't driven my wife's car,

1:34:19

so I'm not that familiar with this. But yeah,

1:34:21

I always have my phone on silent. I mean,

1:34:23

the little silent switches in the silent mode. And

1:34:26

sometimes I forget that that means like most

1:34:29

apps will not make noise. Like

1:34:32

I watch YouTube on it and YouTube ignores the

1:34:34

silent switch and just plays audio. But sometimes I'll

1:34:36

play something and be like, why isn't this making

1:34:38

any sound? And it's because it's honoring the silent

1:34:40

thing. So if you have the silent switch turned

1:34:42

on, as a lot of people do with their

1:34:44

iPhones, and you connect to CarPlay, does your phone

1:34:46

refuse to make sound through the car speakers? It

1:34:49

doesn't do like, you know, broop when you send

1:34:51

a text message, for example, it just sends it.

1:34:53

So that's the best I can think of. And

1:34:55

so this, what this would do is say, you

1:34:58

can leave that switch to silent, but if you set the setting,

1:35:00

what it would do is when you connect the CarPlay, it would

1:35:03

be as if you had switched the switch to not silent and

1:35:05

you hear the bloops. I think

1:35:07

that's correct. It is hard for me to

1:35:09

parse this, but I believe that to be

1:35:11

correct. And like another example is, I

1:35:13

think generally speaking, and this could be my own

1:35:15

settings, like my own focus modes and whatever. So

1:35:18

I might be accidentally lying to you. But like

1:35:20

another example is, I don't think there's an incoming

1:35:22

text message tone, right? Well, if you're in silent,

1:35:24

so you'll see the little banner at the bottom

1:35:27

of the CarPlay screen, but there won't be the,

1:35:29

you know, to the standard ding or whatever you

1:35:31

happen to have your text message sent to set

1:35:33

to, if you have

1:35:36

yourself in silent mode. Moving on,

1:35:38

color filters can help individuals with color blindness

1:35:40

to differentiate colors in the CarPlay interface. Voice

1:35:43

control allows you to control CarPlay entirely

1:35:45

with Siri voice commands through a connected

1:35:47

iPhone. Sound recognition is expanding

1:35:50

to CarPlay to provide notifications for driving related sounds,

1:35:52

such as car horns and sirens. And

1:35:54