ATP

00:00:00 ◼   ► I'm happy to report that we have not yet been classified by the US government as dangerously capable. Therefore, you're still able to listen to our show.

00:00:09 ◼   ► Do you think that Apple has classified us as dangerously capable and that's why we aren't getting invites?

00:00:14 ◼   ► No, that's not it.

00:00:16 ◼   ► I think the world of audio-only podcasts was never super high on Apple's list, but it sure is invisible now.

00:00:26 ◼   ► I think they just don't care.

00:00:29 ◼   ► I mean, we should probably talk about the idea of should we really be doing video.

00:00:34 ◼   ► I don't think any of us are super into that idea, but if we want to continue to have that kind of visibility to the rest of the world, that might be a good idea.

00:00:48 ◼   ► But it would so complicate the production of the show and the editing of the show that I don't think it's enough motivation for us.

00:01:00 ◼   ► But I don't know. How do you guys feel about that?

00:01:01 ◼   ► I don't want to pivot the video.

00:01:04 ◼   ► So I think there's a couple of things here.

00:01:06 ◼   ► It's a little rude of me to say this, but I think all three of us have faces for radio.

00:01:11 ◼   ► And so there's that.

00:01:14 ◼   ► Secondly, I know for me and for John, our physical spaces are not really conducive to doing video.

00:01:23 ◼   ► Like, I have done it.

00:01:24 ◼   ► It can be done.

00:01:25 ◼   ► Every time I'm asked a guest on a podcast, which doesn't happen that frequently, to be clear.

00:01:30 ◼   ► I'm not trying to like humblebrag here, but on the occasions that it does happen, I often ask up front, are you doing video for this?

00:01:37 ◼   ► And annoyingly, the answer is typically yes.

00:01:39 ◼   ► And if the answer is yes, I have to go through this whole like internal dance of, do I really want to clean up the rest of the office?

00:01:47 ◼   ► Do I really want to just, do I don't want to do this at all?

00:01:49 ◼   ► Like, it's just not for me.

00:01:52 ◼   ► It's not what I enjoy.

00:01:54 ◼   ► The flip side of that is, I think it's clear that that's where attention is, certainly for Apple, though honestly, I don't really care if Apple pays that much attention to us.

00:02:03 ◼   ► It would be lovely to be able to go to WBDC, but it's really not that big a deal.

00:02:07 ◼   ► They're not going to pay.

00:02:08 ◼   ► Do you think we suddenly start doing video, they're going to pay attention to us?

00:02:11 ◼   ► No, they're not.

00:02:11 ◼   ► Well, no, no, that's fair.

00:02:13 ◼   ► But they seem to ignore everything that is not video these days, or print for that matter.

00:02:17 ◼   ► I mean, and to be clear, like we were never like super reliably getting press access, but it has really turned to zero in recent years, like the last couple of years.

00:02:26 ◼   ► I think it was always zero.

00:02:27 ◼   ► I think the only time it happened was an aberration, otherwise constant zero, which again, I'm fine with.

00:02:32 ◼   ► Honestly, like it does help to a degree that we don't have to worry about losing our press access because we don't have any.

00:02:40 ◼   ► Yeah.

00:02:40 ◼   ► And so we are able to be really honest and really direct.

00:02:44 ◼   ► That's more difficult.

00:02:46 ◼   ► I know when I have had times in my career where I've had press access with Apple, I have worried about losing it.

00:02:53 ◼   ► I don't think that ever made me do anything major, majorly different from how I would have otherwise done it.

00:02:59 ◼   ► But I'm sure it had to have like a minor impact or some kind of like subconscious biasing that like I maybe I would soften things or not go near certain things because I was afraid of of losing that access.

00:03:11 ◼   ► Whereas when you have nothing to lose, it does give a certain degree of freedom.

00:03:16 ◼   ► But I don't know.

00:03:18 ◼   ► And going back to the video thing for a second, maybe maybe a good guiding principle on this for us to keep in mind is like, I don't think any of our audience has asked us for that.

00:03:26 ◼   ► But that's a very good point.

00:03:28 ◼   ► I think that the reason why YouTubers have pivoted into a lot of podcasts is because they're easier.

00:03:33 ◼   ► You don't need to write and edit nearly as much as you do for for other like formats that succeed on YouTube.

00:03:40 ◼   ► If you can just have a casual conversation and do minimal editing and minimal writing, that is much easier.

00:03:46 ◼   ► So there's a reason why YouTubers like it.

00:03:48 ◼   ► And there's a reason why some podcasters enjoy it, because they think they can get YouTube audience with their podcast.

00:03:56 ◼   ► And that's, of course, very tempting.

00:03:58 ◼   ► What podcast wouldn't like more listeners or more of an audience?

00:04:03 ◼   ► And so I think that's why people are doing it.

00:04:05 ◼   ► But if we are happy doing what we are doing, and our audience is happy doing with us doing it this way, and we're all happy with the numbers and how everything is going, I don't think we should feel compelled to have to push into an area that none of us seem like we actually want to do.

00:04:23 ◼   ► Just for the idea of possibly, basically trying to become YouTubers in a way that it seems like none of us really, really have that in us.

00:04:33 ◼   ► Nobody wants to see us.

00:04:34 ◼   ► It's a young, attractive person's game.

00:04:37 ◼   ► It's not for us.

00:04:40 ◼   ► We have, as is tradition, a whole pile of follow-up.

00:04:43 ◼   ► However, I think we might make it through follow-up.

00:04:46 ◼   ► Typically, the week after WWDC, John basically explains to us that we are going to do a full follow-up episode.

00:04:53 ◼   ► But John, to your credit, I think we're going to make it past follow-up.

00:04:57 ◼   ► So let's see what we can do.

00:04:58 ◼   ► Let's see how it lands.

00:05:00 ◼   ► That's a bold statement.

00:05:01 ◼   ► Yeah, it's a bold statement.

00:05:02 ◼   ► That's like, how many shows, especially member specials, have you started, Casey, by saying,

00:05:07 ◼   ► I think this is going to be a short one?

00:05:08 ◼   ► Every time.

00:05:09 ◼   ► Every time.

00:05:10 ◼   ► We don't have much to say about this.

00:05:11 ◼   ► This is up there with, oh, I'm not going to buy that.

00:05:13 ◼   ► Anyway, have you seen the R2 reviews recently?

00:05:17 ◼   ► They're looking good.

00:05:18 ◼   ► Anyway.

00:05:18 ◼   ► Waiting for the R3X and CarPlay.

00:05:21 ◼   ► Tell me about it.

00:05:22 ◼   ► I'm trying to get through follow-up, and I'm derailing myself.

00:05:25 ◼   ► I'm really rooting for Tesla.

00:05:29 ◼   ► Hear me out.

00:05:30 ◼   ► Hear me out.

00:05:30 ◼   ► What?

00:05:31 ◼   ► Hear me out.

00:05:32 ◼   ► That's a hot take.

00:05:33 ◼   ► I'm really rooting for Tesla, Ellipsis, to start shipping CarPlay.

00:05:38 ◼   ► Because I believe in my bones that the second Tesla ships CarPlay, Rivian's going to be

00:05:43 ◼   ► like, oh, yeah, totes, we're right there, too.

00:05:45 ◼   ► Absolutely.

00:05:45 ◼   ► But they want to control every pixel of their amazing experience.

00:05:49 ◼   ► Don't even get me started.

00:05:51 ◼   ► I think once, if Tesla ships CarPlay, have there been any, I know there was rumor about

00:05:57 ◼   ► that a few months back, but is there anything more recently about that?

00:05:59 ◼   ► Not since like a month or two back where they said it's totally going to be talked about

00:06:03 ◼   ► at WWDC, and then it totally wasn't.

00:06:05 ◼   ► Yeah.

00:06:05 ◼   ► Well, maybe it'll be a fall thing.

00:06:07 ◼   ► Maybe it's a never thing.

00:06:07 ◼   ► I don't know.

00:06:08 ◼   ► But yeah, I think you're right.

00:06:09 ◼   ► I think if Tesla ever ships CarPlay, that will put a lot of pressure on Rivian to finally

00:06:14 ◼   ► cave.

00:06:14 ◼   ► But that's probably going to take years.

00:06:16 ◼   ► I do think that the Rivian R2 is probably going to do what the Model 3 and Model Y did for

00:06:23 ◼   ► Tesla, which is like, it's going to make them all of their money, even though we're all going

00:06:28 ◼   ► to think it's fairly boring.

00:06:30 ◼   ► It's also going to be a really good overall car.

00:06:33 ◼   ► And I'm personally very interested in the R3, but that seems further out.

00:06:38 ◼   ► We'll see.

00:06:40 ◼   ► Yeah.

00:06:40 ◼   ► I don't know.

00:06:41 ◼   ► I haven't, I've been away all this past week, as everyone knows, and I haven't looked too

00:06:46 ◼   ► much into the R2.

00:06:47 ◼   ► But the little bits that I've seen is that it is incredible.

00:06:51 ◼   ► And that the places where they, you know, had to skimp in order to make it cheaper, that's

00:06:55 ◼   ► reasonable.

00:06:56 ◼   ► And the places where you really don't want them to, you know, cut corners and cheap and be

00:07:00 ◼   ► super cheap, they didn't.

00:07:02 ◼   ► So it's supposedly real good.

00:07:04 ◼   ► But anyway, speaking of Rivians, let's talk.

00:07:06 ◼   ► It's road trip.

00:07:07 ◼   ► Jason Paul sent us a link to the Utah road trip demo typo, which is excellent.

00:07:12 ◼   ► And the frustration on the dude typing at the iPad is just palpable.

00:07:16 ◼   ► You can see it right on his face.

00:07:17 ◼   ► It's amazing.

00:07:18 ◼   ► That's what you got to watch the video for.

00:07:20 ◼   ► Just watch, you got to watch the guy's reaction after he does the typo.

00:07:22 ◼   ► It's great.

00:07:22 ◼   ► It's the best because he knows exactly what he did.

00:07:24 ◼   ► And he, you could just see his face.

00:07:26 ◼   ► He's like, oh crap.

00:07:27 ◼   ► Yep.

00:07:27 ◼   ► Like, I can't believe this autocorrect just got me right now.

00:07:31 ◼   ► But anyway, it's, it's delightful.

00:07:33 ◼   ► But Florian also wrote in to say, at the time of that ill-fated demo, I was working for the

00:07:37 ◼   ► company that made that software.

00:07:38 ◼   ► Francois Lagunas, the guy who made the typo, was the CTO and is still a dear friend.

00:07:43 ◼   ► He swears to this day that it was the amount of makeup they made him wear on his hands since

00:07:48 ◼   ► they were prominently featured in the over-the-shoulder device close-ups that caused him to fat finger

00:07:52 ◼   ► the title.

00:07:53 ◼   ► From his recollection, they made the team shoot the do-over as soon as the live event ended

00:07:57 ◼   ► so that they had the footage as soon as possible to edit into the published, quote-unquote, recorded

00:08:02 ◼   ► video of the event.

00:08:03 ◼   ► In any case, it worked out okay in the end.

00:08:05 ◼   ► The company ended up getting acquired by GoPro not long after.

00:08:07 ◼   ► And that software still is the architecture powering the video editing featured in their

00:08:13 ◼   ► quick app, Q-U-I-K.

00:08:14 ◼   ► Hmm, cool.

00:08:15 ◼   ► Yeah, it's like a, you know, it doesn't surprise me they did the recording right after because

00:08:20 ◼   ► they've still got the same clothes on.

00:08:21 ◼   ► It's the same room.

00:08:22 ◼   ► It's the same sound, right?

00:08:23 ◼   ► You know, we're just going to fix this up immediately.

00:08:24 ◼   ► You don't go anywhere, guys.

00:08:26 ◼   ► You still got one more thing to do.

00:08:27 ◼   ► So that's fun.

00:08:28 ◼   ► And hand to make up.

00:08:29 ◼   ► You don't think about hand to make up, but hey, if they're going to shoot your hands

00:08:31 ◼   ► in a close-up, kind of put on that hand makeup.

00:08:33 ◼   ► All right.

00:08:34 ◼   ► There has been a little bit of rumbling.

00:08:36 ◼   ► I've heard this kind of, not literally whispered, of course, but I've heard this kind of being

00:08:40 ◼   ► talked about here and there.

00:08:42 ◼   ► And so we can't credit this to anyone, but an anonymous person said, word on the street

00:08:47 ◼   ► is that iOS, or I guess all the OSes.

00:08:49 ◼   ► All the OSes, 26.6, yeah.

00:08:51 ◼   ► Thank you.

00:08:51 ◼   ► All the OSes, 26.6, actually create the new Spotlight Index.

00:08:55 ◼   ► It just creates it.

00:08:57 ◼   ► It doesn't use it.

00:08:58 ◼   ► So if people update their devices when 26.6 is released, the index will already be there

00:09:03 ◼   ► when they update to iOS 27, or whatever, 27, sorry.

00:09:06 ◼   ► Yeah.

00:09:07 ◼   ► I wonder if something is going to keep the index up to date.

00:09:09 ◼   ► Like that was, again, this is sketchy, unofficial info or whatever, but keep an eye out for 26.6

00:09:15 ◼   ► and see if, I guess you'll have to do it the Marco way, like see if your disk space suddenly

00:09:20 ◼   ► goes down by a lot.

00:09:21 ◼   ► But as people were saying, it's going to take like, you know, it's been taking like a day

00:09:25 ◼   ► or two to build, to rebuild the index after installing like, you know, for example, iOS

00:09:29 ◼   ► 27 beta.

00:09:30 ◼   ► But if they do this and they roll it out in the 26.6 update, by the time everyone updates

00:09:35 ◼   ► at 27, it won't have to rebuild the index.

00:09:37 ◼   ► Maybe it will just incrementally, like maybe it'll just create the index and let it sit

00:09:41 ◼   ► there.

00:09:41 ◼   ► And then when you install 27 OS, it will just incrementally update whatever has happened

00:09:45 ◼   ► since it rebuilt the whole index.

00:09:47 ◼   ► But clever idea.

00:09:48 ◼   ► Watch for that.

00:09:50 ◼   ► Can I tell you guys how much I have to resist putting it on my phone?

00:09:54 ◼   ► Yeah.

00:09:55 ◼   ► Because like, because I'm traveling next week and so, or this week.

00:09:58 ◼   ► And so I'm like, I'm going to, I'm resisting.

00:10:00 ◼   ► I should not.

00:10:01 ◼   ► Do it in the car on the way to the airport.

00:10:02 ◼   ► Yeah, right.

00:10:03 ◼   ► I should not have beta one on my phone when I'm about to travel.

00:10:08 ◼   ► But oh my God, I like, honestly, I'm more tempted on my Mac because now that I've had

00:10:14 ◼   ► it on my, my, like my travel laptop, which I am bringing with me, but now that I have

00:10:19 ◼   ► it on my laptop, um, every time I go back to my, my desktop laptop, I'm just like, oh

00:10:24 ◼   ► my God, everything is so much lower contrast.

00:10:27 ◼   ► The toolbars suck.

00:10:28 ◼   ► Everything is blurry.

00:10:29 ◼   ► The icons are dim.

00:10:30 ◼   ► Like everything, like once you are using Golden Gate for even a small amount of time, when you

00:10:38 ◼   ► go back and see Tahoe, you just like, oh my God, what were they thinking?

00:10:43 ◼   ► It, everything is so much better on Golden Gate.

00:10:45 ◼   ► And it's not like, look, it's not perfect.

00:10:47 ◼   ► There's still a lot about this design that I think they need to keep iterating.

00:10:51 ◼   ► And I'm, and I'm sure they will, but Tahoe looks like such an aberration when you, when

00:10:58 ◼   ► you see Golden Gate and get used to it even for like five minutes.

00:11:01 ◼   ► And do you say that because of functionality, because of the way it looks or yes?

00:11:04 ◼   ► Honestly, I haven't had time to do much functionality with it yet.

00:11:07 ◼   ► It's really just the way it looks.

00:11:08 ◼   ► Like the basics of just like how, how like windows and toolbars and icons look and the

00:11:14 ◼   ► menus without all the, all the weird menu icons.

00:11:16 ◼   ► Like they did like a 1.1 for liquid glass, but it's, it's a good 1.1.

00:11:23 ◼   ► Like it's still, it's still most of the same style, but the tweaks they have made combined

00:11:30 ◼   ► to make a pretty big overall improvement.

00:11:33 ◼   ► And every time I'm back on Tahoe or iOS 26 and I see, you know, a bunch of blurry text

00:11:39 ◼   ► fading under a background list bar.

00:11:41 ◼   ► I'm just like, Oh God, please, please let, let me install these betas on my devices very,

00:11:46 ◼   ► very soon.

00:11:47 ◼   ► One more week.

00:11:48 ◼   ► I got to make it one more week.

00:11:49 ◼   ► So maybe I'll get to beta two, but I'm after once beta two is out, I'm, I'm in that's it.

00:11:54 ◼   ► I'm jumping in.

00:11:55 ◼   ► John, when do you plan on properly running?

00:11:58 ◼   ► You typically don't run the betas at all.

00:12:00 ◼   ► Is that right?

00:12:01 ◼   ► Well, no, I'm, it depends when I, when 26 was coming up, I was running all the 26 betas because

00:12:06 ◼   ► I had to, to get my apps to work in 26.

00:12:09 ◼   ► Uh, once 26 came out, I kept running the betas for a while and I said, why am I doing this?

00:12:12 ◼   ► And so I switched to like the mainstream 26, uh, and I've just been going back and forth

00:12:17 ◼   ► on, you know, on that.

00:12:18 ◼   ► Um, and so, yeah, whatever I have, like 26.5.1 or whatever the latest 26 is.

00:12:22 ◼   ► And on that same machine for reasons that I tooted about it and think I mentioned maybe

00:12:27 ◼   ► on the last show, I don't remember.

00:12:29 ◼   ► Um, I ended up having to partition the main drive so I could install 27 beta because if

00:12:34 ◼   ► you install 27 beta on an external drive, Apple intelligence doesn't work because as we've

00:12:38 ◼   ► discussed on past episodes, Apple intelligence refuses to work if you boot from an external

00:12:42 ◼   ► drive for reasons, I think supposedly security reasons, but I don't know the details, but

00:12:46 ◼   ► anyway, that's just the way it is.

00:12:47 ◼   ► So yeah, I've got a 26, uh, 27 beta and 26.5.1, both on the same machine that I just flopped

00:12:53 ◼   ► back and forth between them.

00:12:54 ◼   ► All right.

00:12:56 ◼   ► So let's talk about the password app, uh, or the passwords app, excuse me.

00:13:00 ◼   ► Uh, Hartley Charlton writes at Mac rumors, the passwords app can now automatically update

00:13:05 ◼   ► weak and compromised passwords.

00:13:06 ◼   ► We talked about this some last episode.

00:13:08 ◼   ► Apple describes the system as agentic with Apple intelligence and Safari securely navigating

00:13:12 ◼   ► through websites, signing in and upgrading accounts to strong passwords without the user

00:13:16 ◼   ► needing to intervene beyond an initial tap.

00:13:18 ◼   ► The feature displays as a live activity when active.

00:13:21 ◼   ► Yeah.

00:13:21 ◼   ► This is the last episode.

00:13:22 ◼   ► I was like, there's no way they could be doing this unless the websites support all of

00:13:26 ◼   ► the, you know, well-known URLs for password changes.

00:13:28 ◼   ► And by the way, ATP.fm now supports that as well.

00:13:31 ◼   ► Cause I mentioned in the last show and I was like, oh yeah, we should do that.

00:13:33 ◼   ► Cause we have optional passwords now anyway.

00:13:35 ◼   ► Um, but surely that's how they have to do it.

00:13:38 ◼   ► And like, well, how can they have a button that says fix these 10 bad passwords?

00:13:41 ◼   ► Does it, what does it expect all 10 of those websites to support this standard?

00:13:44 ◼   ► So it knows where the URLs are.

00:13:46 ◼   ► And the answer apparently is no, it's just going to wing it.

00:13:49 ◼   ► It's just going to say, yep, well, this is the website.

00:13:51 ◼   ► I'm going to, I'm going to run a little safari, headless safari in the background.

00:13:54 ◼   ► And I'm going to let this little, uh, LM powered agent try to find its way to the password

00:14:00 ◼   ► change form and change your password for you.

00:14:02 ◼   ► So yeah, as I said in the, uh, the headline for this item, a reference to Casey's, um, uh,

00:14:08 ◼   ► from our nuggets of wisdom, ATP dev member special passwords app does the hard thing.

00:14:12 ◼   ► It says, nope, we're just gonna, we're just gonna try it.

00:14:15 ◼   ► We're just going to do it.

00:14:15 ◼   ► So good luck passwords out good luck.

00:14:20 ◼   ► All right.

00:14:21 ◼   ► Uh, so let me do a little inside baseball to introduce our next follow-up segment.

00:14:24 ◼   ► Uh, I think we've said many times that John is far and away the person, even when he had

00:14:30 ◼   ► a jobby job far away, the person that puts the most effort into our internal show notes

00:14:33 ◼   ► that we run the show off of.

00:14:34 ◼   ► And you can tell that it's sometimes abundantly obvious that this is the case because I'm going

00:14:41 ◼   ► to talk now, starting now about Mac OS 27 and you look at the time that it takes for

00:14:46 ◼   ► this chapter or these chapters.

00:14:47 ◼   ► There's just lots of images.

00:14:49 ◼   ► First of all, second of all, I still, I still bristle when you characterize me having doing

00:14:54 ◼   ► far and away most of the effort, blah, blah, blah.

00:14:56 ◼   ► Just say, I do it all.

00:14:58 ◼   ► Just, just say, I know occasionally you'll put in an item.

00:15:02 ◼   ► I know occasionally Marco puts in an item, but honestly, just say I do it all.

00:15:06 ◼   ► Yeah.

00:15:06 ◼   ► I mean, yeah, we both know that.

00:15:08 ◼   ► I saw that going a totally different direction.

00:15:10 ◼   ► I saw you being like, no, you know, Casey does this, Marco does that, meaning like broadly

00:15:15 ◼   ► for the whole show, but John, cheesy peasy.

00:15:18 ◼   ► Anyway, so take a look at the timestamp, everyone.

00:15:21 ◼   ► We'll see how long the Mac OS section lasts.

00:15:23 ◼   ► I mean, to give an example, show me the text that you typed in the show notes for this week's

00:15:27 ◼   ► episode.

00:15:27 ◼   ► Excuse me, sir.

00:15:28 ◼   ► Excuse me, sir.

00:15:30 ◼   ► I know you're on vacation.

00:15:31 ◼   ► I know you're on vacation.

00:15:32 ◼   ► Not only that, but I am the chief transcriber in chief for WWDC.

00:15:38 ◼   ► Well, mostly because Google Docs cannot handle both of us doing it at once.

00:15:41 ◼   ► 100%.

00:15:42 ◼   ► And I just, I just backed off.

00:15:44 ◼   ► I said, Casey's Casey wants to do this.

00:15:45 ◼   ► He's going to be taping.

00:15:46 ◼   ► I said that to be fair to Casey, he did do the notes transcribing for the keynote, but he

00:15:50 ◼   ► didn't need to do it because I would have done it.

00:15:51 ◼   ► But I'm still saying for first approximation, it's just.

00:15:53 ◼   ► It's all.

00:15:57 ◼   ► And yes, that does mean that Mac OS comes for, but usually Apple does that too, because

00:16:00 ◼   ► they like save iOS for last when they used to go OS by OS, because like iOS is an important

00:16:04 ◼   ► one and they save the good stuff to the end, you know?

00:16:06 ◼   ► Yep.

00:16:06 ◼   ► I hear you.

00:16:07 ◼   ► Well, anyway, take a look at your podcast player, hopefully overcast and see how long

00:16:11 ◼   ► the Mac OS section takes.

00:16:12 ◼   ► I still contend it's mostly just lots of images that take up a lot of vertical space, but we'll

00:16:15 ◼   ► see all I know is, so I am the, the, the chief note taker in chief of the, uh, of the external

00:16:22 ◼   ► show notes and I've got like 10 or 12 bullets for Mac OS 27.

00:16:26 ◼   ► And I've got one each for iOS, iPad OS, TV OS, watch OS.

00:16:30 ◼   ► So we'll see.

00:16:32 ◼   ► And that's a good example.

00:16:33 ◼   ► By the way, Casey does all of the external facing show notes, even though I occasionally

00:16:37 ◼   ► go in there and add a line or two, Casey doesn't even notice because when he's done

00:16:39 ◼   ► with it, he never looks at them again.

00:16:40 ◼   ► Also, but I would never say that I have anything to do with the external show notes.

00:16:43 ◼   ► It's all Casey.

00:16:44 ◼   ► 100%.

00:16:44 ◼   ► That's why all the bad jokes that are there, that's Casey.

00:16:47 ◼   ► Also very true, rude, but very true.

00:16:50 ◼   ► People sometimes attribute jokes to me and I feel like saying, that's not me.

00:16:54 ◼   ► That's, that's Casey.

00:16:55 ◼   ► All right.

00:16:56 ◼   ► I'm going to stop derailing us or something.

00:16:59 ◼   ► Uh, Mac OS 27.

00:17:00 ◼   ► I'll receive us.

00:17:01 ◼   ► Don't worry.

00:17:01 ◼   ► Yeah, exactly.

00:17:02 ◼   ► Adam DeMassi writes, John said, quote, Golden Gate does sound like a code name rather than

00:17:07 ◼   ► a public name.

00:17:08 ◼   ► Well, he might've been remembering that Golden Gate was actually the code name for Big Sur.

00:17:13 ◼   ► Uh, it was also apparently the code name for eWorld 1.1.

00:17:17 ◼   ► And you even brought receipts to this little endeavor.

00:17:20 ◼   ► Yeah.

00:17:21 ◼   ► Do you guys know what eWorld is?

00:17:22 ◼   ► Yeah.

00:17:23 ◼   ► Wasn't that the like faux AOL?

00:17:25 ◼   ► It's like an online service.

00:17:26 ◼   ► Yeah.

00:17:26 ◼   ► I'm pretty sure I don't know this for a fact, but as having, having used the world when

00:17:30 ◼   ► it was new, I'm pretty sure the company that made the Mac app for AOL, which you probably

00:17:36 ◼   ► aren't familiar with, but anyway, there was a Mac app for AOL.

00:17:38 ◼   ► And I think that company that made that app essentially white labeled it to Apple and said, here, you can have the app.

00:17:43 ◼   ► And then Apple just obviously changed all the strings and stuff and then changed all of the graphics.

00:17:49 ◼   ► The world's whole idea was there was this like a, I don't know how to describe the art style, but sort of a, uh, an impressionistic kind of a cartoony art style of a little village with buildings and stuff.

00:17:58 ◼   ► Back when everyone thought the internet was going to be a, a giant gif of a town with little people in it.

00:18:03 ◼   ► Anyway, that was eWorld.

00:18:05 ◼   ► Uh, and yeah, eWorld 1.1, Golden Gate.

00:18:07 ◼   ► I did not know that.

00:18:08 ◼   ► The Big Sur one, I probably knew and then forgot.

00:18:10 ◼   ► So yeah, there's only so many California names and they get reused.

00:18:13 ◼   ► Do either of you ever use Microsoft Bob?

00:18:15 ◼   ► No, I don't think I've done.

00:18:17 ◼   ► No, I've only seen magazine pictures of it.

00:18:18 ◼   ► I, I used it on, I was, I was traveling to New York when I was a kid once and, and a friend who wasn't Casey, I actually have multiple friends who I would meet in New York.

00:18:28 ◼   ► Um, but I, a friend had it on his computer and I used it for like a couple of days on this trip.

00:18:32 ◼   ► It was wild.

00:18:34 ◼   ► Like, so this was the thing where, you know, again, this was like probably early to mid nineties back when there was still a lot of experimentation in like, what should computer UIs be?

00:18:46 ◼   ► Uh, and, and what should be like the, the metaphor that structures all your applications and documents and everything together.

00:18:53 ◼   ► And Microsoft Bob was an experiment that Microsoft did somewhere in the mid nineties.

00:18:58 ◼   ► is that was like, what if we just arrange things like in a house and you could go to different rooms in the house?

00:19:05 ◼   ► Like very much a, like, you know, like the, the, the early generation of nerds had probably, you know, done a little bit of acid one weekend and came up with this idea in somebody's hot tub.

00:19:16 ◼   ► And it's like, okay, well, that's a cool idea.

00:19:18 ◼   ► It didn't work really well at all.

00:19:20 ◼   ► Um, but I believe that's where Clippy came from.

00:19:23 ◼   ► I think Clippy was like an offshoot of Bob or Clippy escaped to Bob.

00:19:26 ◼   ► But yeah, I mean, but there was like, look, this was back in the day that there, there was still at least some experimentation of like before we had all settled on basically the desktop with applications and files.

00:19:37 ◼   ► Like there, there were other ideas, you know, it wasn't that long after that, that, um, the Palm pilot happened and the Palm pilot also had a totally different structure of like, how should your data and applications be structured?

00:19:50 ◼   ► Which actually was far closer to what we have on iOS today than, than like, you know, PCs were, um, but yeah, e-world was, e-world was a little bit after that, I believe.

00:19:59 ◼   ► But it was, that was back in the day when like nobody was quite sure what the internet would end up being for consumers yet.

00:20:05 ◼   ► And there were things like AOL and CompuServe and everybody wanted their own online service.

00:20:09 ◼   ► And that's, I guess, where e-world came in.

00:20:11 ◼   ► Yeah.

00:20:12 ◼   ► I mean, it was just a hundred percent like the, uh, AOL Mac client, uh, except that they had that graphic where a click on the house and then it would just, I mean, it was very,

00:20:20 ◼   ► really skin deep.

00:20:20 ◼   ► It wasn't like Bob where there was like a, you know, there was very, very few graphics and things fit on a floppy disk, but, uh, yeah, you can find screenshots of it online.

00:20:28 ◼   ► It had a pretty cool art style, but it did not last long.

00:20:30 ◼   ► Thankfully.

00:20:30 ◼   ► Neither did Bob.

00:20:32 ◼   ► I feel, uh, I want to also say that, uh, this, uh, uh, diversion into Microsoft Bob should not be counted towards my time, but I will take full credit for D for the derailment reclaiming my time.

00:20:43 ◼   ► Oh, gracious.

00:20:45 ◼   ► You can now use touch ID for administrator access in Mac OS 27.

00:20:50 ◼   ► John, what does that mean?

00:20:51 ◼   ► I always wondered about this.

00:20:53 ◼   ► I'm like, well, there must be some reason for it.

00:20:55 ◼   ► But if you had like a Mac with touch ID and you needed to like do something, oh, I want to drag a file out of the slash library folder.

00:21:01 ◼   ► And it's like, it was a, in the finder and be like, well, you can't do that because it's not owned by you.

00:21:04 ◼   ► But, uh, if you enter your password, cause you have an admin account, if you enter your password, we'll, we'll, you know, do it for you with administrative privileges or whatever.

00:21:12 ◼   ► And it would always make you enter your password.

00:21:14 ◼   ► Lots of other stuff.

00:21:15 ◼   ► It would be like, oh, you know, you autofill, uh, do touch ID or whatever.

00:21:19 ◼   ► It would ask you for touch ID.

00:21:21 ◼   ► But every time you needed to elevate your privileges, eventually do the equivalent of sudo or sudo or whatever you pronounce that word, uh, it would always throw up a dialogue box.

00:21:30 ◼   ► You had to type in your password.

00:21:32 ◼   ► And I was like, yeah, I don't know.

00:21:33 ◼   ► It's just weird.

00:21:33 ◼   ► What if for whatever reason, 27, you can touch ID to do stuff as admin.

00:21:39 ◼   ► I don't know why they changed their mind.

00:21:41 ◼   ► I'm glad they did.

00:21:42 ◼   ► All right.

00:21:43 ◼   ► You can skip the Siri wait list.

00:21:46 ◼   ► Uh, thanks to front of the show, Steve Trout and Smith.

00:21:48 ◼   ► You can do a defaults, right, which we will put in the show notes and that will let you, I guess, just start using the new Siri.

00:21:54 ◼   ► Is that right?

00:21:55 ◼   ► Yeah.

00:21:55 ◼   ► It doesn't show you the wait list thing.

00:21:57 ◼   ► I was, I was going to say that like, it's fine for us to put this out there because just our super nerdy listeners, which are in the grand scheme of things, few in number, will know about this, but it has since leaked out into the media.

00:22:07 ◼   ► So now everybody knows this.

00:22:08 ◼   ► So hurry up and do it before Apple somehow disables this.

00:22:11 ◼   ► I did it and got immediately into Siri.

00:22:13 ◼   ► Yep.

00:22:13 ◼   ► And it's only for the Mac though, to be clear, it only works on, on Mac OS golden gate.

00:22:16 ◼   ► I mean, where are you going to write, uh, sooner defaults, right command on your phone?

00:22:20 ◼   ► That's the, yeah, I was, I would love to, to skip it on my test phone for iOS 27, but, uh, that there seems to be no way to skip that yet, except I guess be Joanna Stern.

00:22:30 ◼   ► And speaking of Steve Trout and Smith, uh, he also pointed out that there is a change in Apple's documentation.

00:22:36 ◼   ► The key UI design requires compatibility.

00:22:39 ◼   ► Uh, well, let me just read what it says in the documentation.

00:22:41 ◼   ► The system ignores this key when you build for iOS 27 or later iPad OS 20.

00:22:46 ◼   ► Well, basically all the 27 or later, uh, OS's.

00:22:48 ◼   ► So this is the thing that said, don't use liquid glass for my app, please.

00:22:52 ◼   ► And Apple saying tough noogs, you're going to have to use it.

00:22:55 ◼   ► If you compile against iOS, any of the OS is 27.

00:22:58 ◼   ► I think they might've said that last year too.

00:22:59 ◼   ► Like, Oh, we have this, this key.

00:23:01 ◼   ► You can put it in your app and you don't have to upgrade to liquid glass.

00:23:03 ◼   ► We'll just run you using the, basically the old metrics, all the old controls and the old everything.

00:23:08 ◼   ► Uh, that time is over update for 27.

00:23:11 ◼   ► No choice.

00:23:12 ◼   ► Yeah, we, I think we were adequately warned and that kind of key usually only does last a year or two.

00:23:17 ◼   ► So that makes sense.

00:23:20 ◼   ► It's time to get on board.

00:23:21 ◼   ► That being said, like all of the problems of liquid glass from, from 26, like we, like if you were kicking the can down the road and you didn't want to deal with them with your app, you still have to deal with them because even if golden gate fixes some of your biggest, uh, you know, nitpicks or issues, which honestly, in that kind of context,

00:23:42 ◼   ► of like app developers maintaining things, it's not really going to help that much.

00:23:46 ◼   ► Um, it's just nicer, you know, in certain ways, but all of the things about like, you know, having different metrics for controls than iOS 18 and different behaviors and things, you still have to deal with that as long as you support iOS 18 or iOS 26.

00:24:01 ◼   ► And for most apps, that's going to be a little bit like you're probably still dealing with supporting all these things for at least another year.

00:24:09 ◼   ► So, uh, like, like for me with overcast, I'm thinking this is probably a good time.

00:24:14 ◼   ► Like this, like once I launched my 27 version this fall, I can probably drop support for iOS 18 then, but I that's even that's pretty aggressive given where the numbers actually are for most apps these days.

00:24:26 ◼   ► So, uh, I understand why people still want this key to work, um, because then they can keep shipping one UI for their app instead of now kind of two and a half is what they'll have to do.

00:24:39 ◼   ► But, uh, it's, it's going to be a bumpy time.

00:24:41 ◼   ► Like the, the support for 26 and Tahoe is going to be kind of a thorn in app developer sides for probably at least another two years.

00:24:50 ◼   ► Yeah.

00:24:50 ◼   ► I hope you spent the last year working on that issue and not saying, oh, they gave me a key.

00:24:54 ◼   ► I don't have to worry about it because if you did that, you're scrambling now.

00:24:57 ◼   ► Yep.

00:24:57 ◼   ► All right.

00:24:58 ◼   ► During onboarding, and I think this is coming from one, John Syracuse, during onboarding for Golden Gate, the liquid glass slider, where you decide how transparent you want it to be.

00:25:07 ◼   ► Well, that's part of the ongoing of the onboarding process, which is pretty interesting.

00:25:11 ◼   ► That is what that means is, uh, you know, I was just saying, oh, everyone will just leave it on the default.

00:25:16 ◼   ► But now I wonder, I mean, maybe most people probably still leave it on the default because people just hit continue, continue, continue.

00:25:21 ◼   ► I don't care.

00:25:21 ◼   ► Continue later, continue.

00:25:22 ◼   ► They just want to get through it.

00:25:23 ◼   ► But yeah, it's part of, it's part of like the, the onboarding after you install or update to the OS.

00:25:29 ◼   ► So everyone's going to at least see that slider.

00:25:30 ◼   ► And I'm sure Apple will be monitoring where those numbers, I say, I'm sure, but honestly, I don't know what kind of metrics Apple gathers about this stuff.

00:25:38 ◼   ► Apple tends not to gather much info, but we do know that they have some numbers.

00:25:42 ◼   ► So it's not really talked about it.

00:25:44 ◼   ► Like what kind of anonymous usage info is actually gathered from like Mac OS and iOS.

00:25:49 ◼   ► Um, yeah, I hope Apple is looking at where the slider goes.

00:25:53 ◼   ► I bet people will either not touch it or put it all the way to the right.

00:25:56 ◼   ► I put it all the way to the right.

00:25:57 ◼   ► And then some people put all the way to the left because they're young and they think it's cool.

00:26:00 ◼   ► I still, this, this slider still paffles me.

00:26:06 ◼   ► Like it still just feels like we didn't feel like deciding.

00:26:10 ◼   ► So we're going to give you this zero to one floating point setting that changes one thing about liquid glass that honestly isn't even that much.

00:26:20 ◼   ► First of all, it isn't that impactful of a change.

00:26:21 ◼   ► And second of all, the range between zero and one of like how they look is not that different.

00:26:29 ◼   ► But most importantly, going all the way to the right does not make it opaque.

00:26:32 ◼   ► There's reduced transparency for that if you want it.

00:26:34 ◼   ► But just to be clear, the slider makes it less transparent.

00:26:38 ◼   ► And all the way to the left is not fully clear either.

00:26:40 ◼   ► So like whatever you want, this is kind of a like weird milk toast middle of the road kind of like, well, we can give you a little bit of control.

00:26:49 ◼   ► But it feels like a very precise control and it just isn't.

00:26:53 ◼   ► And by the way, they give you like a picture of Apple Park with some crap floating on top of it, you know, liquid glass style.

00:26:58 ◼   ► Be aware that that picture scrolls.

00:27:02 ◼   ► So you can scroll the picture just to see how does it look with an image scrolling underneath the controls.

00:27:07 ◼   ► Oh, interesting.

00:27:07 ◼   ► Before you move on from that, as Marco pointed out, the underlying setting, which is called NS glass tint amount, does go from zero to one.

00:27:15 ◼   ► If you set it to a value higher than one, it is ignored.

00:27:19 ◼   ► Unfortunately, the first thing I did, I'm like, oh, the slider goes from zero to one, 15.

00:27:23 ◼   ► What does 15 look like?

00:27:24 ◼   ► 15 looks like one.

00:27:25 ◼   ► We are sponsored this week by Lisa.

00:27:28 ◼   ► Lisa is tunnel visioned on making great mattresses and furniture that supports those mattresses.

00:27:34 ◼   ► And I really dig that.

00:27:35 ◼   ► Lisa mattresses, they have all sorts of different options, including a pet bed, which isn't exactly a mattress, but it's close enough.

00:27:43 ◼   ► And I think that's adorable.

00:27:44 ◼   ► You should see the little image on their website.

00:27:46 ◼   ► You should go check it out.

00:27:47 ◼   ► We'll tell you the URL at the end.

00:27:49 ◼   ► So you can see this adorable little puppy sitting on one of their pet beds.

00:27:52 ◼   ► But they do mattresses for all sorts of different sizes and shapes of people, including children, including their plus mattress, which has a little bit extra support and cushioning right where it counts.

00:28:02 ◼   ► And that's really cool in my book.

00:28:04 ◼   ► They also have all sorts of different materials that you can choose from to make sure that you're doing what you think is best for your body.

00:28:10 ◼   ► They offered to send a mattress to me so I could give it a shot and make sure that it's worth talking to you folks about.

00:28:17 ◼   ► And so we needed a new mattress for Michaela, actually.

00:28:20 ◼   ► And we ordered a natural hybrid mattress for Michaela.

00:28:23 ◼   ► But I slept on it for a couple of nights before Michaela did, so I could make sure I'd done my due diligence.

00:28:27 ◼   ► And let me tell you, I am a mattress snob.

00:28:30 ◼   ► Most people I know are.

00:28:31 ◼   ► This was a great mattress and a great night's sleep.

00:28:34 ◼   ► And what's great about Lisa is all their mattresses have a lot of these same qualities.

00:28:38 ◼   ► They have a lineup of beautifully crafted mattresses tailored to how you sleep.

00:28:42 ◼   ► Each mattress is designed with specific sleep positions and feel preferences in mind.

00:28:46 ◼   ► They have a little quiz you can go and do, if I'm not mistaken, and you can check out what is the best mattress for you.

00:28:53 ◼   ► It's really, really cool.

00:28:54 ◼   ► So, if you're interested in this, go to Lisa.com for 25% off select mattresses, plus get an extra $50 off if you use the promo code ATP, which is exclusive for our listeners.

00:29:05 ◼   ► That's L-E-E-S-A.com, promo code ATP, for 25% off select mattresses and an extra $50 off.

00:29:13 ◼   ► Support our show and let them know we sent you after checkout.

00:29:16 ◼   ► Again, Lisa.com, promo code ATP.

00:29:19 ◼   ► Thank you to Lisa for sponsoring the show.

00:29:24 ◼   ► All right, this might get me in a lot of trouble, but there was a lot of consternation with regard to the corner radius of macOS Windows.

00:29:35 ◼   ► And, John, would you like to just briefly recap why perhaps you, or if not you, others were upset about this in macOS 26, please?

00:29:44 ◼   ► Yeah, they made the corner radius very large.

00:29:47 ◼   ► It means the corners were very, very rounded, which means there are a larger number of pixels that you don't get to see in the corners of a window.

00:29:54 ◼   ► And this is especially important in Tahoe because the way the UI works is that if, like, if you're looking at a picture, like an image or something, the image goes edge to edge.

00:30:04 ◼   ► So you're losing all four corners of that image if that image is in a window because all four corners are curved.

00:30:11 ◼   ► As opposed to, like, go back...

00:30:13 ◼   ► But, John, it respects your content.

00:30:14 ◼   ► Yeah.

00:30:15 ◼   ► Go back, like, several versions when there was an opaque tile bar.

00:30:18 ◼   ► The title bar had rounded corners.

00:30:20 ◼   ► But when you viewed an image in a window, the image started where the title bar ended.

00:30:26 ◼   ► Like, under the title bar is where the image started.

00:30:28 ◼   ► And those were sharp corners because the title bar ended.

00:30:30 ◼   ► And then there was just a right angle.

00:30:32 ◼   ► And you could see all the pixels in the image.

00:30:34 ◼   ► And then when you went to the bottom of the window for many, many, many years in macOS 10 slash macOS,

00:30:39 ◼   ► the bottom corners of the window were sharp as well.

00:30:41 ◼   ► So if you put an image in a window, you could see all the pixels in the image.

00:30:44 ◼   ► And Apple has slowly eaten away at that.

00:30:46 ◼   ► And 26 was like, we're taking the whole corners.

00:30:49 ◼   ► Like, there could be some...

00:30:50 ◼   ► There could be a big thing in there.

00:30:51 ◼   ► There could be a...

00:30:51 ◼   ► There could be, like, a little signature that you wouldn't even see in the corner of, like, a picture if it's a painting.

00:30:55 ◼   ► They were just chopping it all off.

00:30:56 ◼   ► And some people didn't like how it looked.

00:30:58 ◼   ► Some people did like how it looked.

00:31:00 ◼   ► But the bottom line is the job of the window is to show the content.

00:31:03 ◼   ► And it annoys me when they decide you don't need to see the corners.

00:31:08 ◼   ► And they decided that because of concentricity, which is they had all these little, you know,

00:31:11 ◼   ► their controls had semicircular end caps on them.

00:31:14 ◼   ► And they could make the corners match the radius of the semicircular end caps,

00:31:17 ◼   ► such that if you put a central point where the center of the corner is and you traced out an arc,

00:31:21 ◼   ► the arc on the window edge would exactly match the arc on the controls.

00:31:26 ◼   ► That's concentricity, where you'd see concentric circles drawn from a central point.

00:31:30 ◼   ► And then in macOS 27, as we discussed last episode,

00:31:34 ◼   ► decreased the corner radius of the windows so you can see more of your stuff.

00:31:38 ◼   ► And there was much rejoicing.

00:31:39 ◼   ► And many people who were there live have said since we recorded last week that it was cheering

00:31:43 ◼   ► from the live audience for the fact that they fixed the window radius.

00:31:47 ◼   ► So it's not just me.

00:31:48 ◼   ► But as I pointed out last episode,

00:31:49 ◼   ► they, you know, I'm, I'm, I endorse this change.

00:31:53 ◼   ► They made a smaller corner radius on all the windows,

00:31:56 ◼   ► but it no longer matches the corner radius on the floating controls.

00:32:01 ◼   ► And so now it is the exact opposite of concentricity.

00:32:04 ◼   ► You have two very different corner radii, radiuses.

00:32:07 ◼   ► And yeah, it's hard.

00:32:09 ◼   ► Anyway, all this is difficult for me to explain graphically.

00:32:12 ◼   ► So I'm glad that Sammy Sito posted some images to explain this.

00:32:17 ◼   ► And we'll put this link in the show notes.

00:32:19 ◼   ► When you look at this image, you'll understand what I'm talking about.

00:32:22 ◼   ► 26, you're like, oh, everything looks nice and everything matches in 27 doesn't match anymore.

00:32:27 ◼   ► To be clear, I totally endorse the change in 27,

00:32:29 ◼   ► but it is fighting against the other parts.

00:32:32 ◼   ► They didn't change in liquid glass, which are the little capsule shape buttons.

00:32:35 ◼   ► Yeah.

00:32:36 ◼   ► And I think this is one of those cases where the,

00:32:39 ◼   ► I can see why they went for the Tahoe design,

00:32:43 ◼   ► because from a like visual isolated context, it does look better.

00:32:48 ◼   ► The problem is it doesn't work better in practice with computer UIs.

00:32:52 ◼   ► And that is, I think, where, where Apple has gone a little bit too far in, you know,

00:33:00 ◼   ► when, when their designers get a little bit too, too pure and theoretical about things.

00:33:05 ◼   ► This is kind of their failure mode.

00:33:07 ◼   ► And this has happened numerous times over their history with numerous designers in charge too.

00:33:12 ◼   ► And, you know, fortunately they tend to eventually correct those things.

00:33:16 ◼   ► And in this case, it took a year, but they corrected it.

00:33:19 ◼   ► So I'm glad to see that.

00:33:19 ◼   ► Like, there is no question in my mind, like, yeah, when you're looking at a screenshot,

00:33:23 ◼   ► the Mac OS 26 version does look better.

00:33:26 ◼   ► But unfortunately, design has to interact with the real world.

00:33:30 ◼   ► And sometimes that requires compromises on the pure ideal aesthetics.

00:33:36 ◼   ► And you have to actually run into, you know, what people actually need in reality.

00:33:41 ◼   ► Yeah, again, you have to ask, what is the job of a window?

00:33:43 ◼   ► And this, the design they have now is awkward and bad because also the little floaty buttons

00:33:48 ◼   ► and toolbars were also designed poorly with, you know, if, if they, who designed from the

00:33:53 ◼   ► start and you would say, okay, the job of the window is to show the content and let's let

00:33:57 ◼   ► everything fall from that.

00:33:58 ◼   ► And what they've done is taken a design that totally didn't do that job and fixed one part

00:34:02 ◼   ► of it without fixing the other.

00:34:03 ◼   ► So of course they don't match.

00:34:04 ◼   ► It's still better than it was.

00:34:05 ◼   ► Again, I still endorse the change, but it just highlights, hey, we fixed one thing, but it

00:34:09 ◼   ► looks like maybe this entire idea was bad because now we fixed that one thing and we see some

00:34:13 ◼   ► other stuff that also looks bad with it and needs to be fixed.

00:34:16 ◼   ► It's going to be a while before they rewind, uh, unwind all this stuff, but a baby steps

00:34:21 ◼   ► one thing at a time.

00:34:21 ◼   ► I got to tell you, based only on these screenshots, I agree with the slug of this post, which is

00:34:27 ◼   ► y'all better shut up.

00:34:28 ◼   ► Seriously.

00:34:29 ◼   ► Is this what you really want to see window corners?

00:34:31 ◼   ► And I, I, based on the screenshot alone, it looks fricking terrible.

00:34:36 ◼   ► And I, in Mac OS 27, it looks awful.

00:34:39 ◼   ► Yeah.

00:34:39 ◼   ► Cause, cause you, cause it's a mismatch.

00:34:40 ◼   ► It was, they, they, they fixed the corners and didn't fix the other parts that are next to

00:34:44 ◼   ► the corners.

00:34:44 ◼   ► You know how you could fix the other parts that are next to the corners.

00:34:46 ◼   ► You can make those radii also match the, you know, like there's a way to do it, but they

00:34:50 ◼   ► didn't do it.

00:34:50 ◼   ► Right.

00:34:50 ◼   ► Like, like in this case, like the, like it's a little bit easier on the search box compared

00:34:54 ◼   ► to the toolbar, but like, but yeah, like the search box, the reason why this is a mismatch

00:34:59 ◼   ► is that they made the search box and the toolbar pills instead of round rects.

00:35:05 ◼   ► So, you know, basically like what we used to call ovals in kindergarten, it's like they

00:35:10 ◼   ► are, this is a text field that is a perfect oval around the left and right side.

00:35:14 ◼   ► It's not an oval.

00:35:15 ◼   ► Don't get the geometry to people on you.

00:35:16 ◼   ► Well, okay.

00:35:17 ◼   ► Whatever.

00:35:18 ◼   ► But it's close.

00:35:19 ◼   ► It's a capsule shape.

00:35:20 ◼   ► It has semicircular end caps on a rectangle.

00:35:22 ◼   ► Right.

00:35:22 ◼   ► And the problem is as the, like, as the height of that increases, that means that the radius

00:35:29 ◼   ► has to increase too.

00:35:31 ◼   ► So it does significantly constrain your design in a way that like, well, now you have to

00:35:36 ◼   ► have very, very large, uh, corner radii on the window to match all of these pill shapes

00:35:42 ◼   ► or capsule shapes of the controls.

00:35:44 ◼   ► In previous versions of Mac OS, those weren't exact pill shapes.

00:35:49 ◼   ► They were round rects so that you could, so yes, they had rounded corners, but the entire

00:35:55 ◼   ► left and right side didn't have to be a continuous curve.

00:35:58 ◼   ► So that gives you way more flexibility.

00:36:00 ◼   ► And search fields were pill shapes at one point, but the way they made that work is they just

00:36:04 ◼   ► put bigger margins around them.

00:36:06 ◼   ► The corners of the window were practically square and the search field was pill shaped back in

00:36:10 ◼   ► the day.

00:36:10 ◼   ► They just moved it farther from the edges.

00:36:12 ◼   ► So it didn't look weird.

00:36:13 ◼   ► I mean, just again, it's about holistic design.

00:36:16 ◼   ► Design is like everything together.

00:36:18 ◼   ► Does it work well?

00:36:19 ◼   ► Does it look good?

00:36:20 ◼   ► Does it look harmonious?

00:36:21 ◼   ► And just 26 was like, well, we think it looks good and it's visually harmonious, but nothing

00:36:27 ◼   ► works.

00:36:27 ◼   ► And so they're just trying to fix it without changing everything.

00:36:30 ◼   ► And you get 27.

00:36:31 ◼   ► I think what, what has tentatively sold me on 27 being the improvement is what Marco said

00:36:36 ◼   ► a couple of minutes ago that, yeah, this looks worse in, in this microcosm or in this,

00:36:41 ◼   ► like, you know, zoomed in view, but it actually affords you a much better view of the content

00:36:48 ◼   ► behind all this.

00:36:49 ◼   ► So maybe they really are respecting our content.

00:36:51 ◼   ► Who'd have thunk it?

00:36:52 ◼   ► All right.

00:36:53 ◼   ► 9to5Mac has noted that the Mickey Mouse hands are back in macOS 27 Golden Gate.

00:36:58 ◼   ► And they also indicated or said the world is healing.

00:37:02 ◼   ► Do you know about the Mickey Mouse hands?

00:37:04 ◼   ► Was this news to you?

00:37:05 ◼   ► No, we had talked about that.

00:37:06 ◼   ► I think at some point, semi-recently, I guess when 26 came out.

00:37:09 ◼   ► I, I, I still hate the hand icons in 26 and 27.

00:37:12 ◼   ► They didn't change the shape of the hand icons much.

00:37:15 ◼   ► I think maybe the, it looks like the, the pointing finger is tilted a little bit, but

00:37:18 ◼   ► uh, but yeah, what we're talking about is, uh, Mickey Mouse, the character wears white gloves

00:37:23 ◼   ► with like three little creases on the back of the hand.

00:37:26 ◼   ► And for, I don't know how long for ages and eight, maybe since classic macOS, I think

00:37:30 ◼   ► so.

00:37:30 ◼   ► Uh, the various hand cursors, the pointing hand that you click on links with in a web browser,

00:37:34 ◼   ► uh, the big hand that when you hold on space bar in Photoshop, you can move the canvas around

00:37:39 ◼   ► and the grabby hand where it's like grabbing something.

00:37:42 ◼   ► All those hands look like a white glove cause they were white black outline with a white

00:37:46 ◼   ► cursor with three little, uh, Mickey Mouse creases on the back of the glove.

00:37:51 ◼   ► And in 26, they took off the creases because I don't know that they didn't want it to be

00:37:56 ◼   ► whimsical anymore.

00:37:57 ◼   ► And 27, they brought them back, which to me seems petty, but I endorse it.

00:38:03 ◼   ► That's incredible.

00:38:04 ◼   ► Uh, all right.

00:38:05 ◼   ► In macOS 27, there are a bunch of new icons.

00:38:10 ◼   ► Uh, and I don't know, they're, they're different.

00:38:13 ◼   ► I don't know what else to say, but the, the finder one, I don't know what they're doing here.

00:38:18 ◼   ► The basic apple guy says, honey, the finder icon sucks again.

00:38:22 ◼   ► And I disagree.

00:38:22 ◼   ► I know what he's getting at.

00:38:24 ◼   ► And, uh, we'll get more from the basic apple guy in a little bit.

00:38:27 ◼   ► Um, but if you go to the link we'll have in the notes, you can see the, the Tahoe and

00:38:31 ◼   ► golden eight finder icons.

00:38:32 ◼   ► And the main sin of the golden eight finder icon is they made the blue too light again.

00:38:36 ◼   ► I don't know who's obsessed with making a light blue finder, but just back off, man.

00:38:39 ◼   ► Like it's not, it's gotta be like a deeper blue.

00:38:41 ◼   ► That's the way it's supposed to be.

00:38:42 ◼   ► But they also changed the shape of the profile face.

00:38:45 ◼   ► Like the, the finder icon, which used to be the macOS icon is like, um, uh, uh, face that

00:38:51 ◼   ► is, uh, you know, the two eyes and a smiley face that looks like a face head on, but it

00:38:55 ◼   ► is a blue on the left and white on the right.

00:38:57 ◼   ► And the right hand part of it also looks like a face in profile.

00:39:01 ◼   ► Uh, so it's kind of like a two face thing.

00:39:03 ◼   ► Some people call it two face.

00:39:04 ◼   ► Anyway, uh, in Tahoe, the profile face had a completely vertical line under the nose of

00:39:12 ◼   ► the face.

00:39:12 ◼   ► And now it is more curved.

00:39:14 ◼   ► The original version of this was more curved.

00:39:16 ◼   ► So I think the golden gate shape is actually an improvement.

00:39:19 ◼   ► The color is just too light.

00:39:21 ◼   ► So maybe I should follow feedback on that.

00:39:23 ◼   ► Just darken up the color.

00:39:24 ◼   ► You're, you're on the right path.

00:39:25 ◼   ► Just darken up the color and then maybe extend the white part to the edges of the squircle.

00:39:29 ◼   ► And then maybe it just gives back the old icon.

00:39:31 ◼   ► Yeah.

00:39:32 ◼   ► I think at that point you're at the old icon, which honestly is better.

00:39:34 ◼   ► Yeah.

00:39:34 ◼   ► Well, not really.

00:39:35 ◼   ► Cause this, there's been so many variations.

00:39:37 ◼   ► If you look at the original, like classic Mac OS one from Mac OS 7.6 or whenever it first

00:39:42 ◼   ► appeared, it wasn't the finder.

00:39:43 ◼   ► Rock on it was the booting up Mac OS 7.6, like splash screen face.

00:39:47 ◼   ► It had a vertical line that extended past the bounds of the thing.

00:39:51 ◼   ► It has looked very different over the years.

00:39:53 ◼   ► So anyway, I think golden gates design is better than Tahoe's.

00:39:56 ◼   ► I just missed the color.

00:39:56 ◼   ► I think we just need Steven Hackett to get upset about this and then it'll get changed.

00:40:01 ◼   ► That's how it worked last year.

00:40:02 ◼   ► Right.

00:40:02 ◼   ► Something like that.

00:40:03 ◼   ► All right.

00:40:04 ◼   ► And then we have a bunch of other ones by basic Apple guy and they are side by side in

00:40:09 ◼   ► their post.

00:40:10 ◼   ► And I don't know.

00:40:10 ◼   ► Do you want to go through these?

00:40:11 ◼   ► Do you just have broad comments about?

00:40:13 ◼   ► Yeah, well, we'll go.

00:40:14 ◼   ► I mean, again, follow the link to basically Apple guys post about this.

00:40:16 ◼   ► Cause he posted a ton of images to say show Tahoe versus golden gate.

00:40:20 ◼   ► And I just want to make the point that we talked about this back when Tahoe came out

00:40:24 ◼   ► with, um, when, you know, in WWC 2025 with the advent of Tahoe and icon composer.

00:40:29 ◼   ► And just to recap, they introduced a new format for icons, which is dot icon files.

00:40:33 ◼   ► It's amazing.

00:40:33 ◼   ► No one had ever used that extension before, but Apple got it dot icon.

00:40:36 ◼   ► And it's basically a bunch of resources in a little package.

00:40:42 ◼   ► And instructions for compositing them.

00:40:44 ◼   ► And the resources can be bitmaps or vector images.

00:40:48 ◼   ► And it, what this means is if you go into an application package on Tahoe and you dig around in

00:40:55 ◼   ► there, you're like, where's the icon?

00:40:56 ◼   ► You probably won't find a bit mapped image of the icon unless it has one of those for legacy

00:41:01 ◼   ► reasons or fallback reasons.

00:41:03 ◼   ► What you'll find are these icon assets shoved into an asset catalog that have the ingredients

00:41:07 ◼   ► of the icon and they are composited in real time by macOS to make the icon.

00:41:12 ◼   ► And part of the reason they do that is because Tahoe introduced this idea of, uh, on macOS

00:41:17 ◼   ► anyway, and I guess on iOS as well, of icon themes where you can have a light mode, dark mode,

00:41:22 ◼   ► the clear icons, all that is doing is changing the formula for how those pieces in the dot icon

00:41:29 ◼   ► files are composited to make the final icon.

00:41:32 ◼   ► So ideally in Apple's world, you just have a bunch of vector shapes and a bunch of instructions

00:41:37 ◼   ► for compositing them.

00:41:38 ◼   ► And then we can make that icon look good in light mode, in dark mode, in clear mode and

00:41:44 ◼   ► in, you know, whatever.

00:41:45 ◼   ► Um, another thing that you can do by moving to the icon format to this recipe and ingredients

00:41:51 ◼   ► composited together is that you can change essentially all the icons in every single one of your OS's

00:41:57 ◼   ► without having to redraw your icons.

00:41:59 ◼   ► So that's what we're looking at here.

00:42:00 ◼   ► As far as I can tell, pretty much all these icons and basic Apple guys posts are the same

00:42:06 ◼   ► icons as in the same contents of the dot icon files that were assembled into the assets

00:42:11 ◼   ► dot car file and whatever this, like the resources are the same.

00:42:14 ◼   ► They didn't redraw the icons.

00:42:15 ◼   ► All they're doing is changing the recipe for how they're composited.

00:42:19 ◼   ► And you would think how big a difference could that possibly make?

00:42:22 ◼   ► Obviously the finder icon has changed.

00:42:23 ◼   ► We just got done saying that the shapes are different, but these icons we're looking at

00:42:27 ◼   ► our notes and most of the ones in this blog post that we'll link, I think they're basically

00:42:30 ◼   ► the same icon, but changing the compositing makes the world of difference.

00:42:34 ◼   ► Marco mentioned it before when he said that, um, when he upgraded 27, then when look back

00:42:38 ◼   ► at 26 and everything looked blurry, these are way zoomed in, in our show notes.

00:42:42 ◼   ► So I'd encourage you to like move back from your screen or like shrink them down or whatever.

00:42:45 ◼   ► But the smaller these icons get the, you know, the more they are icon size and not blown up

00:42:50 ◼   ► to this giant size, the more you can see how blurry Tahoe's compositing was.

00:42:54 ◼   ► Like everything is in this weird haze and all the edges are blurred and take that same resource

00:42:59 ◼   ► and rendered in the new way.

00:43:00 ◼   ► And there's two things about the new way of rendering one crisper edges and two, they added

00:43:05 ◼   ► a new refraction, uh, feature to the recipe where each individual layer can say, should

00:43:12 ◼   ► this layer refract what's behind it?

00:43:14 ◼   ► And by refracting, they just mean like make it seem as if the edges of the thing are made

00:43:19 ◼   ► of rounded glass.

00:43:20 ◼   ► So that when you see the things through it, it kind of bends the light a little bit.

00:43:23 ◼   ► Anyway, I think for the most part, uh, the golden gate versions of every single one of

00:43:28 ◼   ► these icons is an improvement because things that were blurry and indistinct and low contrast

00:43:33 ◼   ► and Tahoe become sharp and higher contrast.

00:43:36 ◼   ► And the, the few cases where they added refraction, like the free form icon, I think look great,

00:43:42 ◼   ► especially at large sizes and small sizes kind of lost, but what do you think?

00:43:45 ◼   ► Do you think, uh, this is an upgrade or a downgrade?

00:43:48 ◼   ► Oh my God, I think this looks like an eye test.

00:43:50 ◼   ► When you look at the Tahoe icons and, and then you look at the golden gate icons.

00:43:55 ◼   ► One or two?

00:43:56 ◼   ► Yeah, exactly.

00:43:57 ◼   ► It, it's the Tahoe icons look like I don't have my reading glasses on and I need to put

00:44:01 ◼   ► them like, and then you put the, and then you look over at golden gate, you're like, ah,

00:44:04 ◼   ► crispness.

00:44:05 ◼   ► Like maybe look, Alan Dye is seemed like he's about in his forties.

00:44:08 ◼   ► Maybe he just needs reading glasses and he doesn't know that yet.

00:44:12 ◼   ► It looks like when you, when you see Tahoe, it, it looks like eye strain.

00:44:17 ◼   ► This is part of what I like.

00:44:18 ◼   ► This is part of why I am so anti blurry text ever appearing in a UI because when you are

00:44:25 ◼   ► of middle age and you need reading glasses, when you see blurry text, your eyes, like you

00:44:32 ◼   ► instinctively think I have to squint or, you know, look through the bottom of my glasses

00:44:37 ◼   ► or whatever it is, like whatever your accommodation is for seeing things more sharply.

00:44:42 ◼   ► You know, you, you kind of instinctively think I'm the problem and you, and you do whatever

00:44:47 ◼   ► that accommodation is instinctively and then it doesn't get fixed.

00:44:52 ◼   ► And that's what eye strain is.

00:44:55 ◼   ► And so when you look at blurry text in a UI, it reminds you of eye strain and then also

00:45:02 ◼   ► can cause eye strain, which is part of the many, it's one of the many reasons why that's a bad

00:45:09 ◼   ► idea in a UI design.

00:45:10 ◼   ► And so to, to kind of do that across the whole system in a broad diffuse way to have everything

00:45:18 ◼   ► be a little bit soft and blurry all the time is not a good design.

00:45:24 ◼   ► So I am again, very thankful when I see these icons, like, ah, crispness.

00:45:29 ◼   ► And you know what, when I have to like tweak my overcast icon to, to work well in 26 and 27,

00:45:35 ◼   ► that's going to be a little bit of a pain in the butt.

00:45:38 ◼   ► In this case, I don't care worth it.

00:45:40 ◼   ► I am so happy to see the return of sharpness and contrast.

00:45:46 ◼   ► And that, that's the overall message here.

00:45:49 ◼   ► Some, some of these designs probably need a bit more tweaking to look as, as good as

00:45:53 ◼   ► they can, but overall it is way sharper and higher contrast.

00:45:58 ◼   ► But you don't have to do anything to your overcast icon.

00:46:00 ◼   ► That's, that's the point.

00:46:01 ◼   ► Like the new version of icon composer will show you, if you just take, open up your icon file

00:46:06 ◼   ► that you've been using in Tahoe, it will show you in the app.

00:46:09 ◼   ► Here's how this icon looked in Tahoe.

00:46:11 ◼   ► And here's how it's going to look, uh, in, in golden gate, no changes to the icon.

00:46:16 ◼   ► Again, this is just changing how the software composites your stuff.

00:46:19 ◼   ► Now you can change, I think the one thing they added in 27 is you can flip the little

00:46:24 ◼   ► bit for each layer that says refraction.

00:46:26 ◼   ► Yes, no.

00:46:26 ◼   ► And you can change the amount of refraction and angle and all sorts of other crap like

00:46:30 ◼   ► that.

00:46:30 ◼   ► But I think that is the only new feature.

00:46:32 ◼   ► But like I said, Apple didn't update for the most part, didn't update these icons.

00:46:36 ◼   ► The exact same icon will look different in 26 than it does in 27, which is, you know,

00:46:41 ◼   ► it's wild, but it's like a fallout of them deciding that their icons are going to be

00:46:46 ◼   ► ingredients and recipes instead of baked in bitmaps.

00:46:49 ◼   ► Now there's nothing stopping you from making your icon an icon composer and just putting

00:46:53 ◼   ► in one layer that's a hundred percent opaque bitmap.

00:46:55 ◼   ► Like you can still do that.

00:46:56 ◼   ► There's no one stopping you from making a bitmap icon.

00:46:58 ◼   ► Lots of Apple icons incorporate bitmaps in them.

00:47:01 ◼   ► But it seems like if you look at all of Apple's icons, they prefer you just use a bunch of

00:47:06 ◼   ► vector shapes and that gives Apple the freedom to essentially change the look of every single

00:47:11 ◼   ► icon without asking developers or their own developers to change anything about their icons.

00:47:16 ◼   ► They just change the render.

00:47:17 ◼   ► Now it's somewhat limited.

00:47:18 ◼   ► They're not going to change them in radical ways, but this is the thing to keep an eye on.

00:47:22 ◼   ► It's a bold new regime for icons going from like photorealistic bitmaps in 10.0 to, you

00:47:30 ◼   ► know, a bunch of ingredients and recipes and already from 26 to 27.

00:47:34 ◼   ► 26, they rolled out.

00:47:35 ◼   ► 27, they're changing the look.

00:47:36 ◼   ► So I wonder if next year we're going to have a thing, an icon composer that says, hey, the

00:47:40 ◼   ► same icon with no changes.

00:47:41 ◼   ► Here's how it looks in 26.

00:47:43 ◼   ► Here's how it looks in 27.

00:47:44 ◼   ► Here's how it looks in 28.

00:47:45 ◼   ► And don't remember, remember, you also have to check what is light mode, dark mode, clear

00:47:49 ◼   ► mode.

00:47:49 ◼   ► Is that all of them or is there one more?

00:47:50 ◼   ► I'm forgetting.

00:47:51 ◼   ► Yeah.

00:47:51 ◼   ► It's light, dark and like kind of, yeah, uncolored.

00:47:53 ◼   ► Clear.

00:47:54 ◼   ► Yeah.

00:47:54 ◼   ► So fascinating.

00:47:56 ◼   ► I think some of the icons are downgrades.

00:47:57 ◼   ► I think the little robot guy looks a little bit worse, but he's a pale shadow of his former

00:48:01 ◼   ► self anyway.

00:48:02 ◼   ► Well, that's the thing.

00:48:03 ◼   ► Like when I saw my overcast icon for the first time on 27, I was like, oh, that's a little

00:48:08 ◼   ► it's a little bit too sharp in certain ways because I had designed it to try to fight against

00:48:14 ◼   ► all that softness in 26.

00:48:16 ◼   ► So like there are certain, you know, colors or layer effects or things that like I might

00:48:20 ◼   ► dial back slightly for the 27 version so that it looks more normal.

00:48:24 ◼   ► You know, it's kind of like when you see like when you are when you have like stage makeup

00:48:29 ◼   ► on and then you go into the regular world and you look ridiculous, you know, like in different

00:48:33 ◼   ► contexts, you might actually make different decisions depending on how it is being rendered

00:48:38 ◼   ► and where.

00:48:39 ◼   ► So I think you can make tweaks to the recipe on a per OS basis.

00:48:44 ◼   ► But I'm not entirely sure.

00:48:45 ◼   ► Yeah, I'll have to see.

00:48:46 ◼   ► I'm pretty sure you can't include multiple icons for different OSs.

00:48:51 ◼   ► I did that thing that I talked about in the past episode where I bent over backwards to

00:48:54 ◼   ► use my little bitmap icons on pre-26 and I'm going to still keep doing that if I can.

00:48:58 ◼   ► I'm not.

00:48:59 ◼   ► I know Icon Composer lets you tweak the icons for different modes.

00:49:03 ◼   ► Like, oh, do you want to make a tweak that's in dark mode only?

00:49:05 ◼   ► Like in dark mode, you want to turn off this layer effect, but in light mode, you want

00:49:08 ◼   ► to turn it on.

00:49:08 ◼   ► But I don't know if it lets you do that on a per OS basis.

00:49:11 ◼   ► I know it previews your icon and says, here's how it looks in 26 and 27.

00:49:14 ◼   ► And I think there might be one or two other things they added besides refraction.

00:49:18 ◼   ► I think they might have let you do like inside-outside glint effects on stuff.

00:49:22 ◼   ► Anyway, brave new world of icons.

00:49:27 ◼   ► All right.

00:49:27 ◼   ► There's been a lot of chatter, and understandably so, about how Apple is basically quarterbacking

00:49:33 ◼   ► to get touch Macs ready to rock.

00:49:36 ◼   ► Curtis Hard writes, in Mac OS 27, NS Scroll View has a new refresh controller, which takes

00:49:42 ◼   ► an NS refresh controller, which allows native Pult to refresh.

00:49:46 ◼   ► Try it in Safari on a web page.

00:49:48 ◼   ► Yeah, that's going to feel so weird to do on a Mac, but you know, touch Macs, you got to

00:49:51 ◼   ► do it.

00:49:52 ◼   ► Steve Charlton-Smith appears yet again.

00:49:54 ◼   ► AppKit isn't just getting touch gesture support, but NS Toolbar now has iOS-style fluid morphing

00:49:59 ◼   ► animations when used with a touchscreen, too.

00:50:01 ◼   ► Long presses trigger a right-click, momentum pinch and scroll work great, very clearly telegraphing

00:50:07 ◼   ► upcoming touchscreen Macs.

00:50:08 ◼   ► Anywhere you see a liquid glass stretch or highlight effect on iOS now functions in the

00:50:12 ◼   ► same way on macOS when used with touch.

00:50:14 ◼   ► Here's Calculator on macOS showing off some of the fluid interactions, and we'll put a link

00:50:19 ◼   ► in the show notes.

00:50:20 ◼   ► It has this tweet, or toot, or whatever they're called, and a couple follow.

00:50:24 ◼   ► I have GIFs that indicate what Steve is talking about.

00:50:27 ◼   ► Yeah, all the wiggly effect where you're like on your phone when you're like, most people

00:50:31 ◼   ► don't do it because they just press buttons.

00:50:32 ◼   ► But if you were to press your finger on the button on a phone and move your finger around,

00:50:35 ◼   ► you can see the button kind of stretches and follows your finger.

00:50:38 ◼   ► All that stuff is in macOS controls when used with touch.

00:50:42 ◼   ► And they added a whole bunch of new APIs, a whole bunch of new things you can set on it

00:50:45 ◼   ► to say, this isn't in the notes because it just came out today, but I was recently showing

00:50:50 ◼   ► that there's keys that you can set to make all the menus look essentially like iPad menus.

00:50:54 ◼   ► And if you pull down the menu and keep your finger down on the screen and move it around,

00:50:59 ◼   ► you'll pull the menu a little bit.

00:51:01 ◼   ► I don't think those effects really add anything, but they're part of liquid glass.

00:51:04 ◼   ► That's the liquid part of it.

00:51:05 ◼   ► I mean, they don't take away too much because I feel like they don't bother you if you don't

00:51:09 ◼   ► intentionally trigger them.

00:51:11 ◼   ► But I think it's just basically a waste of time.

00:51:12 ◼   ► But anyway, all that's coming to the Mac.

00:51:14 ◼   ► So do you plan to touch your screens?

00:51:17 ◼   ► Oh, no, definitely not.

00:51:18 ◼   ► Well, John, that's not even a question.

00:51:20 ◼   ► For me, I don't plan to, but there are definitely occasions, especially if I'm like flip-flopping

00:51:23 ◼   ► between iPad and Mac, which doesn't happen often, to be fair, but it does happen.

00:51:28 ◼   ► And there's definitely occasions that I've like started to reach up to like swipe at something

00:51:34 ◼   ► just briefly.

00:51:35 ◼   ► I don't feel like, and here I hear this is Casey 101.

00:51:40 ◼   ► I don't feel like there's anything that this solves in my life.

00:51:43 ◼   ► Like it would be nice, I guess, to occasionally be able to do this.

00:51:45 ◼   ► But sitting here now, I don't yearn for a touchscreen Mac.

00:51:49 ◼   ► That said, I don't remember what report it was, and I'm sure we talked about it on the show,

00:51:54 ◼   ► but when they talked about how when you use touch, there'll be like contextual menus or

00:51:58 ◼   ► something like that that pop up like path style, like around your finger.

00:52:02 ◼   ► I'm sure there's a German.

00:52:03 ◼   ► Yeah, controls that are more suited to fingers than to mouse blinders.

00:52:06 ◼   ► Right, exactly.

00:52:07 ◼   ► That could be interesting.

00:52:09 ◼   ► It could be really cool.

00:52:10 ◼   ► So we'll see what happens.

00:52:10 ◼   ► Yeah.

00:52:11 ◼   ► I mean, it's this type of thing that once you start doing it, you can't go back.

00:52:13 ◼   ► Like that's why everyone wants it.

00:52:15 ◼   ► And judging by my children's laptop screens, they're constantly touching them, even though

00:52:19 ◼   ► they're not touch screens.

00:52:20 ◼   ► So I guarantee both my kids will be touching their screen like the same amount they do now,

00:52:24 ◼   ► which is to say all the time.

00:52:26 ◼   ► I hope they put that oleophobic coating on the screen because that will just help them

00:52:30 ◼   ► keep their screens be less gross.

00:52:31 ◼   ► But I won't do it because I don't tend to use laptops.

00:52:34 ◼   ► And if I did use a laptop, I wouldn't do it because I don't want to get into that habit

00:52:37 ◼   ► because it makes the screen uglier for me.

00:52:39 ◼   ► But, you know, I have an iPad.

00:52:40 ◼   ► I touch it all the time.

00:52:41 ◼   ► So yeah, once you go touch, you can't go back.

00:52:45 ◼   ► I wonder, like, you know, because I, as I've, as I have with my most recent, like laptop

00:52:50 ◼   ► laptop, I have the, the nanotexture screen.

00:52:52 ◼   ► And I've told, I've said on the show many times how much I absolutely love nanotexture.

00:52:57 ◼   ► And I, I don't plan to buy any future laptops without it as long as if they are offered with

00:53:03 ◼   ► it.

00:53:03 ◼   ► And I wonder, like, you know, one of the downsides of nanotexture is that it is harder to clean.

00:53:08 ◼   ► And, you know, it does, just like the, the other type of screen, I guess the glossy

00:53:12 ◼   ► type of screen, it does get those like keyboard imprint, uh, lines in the middle of the screen

00:53:18 ◼   ► that are slowly become impossible to clean.

00:53:21 ◼   ► I don't know how to clean those.

00:53:23 ◼   ► I like once it gets, once you get like to a certain point of those imprint marks, they

00:53:28 ◼   ► seem to be permanent.

00:53:29 ◼   ► And that's always been the case even before nanotexture.

00:53:31 ◼   ► I wonder like, you know, the, I, when, when Apple says nanotexture, they, that refers to

00:53:36 ◼   ► a few different processes they've had over time.

00:53:38 ◼   ► Like the, the original Pro Display XDR was the first one.

00:53:41 ◼   ► Um, and that did one version of it.

00:53:43 ◼   ► And then the current like Mac laptops that offer it are a different version of it.

00:53:49 ◼   ► And then the iPad Pro that has it is yet a different version of nanotexture.

00:53:54 ◼   ► I wonder if maybe they would move to the iPad Pro version of it for the touch laptop.

00:54:00 ◼   ► Um, and, and whether that would be easier to clean.

00:54:02 ◼   ► No, they'll just make a fourth version.

00:54:03 ◼   ► They'll just keep refining it.

00:54:05 ◼   ► Yeah, I imagine it'll be more like the iPad one, but, uh, maybe a newer revision.

00:54:09 ◼   ► All right.

00:54:10 ◼   ► Apple has updated the human interface guidelines.

00:54:12 ◼   ► So the HIG for menus.

00:54:14 ◼   ► Uh, it says in the HIG, use menu item icon sparingly and with purpose.

00:54:20 ◼   ► Icons allow people to find menu items more quickly and help clarify what selecting an item does.

00:54:24 ◼   ► Use an icon to highlight the most common actions and key features of your app.

00:54:27 ◼   ► File system locations, connected devices, visual concepts like rotating or flipping an image

00:54:31 ◼   ► and user generated content like folders and documents.

00:54:34 ◼   ► Don't display an icon if you can't find one that clearly represents menu item.

00:54:37 ◼   ► And they give a very curious example of, uh, the seven days of the week where they've chosen

00:54:43 ◼   ► odd choices for icons for each day of the week with a little gray X below it indicating don't do this.

00:54:49 ◼   ► And then they have a different version.

00:54:50 ◼   ► That's just the text for the seven days of the week.

00:54:52 ◼   ► And they have the little green check mark.

00:54:54 ◼   ► Yes, do that.

00:54:55 ◼   ► Yeah.

00:54:55 ◼   ► I love that.

00:54:56 ◼   ► I love that the, the don't do this is just Tahoe.

00:54:59 ◼   ► Yep.

00:54:59 ◼   ► Yeah.

00:55:00 ◼   ► The idea is you have to have come up with an icon for every item.

00:55:02 ◼   ► So good luck.

00:55:03 ◼   ► Yeah.

00:55:03 ◼   ► Like this thing, we, we literally told you a year ago, this is the right way to do it.

00:55:08 ◼   ► And now we're saying this exact same thing is the wrong way to do it.

00:55:12 ◼   ► And just, this is the clarifying from, uh, last episode where we're like, well, is there a way

00:55:16 ◼   ► to turn off the icons?

00:55:17 ◼   ► Is it some kind of user setting?

00:55:19 ◼   ► Is it a developer choice?

00:55:20 ◼   ► Um, I'm pretty sure it is only developer control.

00:55:23 ◼   ► I didn't see anything in the UI for it, but regardless, the important point is regardless

00:55:27 ◼   ► of how this is implemented in terms of who gets to decide which icons appear where Apple's advice

00:55:33 ◼   ► now reverted essentially to its pre 26 advice, which is don't put an icon in every menu item.

00:55:38 ◼   ► That's not what we want you to do.

00:55:39 ◼   ► So the put an icon in, and again, the 26 interface guidelines also did not say to put an icon in

00:55:46 ◼   ► everything.

00:55:46 ◼   ► We read them on the show to clarify, but they were like, but you know, put them on most stuff,

00:55:51 ◼   ► I guess.

00:55:52 ◼   ► And repeat them in this weird way.

00:55:53 ◼   ► It was like the guidelines were incoherent last time.

00:55:56 ◼   ► The guidelines didn't say you must put an icon in every menu item.

00:55:59 ◼   ► But in practice, what Apple did was basically put an icon in almost every menu item and it

00:56:03 ◼   ► was ridiculous.

00:56:04 ◼   ► And the advice was vague and didn't make it clear.

00:56:07 ◼   ► When am I supposed to use icons or not?

00:56:09 ◼   ► Should I copy what Apple's apps do?

00:56:11 ◼   ► Like the 26 guidelines didn't give developers enough information to know what they should

00:56:15 ◼   ► do.

00:56:15 ◼   ► Uh, the earlier guidelines for all the other years of the Mac OS did give clear advice.

00:56:21 ◼   ► And now the advice is once again, clear.

00:56:23 ◼   ► And not only is it clear, it is the pre 26 advice, which is don't put icons everywhere.

00:56:27 ◼   ► They do say in a subsequent part of this, uh, to, as they say, apply uniform visual treatment

00:56:32 ◼   ► across menu items in the same group for visual consistency and balance provide icons for all

00:56:36 ◼   ► menu items in a group or none of them.

00:56:38 ◼   ► And it's showing like, if there's a group of menu items, like it's like a divider.

00:56:41 ◼   ► And then there's menu items for doing stuff with windows, like move and resize or full

00:56:46 ◼   ► screen or whatever.

00:56:46 ◼   ► If they have icons to let you know, like when this is the tile thing, the tiles to the right

00:56:51 ◼   ► half of the screen, this is the thing and tiles to the left or whatever.

00:56:53 ◼   ► They should all have icons and not just like one or two of them, because these are all commands

00:56:57 ◼   ► where icons are useful.

00:56:57 ◼   ► So again, the advice is clear.

00:56:59 ◼   ► Default, your default should be no icons.

00:57:02 ◼   ► But if you think an icon, as they say, will help clarify what selecting an item does, then

00:57:08 ◼   ► do use an icon, but all icons that do that similar, all menu items that do that similar

00:57:12 ◼   ► thing should also have icons.

00:57:13 ◼   ► So yeah, back to sanity.

00:57:15 ◼   ► Wasn't there a convention in the past?

00:57:18 ◼   ► Maybe I mean, maybe I just like misinterpreted this.

00:57:20 ◼   ► Wasn't there a convention in the past that menu items in macOS would have an icon only if

00:57:26 ◼   ► they matched a toolbar item in that app and it would be the same icon as the toolbar item?

00:57:30 ◼   ► I think the Hig might have said that at one point, but I don't think that was strongly

00:57:35 ◼   ► suggested for a while.

00:57:36 ◼   ► Maybe it was in the pre-26 ones.

00:57:38 ◼   ► But yeah, the thing with toolbar icons is they have changed so much over the years that I think

00:57:43 ◼   ► they probably just deleted that advice in menus if it was ever there, just because it

00:57:47 ◼   ► adds confusion.

00:57:48 ◼   ► It's like, well, you know, in the current design system, it's impossible to do that because

00:57:54 ◼   ► the pictures on the toolbars would never fit in toolbar icons.

00:57:57 ◼   ► Like if you think about the toolbar images in like macOS 10.0, those wouldn't fit in menus,

00:58:02 ◼   ► especially in the pre-retina days.

00:58:03 ◼   ► This was just literally impossible, but yeah.

00:58:06 ◼   ► All right.

00:58:07 ◼   ► That is the end of the macOS section.

00:58:09 ◼   ► You can look at your timestamp now and compare it to where we were so very, very long ago.

00:58:13 ◼   ► Don't forget to subtract the Microsoft Bob time.

00:58:17 ◼   ► Oh yeah, that's very important.

00:58:18 ◼   ► Also, much of what we were saying was way beyond macOS.

00:58:22 ◼   ► Like the icons are universal.

00:58:23 ◼   ► Whose side are you on?

00:58:24 ◼   ► I'm just saying, look, we're talking about Apple OS's 27 and Microsoft Bob.

00:58:28 ◼   ► Well, that was the macOS section.

00:58:30 ◼   ► The icons were, yes, but that's, we have the same problem as Apple.

00:58:33 ◼   ► Maybe we should have just talked about improvements, trust and safety in AI.

00:58:37 ◼   ► Well done.

00:58:38 ◼   ► Well done.

00:58:40 ◼   ► All right.

00:58:40 ◼   ► With regard to iOS 27, Mario Guzman writes, the new pulldown animation for Siri AI in iOS 27

00:58:46 ◼   ► is pretty nice.

00:58:47 ◼   ► It's basically the old pull to refresh animation from iOS 6.

00:58:50 ◼   ► Nice, fun, fluid.

00:58:52 ◼   ► The animation will also start from where you drag your fingers.

00:58:54 ◼   ► So it won't always be animating in from the center.

00:58:56 ◼   ► And at a glance, this looks pretty great.

00:58:58 ◼   ► Yeah.

00:58:59 ◼   ► Check out the movie and the toot that we'll link in the show notes is basically you pull

00:59:02 ◼   ► down from the top and it's like, you're pulling a black blob from the top, but like it bends

00:59:07 ◼   ► down the top of the screen.

00:59:08 ◼   ► And once the black blob gets far enough away from the top of the screen, the screen like

00:59:12 ◼   ► snaps closed behind it.

00:59:13 ◼   ► And that blob becomes the floating Siri thing.

00:59:15 ◼   ► I'm still not a big fan of the glossy black, transparent, refracting look for Siri stuff.

00:59:21 ◼   ► But it is at least consistent and it is mostly legible.

00:59:25 ◼   ► And these little fun bits where you get to pull out a little blob of ink are also fun.

00:59:29 ◼   ► All right.

00:59:31 ◼   ► That was the end of the iOS 27 section.

00:59:33 ◼   ► Let's move on to iPadOS 27.

00:59:35 ◼   ► And Steve Trout and Smith writes, the iPad's menu bar is now left aligned and can be set

00:59:39 ◼   ► to display permanently.

00:59:40 ◼   ► When hidden, it will still show the app name.

00:59:43 ◼   ► So you know which app is currently front most.

00:59:44 ◼   ► Yeah.

00:59:45 ◼   ► Whoever was insisting, we can't just put a menu bar on the iPad.

00:59:49 ◼   ► Like, isn't that surrendering?

00:59:50 ◼   ► It's like, well, what if we center it?

00:59:51 ◼   ► Then it's like, we're doing something different.

00:59:52 ◼   ► And then 27 said, just put, just left align it.

00:59:55 ◼   ► Like just the menu bar is left aligned.

00:59:58 ◼   ► Just do it and make an option to have it permanently visible.

01:00:01 ◼   ► And when it's not visible, make sure you show the app name.

01:00:04 ◼   ► It's like, there's just slowly, like we, you know, we, we tried for a decade or whatever

01:00:07 ◼   ► to come up with a UI that's better than the Mac UI.

01:00:10 ◼   ► We couldn't do it.

01:00:11 ◼   ► Not to say it's not possible, but all I can say is we didn't do it.

01:00:13 ◼   ► So let's just do the things we know work.

01:00:16 ◼   ► That is the end of the iPadOS 27 section.

01:00:18 ◼   ► Let's talk about tvOS 27, which in John's defense, there was almost no mention of in the

01:00:22 ◼   ► keynote.

01:00:23 ◼   ► We should talk about hardware support, which is to say tvOS 27 drops support for the Apple

01:00:28 ◼   ► TV HD from 2015 and the Apple TV 4K first generation from 2017.

01:00:32 ◼   ► As I have lamented many times recently and having come fresh off of a vacation where I

01:00:39 ◼   ► brought, I believe it's the HD from 2015, whatever the, I think this is the first one that allowed

01:00:45 ◼   ► for, um, apps.

01:00:47 ◼   ► Wow.

01:00:48 ◼   ► Uh, this is an old Apple TV and it shows, uh, yeah, it's not great.

01:00:52 ◼   ► And I would love, as I've said so many times, I would love to trickle down and upgrade my travel

01:00:59 ◼   ► slash tailgate Apple TV to something nicer.

01:01:01 ◼   ► And apparently this is as nice as it will ever be because it will not get the tvOS 27.

01:01:05 ◼   ► And I'm not really sad about that since it's already hurting so bad, but give me a new Apple

01:01:10 ◼   ► TV, Apple, please, please.

01:01:12 ◼   ► I want it.

01:01:12 ◼   ► Please.

01:01:13 ◼   ► I'm kind of surprised that the Apple TV 4K got dropped, but then I saw what year it came

01:01:16 ◼   ► out.

01:01:17 ◼   ► I was like, okay, it seems, it seems so recent.

01:01:19 ◼   ► Apple TV 4K.

01:01:20 ◼   ► That's the new one.

01:01:21 ◼   ► No.

01:01:21 ◼   ► First generation was 2017.

01:01:22 ◼   ► Yeah.

01:01:23 ◼   ► Yep.

01:01:23 ◼   ► Turns out Apple TVs don't come out that often.

01:01:25 ◼   ► Yep.

01:01:26 ◼   ► Who knew, uh, watchOS 27, we're done with tvOS, which in, again, in John's defense, that

01:01:31 ◼   ► makes sense.

01:01:31 ◼   ► WatchOS 27.

01:01:32 ◼   ► Let's also talk about their hardware support.

01:01:34 ◼   ► Uh, Hartley Charlton, Mac rumors says Apple confirmed that the only Apple watch models compatible

01:01:38 ◼   ► with watchOS 27 are series nine, series 10, series 11, ultra two, ultra three, and SE three.

01:01:43 ◼   ► Uh, initially they had incorrectly said that the series nine was not supported, but that was

01:01:48 ◼   ► an oops.

01:01:49 ◼   ► It is supported.

01:01:50 ◼   ► So again, nine, 10, 11, not the ultra one, but yes for ultra two and three and yes for the

01:01:55 ◼   ► SE three.

01:01:55 ◼   ► When they said the nine wasn't supported, I'm like, really?

01:01:58 ◼   ► Jeez, this is harsh.

01:01:59 ◼   ► I'm glad the nine came back in, but still this is, I mean, Marco would know better than I,

01:02:04 ◼   ► is this the most harsh hardware cutoff ever for a watchOS release?

01:02:08 ◼   ► Yeah.

01:02:09 ◼   ► And, and I think that's, it's a casualty of the very slow update pace they chose for the

01:02:16 ◼   ► Apple watch SOCs.

01:02:17 ◼   ► You know, as we talked about in when, when, whenever new Apple watches come out, they have

01:02:22 ◼   ► gone like two or three versions in a row having the same processor on, on Apple watches.

01:02:27 ◼   ► So this is, this is what happens is like all of a sudden when you have, when you want to

01:02:32 ◼   ► cut off a processor that was used for three generations in a row, uh, you end up slicing

01:02:37 ◼   ► off a whole lot of models.

01:02:39 ◼   ► So that's, that's why this is happening.

01:02:41 ◼   ► We are sponsored this week by Quince.

01:02:45 ◼   ► Now you've potentially heard one of our former Quince spots or maybe on a different podcast,

01:02:50 ◼   ► and you might think to yourself, well, Quince just does clothing, right?

01:02:54 ◼   ► Oh no, my friend.

01:02:55 ◼   ► Oh no, not at all.

01:02:56 ◼   ► Not only does Quince do clothing for people who buy men's clothing, for people who buy women's

01:03:01 ◼   ► clothing, but they do child clothing.

01:03:03 ◼   ► In fact, Declan got, uh, some like stretchy pants, some, some like kind of sweatpants that

01:03:09 ◼   ► he really, really enjoys.

01:03:10 ◼   ► And Michaela got this absolutely adorable, uh, long sleeve dress with like hearts and different

01:03:16 ◼   ► colors all over it.

01:03:17 ◼   ► And pockets, which I don't know anything.

01:03:19 ◼   ► And even I know that that's really exciting.

01:03:21 ◼   ► They do stuff for the home, for travel.

01:03:23 ◼   ► They do bags and accessories, jewelry, beauty and wellness, all sorts of different things.

01:03:27 ◼   ► And what's great about Quince is that all of their clothes are really well made.

01:03:30 ◼   ► I had gotten a set of linen pants at the end of last summer.

01:03:34 ◼   ► I've been so excited to bring them to the beach this summer so I can use them in their like

01:03:38 ◼   ► native habitat.

01:03:39 ◼   ► And then I forgot them at home.

01:03:42 ◼   ► I'm genuinely upset by this.

01:03:44 ◼   ► I literally came home from the beach today and I completely forgot to bring them.

01:03:48 ◼   ► I'm very sad, but I've got plenty of summer left, including one more beach trip.

01:03:51 ◼   ► So there's plenty of opportunity to put on these excellent linen pants that I got from

01:03:54 ◼   ► Quince.

01:03:54 ◼   ► So if you are interested in any of the stuff that they offer, including, but not limited

01:04:00 ◼   ► to clothing.

01:04:00 ◼   ► One of the great things that, that Quince does is that they work directly with ethical

01:04:05 ◼   ► factories.

01:04:05 ◼   ► So this way you cut out the middle people and you're paying for quality, not brand markup.

01:04:10 ◼   ► Like I said, they have things beyond clothes.

01:04:12 ◼   ► They have upholstered sofas, they have ceramic cookware, premium bedding, et cetera.

01:04:16 ◼   ► You can almost do your entire house with just Quince stuff.

01:04:21 ◼   ► And that's pretty awesome in my book.

01:04:22 ◼   ► So elevate your summer wardrobe.

01:04:24 ◼   ► Go to quince.com slash ATP for free shipping on your order and 365 day returns.

01:04:30 ◼   ► It's now available in Canada as well.

01:04:32 ◼   ► That's Q-U-I-N-C-E dot com slash ATP for free shipping and 365 day returns.

01:04:38 ◼   ► Quince.com slash ATP.

01:04:40 ◼   ► Thank you to Quince for sponsoring the show.

01:04:42 ◼   ► Apple's AI tech talk.

01:04:48 ◼   ► So when we were there two years ago, I believe it was, what they did was they had the keynote

01:04:54 ◼   ► and then most of us, I don't remember if we were told to, if it was optional or what the

01:04:58 ◼   ► deal was, but most of us were shuffled into the Steve Jobs theater.

01:05:01 ◼   ► And that's when we saw, um, I just seen, who were the, who was she talking to?

01:05:05 ◼   ► I can't recall.

01:05:05 ◼   ► Talking to, uh, John G.

01:05:06 ◼   ► Andrea and maybe Mike Rockwell.

01:05:09 ◼   ► And I forget it was that some Apple executives.

01:05:11 ◼   ► Yeah.

01:05:12 ◼   ► And she, I mean, I really like I just seen, but she was clearly giving nothing but softball

01:05:16 ◼   ► questions to Apple.

01:05:16 ◼   ► Uh, I don't know if they did anything last year or not.

01:05:19 ◼   ► It doesn't really matter.

01:05:19 ◼   ► But this year what they did was, uh, selected media was invited to an on the record technical

01:05:25 ◼   ► deep dive into the bold new architecture, enabling Apple intelligence capabilities.

01:05:29 ◼   ► That's Apple's description.

01:05:30 ◼   ► I'm assuming that quote is from the verge, uh, but there has quotation marks around it.

01:05:34 ◼   ► So I'm assuming it's Apple talking when they say that.

01:05:36 ◼   ► So it was Craig Federighi, Amar Subramanya, I think I got that close to right, uh, who's

01:05:42 ◼   ► the vice president of AI, Mike Rockwell, who's the Siri lead and Sebastian Marino Mess, uh,

01:05:46 ◼   ► the software VP.

01:05:47 ◼   ► There are some images from tech radar, which we'll put in the show notes of Steve talking

01:05:51 ◼   ► in front of various like diagrams and whatnot.

01:05:54 ◼   ► That's Craig, not Steve.

01:05:56 ◼   ► Oh God.

01:05:56 ◼   ► Wow.

01:05:56 ◼   ► Sorry.

01:05:57 ◼   ► Yes.

01:05:57 ◼   ► Thank you for the correction.

01:05:58 ◼   ► Good grief.

01:05:59 ◼   ► Uh, yeah, I, I, I, this was given in the, like the, what was it?

01:06:02 ◼   ► The developer center.

01:06:03 ◼   ► It's a tiny, I've never even been in this room, although I, maybe I've walked by it,

01:06:07 ◼   ► but a tiny little room.

01:06:08 ◼   ► Like it's not, Steve jobs theater is big.

01:06:10 ◼   ► Oh, this wasn't in the theater.

01:06:11 ◼   ► My bad.

01:06:12 ◼   ► That's where I thought it was.

01:06:12 ◼   ► No, this was in, it was in the developer center, that tiny room that has like four rows.

01:06:17 ◼   ► It's where the talk show live was like during like the COVID lockdown year.

01:06:20 ◼   ► Do you remember?

01:06:20 ◼   ► My bad.

01:06:21 ◼   ► I, I didn't realize.

01:06:22 ◼   ► Yeah.

01:06:22 ◼   ► But, but it's, but it's, it's so small.

01:06:24 ◼   ► It's like four or five rows deep.

01:06:26 ◼   ► It's just a one tiny little room.

01:06:27 ◼   ► And I don't, I don't understand why they didn't do this as a WWDC session or publicly broadcast

01:06:33 ◼   ► or whatever, because it was clearly on the record.

01:06:35 ◼   ► Press was there and you know, people were taking pictures.

01:06:39 ◼   ► I don't know if anyone just recorded the whole thing on their phone, but I kind of wish they

01:06:41 ◼   ► did because everybody was forced to say, here's what my AI transcription note taking app wrote.

01:06:46 ◼   ► Here's what the fastest typist at our company wrote.

01:06:48 ◼   ► Like we're going to read a little bit from a chance Miller from nine to five macro, assuming

01:06:52 ◼   ► had really good fingers and typed all this stuff up, but even that had typos or maybe it

01:06:56 ◼   ► was anyway, tons of info in this AI tech talk that I guess is not really, I mean, is it relevant

01:07:04 ◼   ► to developers?

01:07:05 ◼   ► There's, there's so many WWDC sessions.

01:07:07 ◼   ► You could argue like, do developers need to know this?

01:07:09 ◼   ► Or is this just like Apple talking about a cool thing that they did?

01:07:12 ◼   ► Sometimes it's a little bit of each.

01:07:14 ◼   ► You're like, well, it does help if developers know what's under the covers.

01:07:16 ◼   ► It'll give them an idea to how to correctly assess a feature and know if it's right for them,

01:07:20 ◼   ► but it's just cool.

01:07:21 ◼   ► Um, and so they did this to explain their AI architecture fascinating, but I've had to piece

01:07:27 ◼   ► it because I wasn't there.

01:07:28 ◼   ► I had to piece it together from everybody's reports.

01:07:31 ◼   ► And some of the, they, they only invited a tiny select set of press because the room is

01:07:36 ◼   ► tiny.

01:07:36 ◼   ► You can't fit that many people.

01:07:37 ◼   ► I don't know why they didn't do this in the Sheep's App Theater.

01:07:39 ◼   ► Maybe they thought most people aren't interested in this nerdy crap.

01:07:42 ◼   ► Uh, Craig Federee is interested in this nerdy crap, clearly.

01:07:44 ◼   ► Um, but like the, what the, the press that they invited, some of them were into it and some

01:07:50 ◼   ► of them were like, ah, they just talked about AI.

01:07:51 ◼   ► And like, they latched onto a few of like the intercompany drama things, but I would have

01:07:56 ◼   ► loved to learn even more.

01:07:57 ◼   ► So I learned as much as I could by looking at what other people reported and now we'll piece

01:08:01 ◼   ► it together here.

01:08:01 ◼   ► All right.

01:08:03 ◼   ► So we have those images, like we said, and I'm going to read a bunch from Chance Miller,

01:08:07 ◼   ► as John said.

01:08:08 ◼   ► So according to Federighi, and these aren't as best as we know, either verbatim or near

01:08:13 ◼   ► verbatim quotes.

01:08:14 ◼   ► Yeah, this is people attempting to transcribe.

01:08:16 ◼   ► So if there are errors, it's just errors in transcription.

01:08:17 ◼   ► Of course, we don't have the Gemini app as our app.

01:08:21 ◼   ► In fact, none of that client code is part of how we run on iOS.

01:08:24 ◼   ► For these models, we use none of the models that Google deploys to their customers, nor

01:08:29 ◼   ► do we use the infrastructure and means by which they deploy models to their customers.

01:08:33 ◼   ► And then when it comes to the knowledge base, we of course don't use Google search or anything

01:08:37 ◼   ► like that as the foundation of our system.

01:08:39 ◼   ► So I hope that's clear.

01:08:40 ◼   ► The amount of Google assistant we use is none.

01:08:43 ◼   ► So here, Craig, and I encourage people to look at these pictures in a sec, right?

01:08:46 ◼   ► This is the best images I could find because I think they were in the front row and there

01:08:49 ◼   ► are slides behind them, which I just released the slides, man.

01:08:51 ◼   ► But anyway, this shows an architecture diagram of like, like the phone on the left with like

01:08:56 ◼   ► a stack of round rec sewing, like the system experience and the AI assistant.

01:09:02 ◼   ► And then down from that are the apps.

01:09:03 ◼   ► And then down from that is the system orchestrator.

01:09:05 ◼   ► And like, it's showing like the block diagram.

01:09:07 ◼   ► And then there's a line going to the right showing the cloud part of it.

01:09:09 ◼   ► And we're going to talk about both parts of them.

01:09:10 ◼   ► But this first part where Craig's talking, he's like, look, you know, I know we did a deal

01:09:15 ◼   ► with Google, but he's trying to explain, and I'm not sure how good a job he's doing, but

01:09:19 ◼   ► I think what he's trying to explain is when we say we're using Google Gemini, um, that

01:09:24 ◼   ► doesn't mean that like, you know, when you go to gemini.google.com and you type in that

01:09:27 ◼   ► box, that's not what we're using.

01:09:29 ◼   ► That is an entire like piece of software that under the cover uses Gemini models and stuff,

01:09:36 ◼   ► but we're not using Google assistant.

01:09:39 ◼   ► We're not using the Gemini chat bot.

01:09:41 ◼   ► Uh, and he goes further to say, we're not using the infrastructure or that they use to

01:09:47 ◼   ► run the backend for those services either.

01:09:49 ◼   ► That is entirely a Google thing.

01:09:51 ◼   ► As he says, the amount of Google assistant we use is none.

01:09:54 ◼   ► What he's trying to say is we wrote our own like thing that uses models behind the scene

01:10:00 ◼   ► that shares zero code with Google.

01:10:02 ◼   ► And when we talk to models behind the scenes, those models are not any of the models that

01:10:07 ◼   ► Google uses to power any of its customer facing products.

01:10:10 ◼   ► So this is Apple really trying to draw the line.

01:10:13 ◼   ► So you don't think like, oh, it's a shiny black blob where I get to talk to Google Gemini.

01:10:17 ◼   ► Just like I do at Gemini at Google.com.

01:10:19 ◼   ► That's not it at all.

01:10:20 ◼   ► Continuing from Craig, the system orchestrator is key to the privacy architecture of our

01:10:25 ◼   ► entire system.

01:10:26 ◼   ► It's what coordinates requests against things like the app toolbox that provides access to

01:10:30 ◼   ► actions within your apps.

01:10:31 ◼   ► The semantic, the spotlight semantic index to access personal content, to help fulfill your

01:10:35 ◼   ► request.

01:10:36 ◼   ► And even things like on-screen context to understand what you might be looking at at the

01:10:39 ◼   ► moment you're making a request.

01:10:40 ◼   ► This in turn is built on a set of powerful on-device models.

01:10:44 ◼   ► These handle everything from understanding speech to synthesizing the voice that speaks

01:10:48 ◼   ► back to you, to understanding visually the environment in the on-screen context, understanding

01:10:52 ◼   ► text that might be on the screen, as well as a whole set of other models.

01:10:55 ◼   ► For some requests, models are capable of processing your Siri requests entirely locally on the device.

01:11:02 ◼   ► But sometimes the system orchestrator realizes that it's a more sophisticated question, and

01:11:06 ◼   ► then it wants to draw on greater intelligence.

01:11:09 ◼   ► It does that by contacting our models running in Private Cloud Compute.

01:11:13 ◼   ► Private Cloud Compute hosts our third generation of Apple Foundation models, from our AFM Cloud

01:11:19 ◼   ► and AFM Cloud Pro models to our AFM Fusion model and Image model.

01:11:23 ◼   ► These are the models that are the product of our collaboration with Google.

01:11:26 ◼   ► These are models designed specifically for our Apple intelligence experiences.

01:11:31 ◼   ► Finally, when you make a request involving current events or other elements of world knowledge,

01:11:35 ◼   ► those responses are grounded by accessing Apple's world knowledge service.

01:11:39 ◼   ► This is something that we've built over many years and provides a great source of information

01:11:43 ◼   ► to satisfy your request.

01:11:45 ◼   ► So that is kind of describing the layer cake to saying, we have a whole bunch of models

01:11:50 ◼   ► that are running on your device.

01:11:51 ◼   ► And we wrote software, the system orchestrator that's above them, that when you give a query

01:11:54 ◼   ► through Siri, which is an app that we wrote, and it goes to the system orchestrator, which

01:11:57 ◼   ► is an app that we wrote, and it decides where is it going to send your request.

01:12:00 ◼   ► And we have a bunch of different models that run on device that do all these things that

01:12:03 ◼   ► we listed and even more.

01:12:04 ◼   ► And it decides, you know, that's what happens when you're like, I'm using, I'm using, you

01:12:08 ◼   ► know, Google Gemini to do something, or I'm using like a codex coding agent, like, and

01:12:13 ◼   ► you're just typing, it's not like it takes what you type, sends it to an LLM, gets a

01:12:17 ◼   ► response.

01:12:17 ◼   ► There's a whole piece of software under there that is deciding, you know, there's the system

01:12:21 ◼   ► prompt and all the other stuff, but like, it's composing prompts to send to different

01:12:25 ◼   ► agents, getting the responses back, using those to figure out where and when it's going to

01:12:30 ◼   ► send the next response.

01:12:31 ◼   ► Like, that's, that's the product.

01:12:32 ◼   ► Like, we've talked a lot about like, LLMs being commoditized.

01:12:35 ◼   ► Is this a, is this thing where anyone can have an LLM and it's fine?

01:12:38 ◼   ► LLMs may be commoditized, but the products that we're all using are not,

01:12:43 ◼   ► not commodities.

01:12:44 ◼   ► It's kind of like saying like the underlying search engine technology may be commoditized,

01:12:49 ◼   ► but like the things that make the app or the things that, you know, index the web and all

01:12:54 ◼   ► that stuff are not commodities because that's what makes a product good.

01:12:57 ◼   ► And here's Apple describing the app that it made on all our phones and everything that does

01:13:03 ◼   ► all this stuff that eventually, yes, sends requests to either an on-device model or a cloud

01:13:06 ◼   ► model and private cloud compute.

01:13:08 ◼   ► But that is the very, very last link in a long chain and, you know, requests are going

01:13:14 ◼   ► back and forth all internally with all this stuff many times over before you get a response.

01:13:19 ◼   ► And that's, you know, that's what Apple's saying.

01:13:21 ◼   ► That's what we wrote.

01:13:21 ◼   ► And so you, again, you should not expect this to look, behave, or be anything like Google

01:13:26 ◼   ► Gemini because it just doesn't share anything with it except for the fact that the various

01:13:31 ◼   ► underlying models that it is orchestrating among were created in collaboration with Google

01:13:36 ◼   ► Google in a fuzzy degree that we'll get to in a little bit.

01:13:38 ◼   ► And then even going so far as to say, like, when we have to find something like the world

01:13:43 ◼   ► knowledge thing, which is Apple's fun phrase, it doesn't go to Google search either.

01:13:46 ◼   ► Like, yes, I know they did a collaboration with Google, but it's not going to Google search

01:13:49 ◼   ► Apple made.

01:13:50 ◼   ► And I think this might've been in the notes and we never got to it in the past year.

01:13:53 ◼   ► But Apple's world knowledge service, I don't know if we ever talked about it on the show,

01:13:57 ◼   ► but like the rumor was they're making a service that knows things about the world.

01:14:01 ◼   ► And so when you ask a question about the world, it doesn't go to Google search.

01:14:05 ◼   ► It goes to Apple's word knowledge service.

01:14:07 ◼   ► And Dan Morin's had some things to say about that.

01:14:10 ◼   ► Righto.

01:14:11 ◼   ► So Dan tooted.

01:14:12 ◼   ► I asked Siri AI what it knew about me.

01:14:15 ◼   ► Well, about, quote, Dan Morin, quote.

01:14:17 ◼   ► And it did a pretty well until I clicked on my hyperlinked name and got a Wikipedia entry

01:14:22 ◼   ► for an entirely different Dan M.

01:14:23 ◼   ► Fascinatingly, since Siri AI seems to largely be relying on Apple's web index, it doesn't appear

01:14:30 ◼   ► to have the ability to access the web directly.

01:14:31 ◼   ► Case in point, after my mistaken Wikipedia identification, someone did go out and create a Wikipedia page

01:14:36 ◼   ► for me, but Siri AI can't see it yet, even if you provide it with URLs.

01:14:42 ◼   ► So Dan asks, what about this?

01:14:44 ◼   ► And pastes in the URL.

01:14:45 ◼   ► And the response is, that URL currently leads to a, quote, Wikipedia does not an article with

01:14:49 ◼   ► this exact name, quote, page.

01:14:51 ◼   ► If you're seeing a live page of that link, it may have been created very recently and hasn't

01:14:55 ◼   ► been indexed by my search engines yet, or it might be a draft in the Wikipedia sandbox.

01:14:58 ◼   ► If you're able to open it, what does the first sentence say?

01:15:01 ◼   ► Why don't you tell me what it says?

01:15:02 ◼   ► So this is the blessing, the blessing and the curse of Apple creating its old world knowledge

01:15:08 ◼   ► thing.

01:15:08 ◼   ► Part of what makes Google valuable is their ability to crawl the web and keep their index

01:15:16 ◼   ► up to date.

01:15:17 ◼   ► And Apple, Google has made tremendous advances in that.

01:15:19 ◼   ► Back in the old days, you know, you'd make a change to a page and you'd have to wait for

01:15:24 ◼   ► it to appear in the Google index.

01:15:25 ◼   ► These days, Google indexes aggressively and amazingly and in proportion to how much change

01:15:30 ◼   ► in traffic things get.

01:15:32 ◼   ► And you can bet Google is indexed in Wikipedia like crazy.

01:15:35 ◼   ► How good is Apple going to be at keeping their world knowledge index up to date?

01:15:40 ◼   ► Now, granted, this is asking a lot.

01:15:42 ◼   ► Wikipedia page was probably created that day.

01:15:44 ◼   ► And then Dan Martin's asking it.

01:15:46 ◼   ► But just to show the way, you know, the Siri AI works, even if given the URL, it still won't

01:15:54 ◼   ► just go out and get that URL.

01:15:56 ◼   ► It says, well, that leads to a page that says there's no article with its name.

01:15:59 ◼   ► Yeah, I bet it does.

01:16:00 ◼   ► And the Apple world knowledge index that was created for some time in the past.

01:16:04 ◼   ► But right now it doesn't.

01:16:06 ◼   ► Hey, Siri AI, why don't you go to that web page right now and read what it says on it?

01:16:11 ◼   ► And it's like, I can't do that.

01:16:14 ◼   ► But if you open it up, what does it say?

01:16:16 ◼   ► Oh, so I just it's I'm not sure how they're going to resolve this.

01:16:20 ◼   ► Here's the latest top news.

01:16:23 ◼   ► That's staying in, baby.

01:16:25 ◼   ► Switch to Fox News podcasts.

01:16:27 ◼   ► Switch to CNN.

01:16:28 ◼   ► Oh, my God.

01:16:29 ◼   ► This is terrible.

01:16:30 ◼   ► Oh, no.

01:16:30 ◼   ► This is all together terrible.

01:16:31 ◼   ► What is this playing?

01:16:36 ◼   ► I guess I said the SIRI.

01:16:38 ◼   ► This is a good opportunity, Marco, for you to address the thing that people kept posting

01:16:42 ◼   ► to you this week on Mastodon.

01:16:45 ◼   ► Yeah.

01:16:45 ◼   ► So this has been going around that if you if you open up the audio for the WBC keynote

01:16:52 ◼   ► in an audio editor and show a frequency spectrogram.

01:16:56 ◼   ► And I verified this is this is actually true.

01:16:58 ◼   ► Whenever they say the word SIRI.

01:17:01 ◼   ► They strip out these tight, narrow frequency bands at 3K, 4K, 5K and 6K hertz.

01:17:10 ◼   ► It's a it's a tight enough strip that you don't really hear a difference.

01:17:16 ◼   ► Like as a human listening to the presentation, you don't it doesn't sound weird to you.

01:17:20 ◼   ► It doesn't sound muffled or muted because they're only doing these little tiny slices

01:17:24 ◼   ► of the frequency spectrum.

01:17:25 ◼   ► And the idea behind this is theoretically to be just enough of a reduction in the signal

01:17:33 ◼   ► in key areas that it defeats the recognition of everyone's devices listening to that keynote

01:17:40 ◼   ► like in the background so that everyone's phones don't activate when they hear somebody

01:17:44 ◼   ► in, you know, in the video, say, Siri, because they said that a lot in the skin.

01:17:49 ◼   ► Yes.

01:17:49 ◼   ► And this is actually there.

01:17:51 ◼   ► There was a similar trick done back when Amazon did a Super Bowl commercial for the Alexa service.

01:17:58 ◼   ► I believe it was in 2018 or 2019, something like that.

01:18:01 ◼   ► They did something similar with the audio of that commercial where when they when they said

01:18:07 ◼   ► their hail word to their assistant, they they also stripped out some frequencies to to fix

01:18:13 ◼   ► the same problem so that that way everyone's Amazon Echoes wouldn't wake up in response to that.

01:18:17 ◼   ► And they even and Amazon even blogged about it, about like the science behind why they were doing

01:18:21 ◼   ► this and how it how it like, you know, subtly defeated their recognition to serve this role.

01:18:26 ◼   ► So everyone said, since Apple's doing this, that must be why.

01:18:29 ◼   ► And then everyone else started saying, well, wait a minute, my phone and my wife's phone

01:18:34 ◼   ► and all these other phones in the room kept activating during the keynote.

01:18:37 ◼   ► And I tried.

01:18:39 ◼   ► The reason why I paid attention to this is that even even back when Amazon did their version

01:18:44 ◼   ► of this back, you know, a few years ago, I thought, wouldn't it be a great overcast feature

01:18:49 ◼   ► if I just detect like if you are playing out loud on a speaker or in a car, if I try to

01:18:56 ◼   ► detect when a podcast contains the word Siri or Alexa and apply that filter myself in overcast

01:19:04 ◼   ► as it's playing so that podcasts don't trigger all your devices accidentally, wouldn't that

01:19:09 ◼   ► be a cool feature?

01:19:10 ◼   ► So I tried recreating this effect, even just like first in an audio editor.

01:19:14 ◼   ► So I would record myself saying Siri or Alexa or whatever and giving a command and I would

01:19:19 ◼   ► play it over speakers and then I'd see my devices activate and then I would say, all

01:19:22 ◼   ► right, what filters in Adobe Audition or whatever do I need to apply to replicate this effect

01:19:27 ◼   ► so it defeats that recognition?

01:19:29 ◼   ► And back when Amazon did their Super Bowl commercial, I could not do it.

01:19:34 ◼   ► No matter what I tried, I could not get the recording of my voice to not trigger the speakers.

01:19:42 ◼   ► So what did I do when all this came out this week about Siri?

01:19:46 ◼   ► I tried the exact same thing and I went in and I was able to

01:19:48 ◼   ► able to compare because now I had the real audio from Apple.

01:19:51 ◼   ► So, you know, so I opened that real audio up and I saw exactly how they stripped out those

01:19:56 ◼   ► frequencies and exactly which ones were stripped out with what kind of response pattern and I

01:20:01 ◼   ► figured out exactly how to replicate like as close as possible to that process.

01:20:06 ◼   ► And I applied it to a test file and I still couldn't replicate the effect.

01:20:11 ◼   ► It would still wake up all my devices whenever I would play my voice coming over the speakers.

01:20:16 ◼   ► Now, what some people have theorized is that maybe by recognizing my voice, maybe I'm giving it such a strong signal because it's me saying it that my device would start to it.

01:20:27 ◼   ► But maybe like somebody else wouldn't be able to activate my devices as easily.

01:20:32 ◼   ► And that's certainly possible.

01:20:33 ◼   ► But also with everybody saying that all of their devices were being woken up by the keynote anyway, it seems like this is not necessarily a very strong filter that basically the recognition of the term Siri by all the modern Apple devices out there seems like it actually might be sophisticated enough and sensitive enough that maybe this trick of muting these few frequencies doesn't even defeat it necessarily.

01:21:00 ◼   ► Or it doesn't defeat it enough if the voice is close enough to you.

01:21:05 ◼   ► Or maybe if you haven't done like the personal voice training with the Siri setup process, maybe yours is more sensitive or less sensitive.

01:21:13 ◼   ► I don't know.

01:21:14 ◼   ► But there's a lot of variables and it's not it doesn't seem super clear cut that this process works 100% of the time.

01:21:20 ◼   ► Yeah, they should add some other way to mix it in there like some frequency you could add or something.

01:21:25 ◼   ► But I mean, the main problem is, as we've discussed every time we've mentioned audio, that waveform that you're looking at your computer.

01:21:30 ◼   ► screen doesn't magically transfer into your brain.

01:21:33 ◼   ► It has to go through a whole bunch of other stuff.

01:21:34 ◼   ► And guess what?

01:21:35 ◼   ► The frequency response of speakers is not perfect.

01:21:37 ◼   ► So, yeah, you could put out the slice out these narrow slices.

01:21:40 ◼   ► But by the time that makes it through your audio chain to your speakers, to the physical world, through the air to your ears, how much of that slicing out survives?

01:21:49 ◼   ► The finer you slice it, the more chance there is that just the quality of the amplification and the speaker cone wiggling around end up mushing some other frequencies into the range where you could notch them out.

01:22:03 ◼   ► Now there's stuff there and it's it's really complicated.

01:22:05 ◼   ► I mean, to be clear, like this particular slice out would be very difficult to do in any kind of analog way.

01:22:13 ◼   ► If you try to do it like with a parametric EQ, you won't be able to.

01:22:16 ◼   ► You need like an FFT filter with.

01:22:18 ◼   ► That's what I'm saying.

01:22:18 ◼   ► But it's so narrow that, like, you think, OK, well, now that I've cut that out, there'll be nothing in that frequency band when the sound hits my ear.

01:22:24 ◼   ► And it's like, well, is the signal chain from that audio file to your ears so perfect that it's tiny, narrow slice of frequency is going to have no energy in it whatsoever because your speakers are that perfect?

01:22:37 ◼   ► I, you know, especially if you're playing over a phone speaker or something, I'm not sure it's going to work.

01:22:41 ◼   ► So it doesn't surprise me.

01:22:42 ◼   ► What I'm saying is it doesn't surprise me that it is unreliable.

01:22:44 ◼   ► Yeah.

01:22:46 ◼   ► All right.

01:22:46 ◼   ► So coming back around, let's talk about what Amar had said during the same presentation.

01:22:53 ◼   ► Again, they are the vice president of AI.

01:22:55 ◼   ► We're super excited about our third generation of Apple Foundation Models, or AFM, in partnership with Google.

01:23:00 ◼   ► We've built a family of models spanning on device to the cloud.

01:23:03 ◼   ► AFM Core, Core Advanced Cloud, and Cloud Image are custom builds for Apple Silicon, trained using proprietary data and refined using techniques.

01:23:11 ◼   ► This was a potential transcription thing.

01:23:13 ◼   ► Yeah, I don't know.

01:23:14 ◼   ► I think the word they had for the transcription was different, so I looked at other sources, and another source had techniques.

01:23:19 ◼   ► But anyway, refined using something.

01:23:21 ◼   ► From Gemini Frontier Models.

01:23:23 ◼   ► And also, our friend Stephen Robles was there and did a video about all this, about a bunch of stuff, including this meeting.

01:23:31 ◼   ► We'll put a timestamp link in.

01:23:33 ◼   ► But what Stephen said was that he thinks refined is another way of saying that they distilled the Gemini models to train Apple Foundation Model models.

01:23:42 ◼   ► And what that means is basically you have some finished model that someone's already trained, and you don't want to have to go through all the effort of training your thing the same way.

01:23:52 ◼   ► So instead, you ask questions of that other finished model and use the answers to train your model.

01:23:57 ◼   ► That's distilling.

01:23:59 ◼   ► Lots of companies are, you know, everybody does it, apparently.

01:24:01 ◼   ► It's just an open secret.

01:24:02 ◼   ► You know, if your competitor comes out with a better model than you, you ask that model questions and use the answers to train your model so your model is as good as theirs.

01:24:10 ◼   ► It's not clear to me when they talk about all these models.

01:24:13 ◼   ► First, they give them all Apple names.

01:24:14 ◼   ► It's, you know, AFM, Apple Foundation Model, Core, Core Advanced, Core Advanced Cloud.

01:24:19 ◼   ► The names sound like old Intel processors, honestly.

01:24:21 ◼   ► But Apple gives them names.

01:24:24 ◼   ► And so they're created in partnership with Google, trained using proprietary data.

01:24:31 ◼   ► So it's not like a finished model from Google because they're training it.

01:24:34 ◼   ► And is the proprietary data Apple's proprietary data, Google's proprietary data?

01:24:39 ◼   ► And then refined using techniques from Gemini Frontier Models?

01:24:43 ◼   ► Are you distilling bigger?

01:24:44 ◼   ► Like, it's hard to tell what they're doing physically speaking.

01:24:48 ◼   ► Like, are they starting?

01:24:50 ◼   ► They're all called Apple Foundation Model.

01:24:52 ◼   ► So they were called that before the Google deal.

01:24:54 ◼   ► Google, Apple had Foundation Models.

01:24:55 ◼   ► This is AFM 3, but they dropped the 3 for most of the stuff, right?

01:24:58 ◼   ► So there was AFM 1 and 2 that Apple made all on its own and now they're doing 3.

01:25:02 ◼   ► Do they taking the models from AFM 2 and just using those models and then distilling against Google Gemini Models to train them?

01:25:08 ◼   ► Or are they taking Gemini Models and training them with Apple?

01:25:12 ◼   ► It's not clear.

01:25:12 ◼   ► But all I'm saying is that they were not exactly clear in, like, what is the origin of these models?

01:25:19 ◼   ► Were they made it Apple and refined with data from Google?

01:25:22 ◼   ► Were they made it Google and refined with data from Apple?

01:25:24 ◼   ► Like, clear as mud.

01:25:26 ◼   ► But they're basically saying this.

01:25:28 ◼   ► Both cooks have some participation into this meal.

01:25:32 ◼   ► We're just not sure what they do.

01:25:33 ◼   ► All righty.

01:25:34 ◼   ► Then some more details.

01:25:37 ◼   ► Oh, this little bit here.

01:25:39 ◼   ► I know the names are just like AFM Core or Cloud, blah, blah, blah.

01:25:42 ◼   ► AFM Cloud Pro, Apple says, this is Ammar talking again, is our most capable model with quality similar to Gemini Frontier Models.

01:25:52 ◼   ► Again, does that mean it is a Gemini Frontier Model?

01:25:56 ◼   ► Is that like, because when they say Frontier Model, they're saying, like, the best Google models that they have.

01:26:01 ◼   ► This is AFM Cloud Pro.

01:26:03 ◼   ► It runs only on servers.

01:26:04 ◼   ► It's presumably a big model.

01:26:06 ◼   ► It has an Apple-given name, AFM Cloud Pro.

01:26:09 ◼   ► And they say, this one is the same quality as the best models Google has.

01:26:13 ◼   ► Is it just those models?

01:26:16 ◼   ► Is it those models changed in some way?

01:26:18 ◼   ► Because we have those stories about how Apple is given access to these Gemini models and is allowed to tweak them and modify them.

01:26:24 ◼   ► Or is it an Apple model that is distilled against Gemini models?

01:26:28 ◼   ► Not entirely clear, but Apple's trying to characterize, when we say all this AFM stuff, here's how you should think about it.

01:26:35 ◼   ► And AFM Cloud Pro is basically their, like, their, you know, ChatGPT 5.5, their, whatever.

01:26:42 ◼   ► Like, it's their biggest thing.

01:26:43 ◼   ► All right, then some more details with regard to this same tech talk from above.

01:26:48 ◼   ► John, can you go through some definitions for us to set the foundation, though, please?

01:26:51 ◼   ► Yeah, I had to look this up because they just started talking about them in the talk.

01:26:55 ◼   ► And we're going to start reading from an Apple press release, which makes more sense in light of this talk or give some background for it.

01:27:00 ◼   ► So the first term is dense model.

01:27:02 ◼   ► And that, the definition of that that I found is that every parameter in the network is activated for every input.

01:27:09 ◼   ► For example, if a model has 70 billion parameters, all 70 billion, is that correct from that?

01:27:15 ◼   ► I might have miscopy-pasted.

01:27:16 ◼   ► All 70 billion process each token.

01:27:18 ◼   ► That is correct.

01:27:20 ◼   ► So, yeah, you've got the models are measured in a number of parameters.

01:27:23 ◼   ► And if it is a dense model, anytime you give it any request, all 70 billion parameters participate in every single token that it produces.

01:27:31 ◼   ► Tokens are sequences of words.

01:27:33 ◼   ► Again, I should find those links to the YouTube videos that explain how LLMs work.

01:27:38 ◼   ► But it is a good idea to know how they work.

01:27:39 ◼   ► And that's a dense model.

01:27:41 ◼   ► Which may be how you think all models work, but that's not, you know, that was just like the sort of the most naive version of it.

01:27:46 ◼   ► Then you have sparse models or mixture of experts or MOE.

01:27:50 ◼   ► The model is divided into many expert subnetworks.

01:27:54 ◼   ► A small router network decides which experts are relevant for each token.

01:27:59 ◼   ► And only those experts are activated.

01:28:01 ◼   ► So that's where, you know, for each token that's going through, all 70 billion parameters don't participate in the processing of that token.

01:28:08 ◼   ► Only the ones that another network decides are relevant.

01:28:11 ◼   ► So you don't have to activate all of them at once.

01:28:13 ◼   ► So that's dense model versus sparse model.

01:28:15 ◼   ► And you'll hear those terms in this big Apple press release about their third generation foundation models.

01:28:21 ◼   ► All right, so with that in mind, Apple's announcement of its third generation foundation models happened over at machinelearning.apple.com with, you know, we'll put the link in the show notes.

01:28:30 ◼   ► With regard to the on-device models, AFM Core is a 3 billion parameter dense model.

01:28:36 ◼   ► So that's the one where if you ask it anything, it's using all 3 billion parameters.

01:28:40 ◼   ► AFM Core Advanced is a 20 billion parameter sparse model and our most powerful on-device model.

01:28:47 ◼   ► It's natively multimodal, enabling helpful features like expressive voices and higher accuracy dictation.

01:28:53 ◼   ► Built on cutting-edge Apple research, this model activates just 1 to 4 billion parameters at a time, depending on the request.

01:28:59 ◼   ► AFM 3 Core Advanced is unlocked and optimized for our most capable Apple Silicon systems.

01:29:05 ◼   ► So this is the model that you only get to run on whatever it is.

01:29:08 ◼   ► M3 or better.

01:29:09 ◼   ► iPhone 17 Pro or better.

01:29:11 ◼   ► iPad M4.

01:29:13 ◼   ► That's AFM Core.

01:29:15 ◼   ► AFM 3 Core Advanced.

01:29:16 ◼   ► I took out the threes.

01:29:17 ◼   ► So just compare.

01:29:18 ◼   ► AFM Core is their 3 billion parameter dense model, which is like all the parameters all the time.

01:29:24 ◼   ► It runs on lower devices.

01:29:26 ◼   ► And AFM Core Advanced is 20 billion, but it's sparse.

01:29:29 ◼   ► So that's a big jump.

01:29:31 ◼   ► And you can kind of understand why only Apple's top-end devices can handle it.

01:29:34 ◼   ► Even though all 20 billion aren't activated once, it's 1 to 4 billion, depending on the request.

01:29:39 ◼   ► But it's a much more sophisticated model.

01:29:40 ◼   ► So that's the difference they were talking about in those slides, about how you can't get the expressive voice and stuff.

01:29:47 ◼   ► That's all AFM Core Advanced.

01:29:49 ◼   ► Righto.

01:29:50 ◼   ► And then server-based models running on private cloud compute.

01:29:54 ◼   ► There's AFM Cloud, AFM Cloud Image, and AFM Cloud Pro.

01:29:57 ◼   ► AFM Cloud is optimized for speed, efficiency, and performance.

01:30:00 ◼   ► AFM Cloud Image is for image generation and editing.

01:30:04 ◼   ► Go figure.

01:30:04 ◼   ► And AFM Cloud Pro powers our most demanding use cases like agentic tool use and complex reasoning.

01:30:10 ◼   ► Yeah.

01:30:10 ◼   ► So Apple, if you look at their talk and all the description of this, like, they don't come out and say this.

01:30:16 ◼   ► But every part of their architecture, and I imagine most things like this, is trying as hard as it can to use the smallest possible model it can get away with.

01:30:28 ◼   ► This is as opposed to, for example, me, whenever I use any of these products.

01:30:31 ◼   ► I'm using, like, you know, codex from the command line or something.

01:30:34 ◼   ► And it has, like, a slash models thing where you can pick which model you want.

01:30:37 ◼   ► Whatever the biggest is, I always pick biggest, highest effort, most thinking.

01:30:44 ◼   ► Because I'm like, I want the smartest one.

01:30:46 ◼   ► I'm not going to.

01:30:47 ◼   ► It's like, well, it would be much faster if you use a smaller one.

01:30:49 ◼   ► It's a waste of time to use the big one on these little dinky tasks.

01:30:52 ◼   ► I'm like, I just want the smartest all the time, which means I run out of tokens a lot.

01:30:55 ◼   ► Which is fine.

01:30:56 ◼   ► That's my own stupid choice.

01:30:57 ◼   ► That is not what Apple wants billions of iPhone users to do.

01:31:01 ◼   ► So what they're going to do is you make a request and it's going to be like, if we can get away with doing this with AFM Core, our 3 billion parameter dense model, we're going to.

01:31:09 ◼   ► Even though if we gave it to AFM Cloud Pro, it would do a way better job.

01:31:14 ◼   ► We're not going to do that because computing is a scarce resource.

01:31:19 ◼   ► It costs money.

01:31:20 ◼   ► It takes time.

01:31:21 ◼   ► And even though it probably would give a better result, although that is somewhat debatable because sometimes they say, well, if you give it to one of these big thinking models, not only does it take longer, but actually gives a worse result because it overthinks it and blah, blah, blah.

01:31:31 ◼   ► But setting that aside, my experience has been the bigger, more powerful models do better no matter what you ask them.

01:31:38 ◼   ► I'm sure there are counterexamples, but my experience with code stuff is they do better.

01:31:43 ◼   ► But that's not what Apple software wants to do.

01:31:45 ◼   ► So it's going to figure out what is the wimpiest model that I can send this to.

01:31:49 ◼   ► If none of the on-device ones will do it, then I have to send it to the cloud.

01:31:52 ◼   ► But even when I send it to the cloud, do we really want to send it to the one, the AFM Cloud Pro, the one we said that is comparable to Google stuff?

01:32:00 ◼   ► That is like the line of last resort.

01:32:02 ◼   ► But when I'm sitting there talking to like, you know, you know, ChatGPT or whatever in a web browser, I'm at least picking the pop-up menus and say, yeah, go to ChatGPT, super hard thinking, maximum, whatever, all the time.

01:32:16 ◼   ► No matter what I'm asking it, Apple does not give you that choice.

01:32:19 ◼   ► They only have one model, AFM Cloud Pro, that is on that level.

01:32:23 ◼   ► And I bet it gets very few requests by design.

01:32:26 ◼   ► The only other thing is like if you do anything with images, they just have one image model.

01:32:29 ◼   ► So everything there is going to AFM Cloud Image, but everything else, they have so many choices before they get down to the quote-unquote good model underneath it all.

01:32:38 ◼   ► And I think that may come up when we get to another Dan Morin example.

01:32:42 ◼   ► Maybe we skipped over it or maybe I removed it from the notes.

01:32:45 ◼   ► But sometimes, yeah, maybe I get to get clipped out of the notes here.

01:32:49 ◼   ► Sometimes when you ask Siri AI in the current betas a question and it gives you an answer.

01:32:55 ◼   ► Actually, that's from my experience, not Dan Morin, sorry.

01:32:59 ◼   ► Casey, my goodness.

01:33:01 ◼   ► Even you confuse you and Dan Morin.

01:33:03 ◼   ► Yeah, yeah.

01:33:03 ◼   ► I don't know where it went in the notes, though.

01:33:05 ◼   ► Anyway, sometimes when you ask a question and you get an answer, as I did with Siri AI and macOS, you'd be like, oh, that's a pretty good answer.

01:33:13 ◼   ► And then when you ask the same question later to, for example, demonstrate it to a family member, it gives you a totally different, worse answer.

01:33:20 ◼   ► And when I see that, I'm like, did you go to AFM Cloud Pro the first time?

01:33:24 ◼   ► But then the second time I ran it, you did like a local on-device model and gave me a crap answer because that first answer was so much better.

01:33:31 ◼   ► So, yeah, Apple doesn't seem to give you control over that routing.

01:33:35 ◼   ► I don't think there's a pop-up menu in Siri where you can pick the model that you want.

01:33:39 ◼   ► And sometimes that difference shows through.

01:33:41 ◼   ► Maybe if you're not, like, experimenting like I am and asking the same question multiple times with the exact same wording, you won't notice that.

01:33:47 ◼   ► But results may vary.

01:33:50 ◼   ► Traditional large language models, whether dense or sparsely activated, require all weights to reside in active memory or DRAM.

01:33:56 ◼   ► To break this barrier, AFM Core Advanced introduces a novel sparsely activated architecture built on instruction-following pruning, or IFP, a technique developed by Apple researchers.

01:34:07 ◼   ► And, in fact, there's instruction-following pruning of large language models for large language models from June 2025.

01:34:14 ◼   ► That's also at machinelearning.apple.com.

01:34:16 ◼   ► I forget if we ever talked about that paper, that Apple said it had a bunch of papers on this topic.

01:34:21 ◼   ► But here is the fruits of that labor.

01:34:23 ◼   ► Apparently, AFM Core Advanced does, in fact, this thing.

01:34:26 ◼   ► And this is this paper, yeah, from June, from basically a year ago.

01:34:29 ◼   ► All right.

01:34:30 ◼   ► Then there's also LLM in a Flash, Efficient Large Language Model Inference with Limited Memory, again, on machinelearning.apple.com.

01:34:38 ◼   ► Filippone Stefano wrote an article about it, which we'll also link from Filippone.

01:34:43 ◼   ► Finally, the LLM in a Flash paper addresses challenges and solutions for running LLMs on devices with limited DRAM capacity.

01:34:49 ◼   ► It presents an approach for efficiently executing LLMs that exceed available DRAM capacity by storing model parameters in flash memory and loading them into DRAM on demand.

01:34:58 ◼   ► Yeah, I think we did also mention that paper ages ago.

01:35:01 ◼   ► Apple explicitly calls out the instruction following pruning for large language models sparse things.

01:35:06 ◼   ► So they're definitely doing that because that was reading from a direct quote from their VP of AI or whatever.

01:35:12 ◼   ► What is his name?

01:35:13 ◼   ► Amar?

01:35:15 ◼   ► Yeah, yeah, yeah.

01:35:16 ◼   ► Yes.

01:35:17 ◼   ► And there's other people about flash memory, about like, oh, you don't have enough RAM, but you want to use a bigger model.

01:35:23 ◼   ► You can store part of it in a flash and like swap it in or whatever.

01:35:25 ◼   ► That makes me wonder if one of the reasons Apple intelligence is disabled on external drives is that Apple just doesn't want to deal with having to guess at the performance characteristics of external drives.

01:35:36 ◼   ► If they're doing if they are, in fact, doing this technique, this LLM in a flash thing where some of the model is in RAM and some of it is on flash.

01:35:44 ◼   ► Apple knows the character, the speed characteristics of all of its internal SSDs.

01:35:47 ◼   ► And so maybe it's one that's an external drive.

01:35:51 ◼   ► They're like, how about we just disable entirely?

01:35:52 ◼   ► Because I'm sure plenty of the drives are fast enough.

01:35:54 ◼   ► But if they're not, it has a terrible experience or like a performance falls off a cliff.

01:35:59 ◼   ► I'm just speculating.

01:36:00 ◼   ► I have no idea if this is true.

01:36:01 ◼   ► If someone knows for sure, please write in and tell me why.

01:36:04 ◼   ► Why can't I boot from an external drive and use Apple intelligence?

01:36:07 ◼   ► But you can't.

01:36:08 ◼   ► And maybe this is one of the reasons.

01:36:09 ◼   ► Or it could be that this LLM in a flash thing isn't even used in any of the 27 OSs.

01:36:12 ◼   ► I don't know that either, but I just thought I'd throw that out there.

01:36:15 ◼   ► All right.

01:36:15 ◼   ► With that in mind, coming back to Apple's Foundation Models announcement, instead of forcing the entire model into DRAM, the full model is stored in flash memory or NAND.

01:36:23 ◼   ► Because NAND to DRAM bandwidth is too slow to swap weights token by token as standard MOE or mixture of experts models require, AFM3 Core Advanced makes routing decisions per prompt.

01:36:34 ◼   ► A lightweight, dense block selects a fixed set of experts during initial processing, periodically reselecting them during generation.

01:36:40 ◼   ► To minimize data movement, the model relies on a high percentage of always active, shared experts, alongside input-dependent, routed experts, swapped into DRAM only when needed.

01:36:50 ◼   ► This design also introduces crucial inference time elasticity.

01:36:55 ◼   ► Rather than using a single model for all tasks or managing an ensemble of smaller models, AFM Core Advanced uses a predetermined number of active parameters tailored to each specific use case.

01:37:05 ◼   ► This allows weights to be loaded incrementally across requests of varying difficulty, scaling the model size for beyond traditional DRAM limits while minimizing latency.

01:37:15 ◼   ► So this is them saying that this specific model, AFM3 Core Advanced, does use the things from the flash paper, it sounds like.

01:37:22 ◼   ► But of course, Apple Intelligence hasn't been able to run when booted from an external drive since its introduction in 2024.

01:37:28 ◼   ► So this is, and also the AFM Core Advanced only runs on the highest-end hardware.

01:37:33 ◼   ► So that can't be the root reason why they do this, but I wonder if it contributes to it.

01:37:37 ◼   ► But anyway, all of that is to say, if you think Apple just, like, bought models from Google, stuck them on servers, and sent requests to them, that's not how this works at all.

01:37:45 ◼   ► Like, I feel like it's the biggest point of this tech talk is with all those things on the diagram and Craig Fadery pointing out, like, you know, Google's participation, they were an important part of this.

01:37:55 ◼   ► And we'll get to some even more important parts of them in a second.

01:37:57 ◼   ► But Apple is the one writing the software that's on top of it.

01:38:01 ◼   ► And Apple, in fact, has made many advances, all those AI people that Apple hired that, you know, eventually left to go to other AI companies, they actually did interesting novel work.

01:38:10 ◼   ► I'm not sure if this stuff is, like, the equivalent things are happening at all the cutting-edge AI companies, but I do want to give Apple credit for, like, they're not just, like, we failed, we can't do anything, let's just use a third-party product and slap a Siri face on it.

01:38:23 ◼   ► Not at all.

01:38:24 ◼   ► Like, these papers and the fact that they're using them and the description of how they use them and how it gets them to be able to run models that otherwise wouldn't run on device.

01:38:33 ◼   ► And essentially how it makes it possible for Apple to ship 27 OSs to literally billions of Apple devices and not destroy any, all their servers.

01:38:43 ◼   ► Like, they can't do the thing that everyone else is doing, which is, like, every request goes to a server, don't run anything on device.

01:38:48 ◼   ► They're trying so hard to run everything on device that they possibly can, using lots of interesting techniques to make models that shouldn't fit on your phone actually fit on your phone, and shuffling between them.

01:38:59 ◼   ► And, like, it's fascinating, and it shows that they are not, I mean, they're behind, like, because they're not the cutting edge like their competitors are.

01:39:06 ◼   ► But they're using what they're good at, basically writing client-side software, to innovate in that area on top of the underlying technologies that makes the models themselves and trains them and stuff like that.

01:39:18 ◼   ► Then coming back to Amar, to bring this model to production, we work with both Google and NVIDIA to extend our private cloud compute infrastructure to NVIDIA GPUs in Google's cloud while maintaining Apple's unmatched privacy guarantees.

01:39:33 ◼   ► Andrew Cunningham over at ArsTechnica writes,

01:39:35 ◼   ► To do this while still making the same privacy promises, Apple's new iteration of private cloud compute is using NVIDIA's confidential computing, Intel's trust domain extensions, and Google's Titan security chip to provide layers of protection similar to what Apple provides for its own servers.

01:39:49 ◼   ► To provide additional protection, Apple keeps a cryptographically verifiable append-only ledger of all Google Cloud hardware that is part of the PCC fleet.

01:39:57 ◼   ► And Apple's devices will only trust hardware, excuse me, software on these servers that is signed by Apple.

01:40:03 ◼   ► The Google Cloud servers don't yet support all the same protections as Apple's own private cloud compute servers, but Apple says it will gradually be ramping towards the complete set of protections throughout the summer preview period.

01:40:15 ◼   ► So, what does a cryptographically verifiable append-only ledger sound like to you?

01:40:22 ◼   ► Yeah, it's not blockchain.

01:40:24 ◼   ► It sounds like it.

01:40:25 ◼   ► Isn't that exactly what a blockchain is, right?

01:40:27 ◼   ► Don't say blockchain.

01:40:28 ◼   ► Yeah.

01:40:30 ◼   ► I mean, all this stuff, as has been pointed out when we talked about private cloud compute, like, Apple is doing everything that it possibly can to mathematically show and prove and guarantee that you are talking to servers that behave in the way that Apple says.

01:40:48 ◼   ► And that way is, Apple can't see it because it's all end-to-end encrypted, so Apple can't see it.

01:40:53 ◼   ► And obviously, the server that receives a request has to, of course, decrypt it because how could it, you know, it has to know what you said to do the work.

01:41:00 ◼   ► But the other guarantee that Apple gives is the software that runs on those servers will never save any of that.

01:41:08 ◼   ► It doesn't even log stuff.

01:41:09 ◼   ► Like, it doesn't, like, no data that comes into that server that is decrypted in memory and then processed and chucked back out is saved in any way.

01:41:16 ◼   ► And so, that's private cloud computing.

01:41:18 ◼   ► It was 2024 where they talked about that.

01:41:19 ◼   ► They gave a big paper on it and everything.

01:41:20 ◼   ► And that's part of the promise is we, our software behaves in this way.

01:41:25 ◼   ► We can't see your data.

01:41:26 ◼   ► We don't save your data.

01:41:27 ◼   ► We can never see it.

01:41:29 ◼   ► If anyone asks us to get it, we don't have it.

01:41:31 ◼   ► We never kept it.

01:41:32 ◼   ► It is transiently there and disappears.

01:41:34 ◼   ► And that entire time, it's encrypted and kept in this hardware.

01:41:36 ◼   ► And then we don't save anything about it.

01:41:38 ◼   ► It's thrown away.

01:41:39 ◼   ► Apple has its own servers with those, you know, M3 Ultra, M2 Ultras, whatever they were, that we saw being made in that factory.

01:41:46 ◼   ► Runs Apple Silicon, runs their models.

01:41:48 ◼   ► They've had that for, I guess, since 2024.

01:41:51 ◼   ► That's private cloud computing.

01:41:52 ◼   ► They just continued to call everything on their servers private cloud compute.

01:41:58 ◼   ► So we're wondering about this several episodes ago.

01:41:59 ◼   ► What would they do?

01:42:00 ◼   ► They're just calling it private cloud compute.

01:42:02 ◼   ► And what they mean by that is those M2 Ultra Apple Silicon servers that Apple made.

01:42:06 ◼   ► And also, NVIDIA GPUs running in Google's cloud.

01:42:11 ◼   ► Because as we saw, Google has a similar architecture to PCC.

01:42:15 ◼   ► And those things that Casey listed there, if you go to the Andrew Cunningham article at Ars Technica that will be in the show notes, those are links.

01:42:23 ◼   ► So if you want to learn what is NVIDIA's confidential computing, what is Intel's trust domain extensions, what is Google's Titan security chip?

01:42:30 ◼   ► Those are the pieces of the puzzle that are building up towards making Google able to run servers that Apple calls private cloud computing, even though they share essentially nothing with Apple's private cloud computing other than the promises that they fulfill.

01:42:46 ◼   ► And as this article says, that they're not quite up to the standards of Apple's private cloud compute.

01:42:53 ◼   ► But the idea is they will be before the 27 OS's ship.

01:42:57 ◼   ► If they're not, how will we know?

01:42:58 ◼   ► Like, again, Apple gives these images to security researchers so they can mathematically prove that when you run a request, it's really running against this image.

01:43:09 ◼   ► And security researchers have at it.

01:43:11 ◼   ► Prove to yourself that we're not saving any data.

01:43:12 ◼   ► There's no secret NSA side channel that sends every request in plain text to some server or logs it.

01:43:18 ◼   ► Like, you have the binary image.

01:43:20 ◼   ► This is it.

01:43:21 ◼   ► Prove to yourself.

01:43:22 ◼   ► Because the executable binary is, you know, it's just machine code.

01:43:25 ◼   ► You can see what it does.

01:43:26 ◼   ► You can decompile it and you security researchers should be able to prove, see, we're not saving it anymore.

01:43:31 ◼   ► And then also what you should be able to prove is that thing that we gave you, that's the thing that's running on the server.

01:43:36 ◼   ► And all the stories I've said about this is that in the end, you have to, there is a root of trust in believing that Apple is not lying to you.

01:43:42 ◼   ► That's true of every piece of software in the world.

01:43:45 ◼   ► Lots of times people send an ask ADP questions.

01:43:47 ◼   ► Like they say, Apple says my message is encrypted, but do we have to just take their word for it?

01:43:52 ◼   ► The answer is yes.

01:43:53 ◼   ► Like, you know, like they write the software.

01:43:56 ◼   ► They could lie and say it's end to end encrypted and just be lying through their teeth.

01:43:59 ◼   ► But security researchers would discover that they were lying in most cases.

01:44:04 ◼   ► In the end, there is some root of trust.

01:44:06 ◼   ► You know, if you keep digging down and like, what is that paper that we've linked to a few times?

01:44:13 ◼   ► Musing on trust or whatever, something like that, about talking about if you had a compiler that was corrupted in some way that you couldn't trust anything because the compiler builds all your other programs.

01:44:23 ◼   ► I wish I could remember what that one.

01:44:24 ◼   ► Yeah, I know what you're thinking of.

01:44:25 ◼   ► Find it for the notes, maybe.

01:44:27 ◼   ► So, yeah, but you do have to trust it when they say we want some software to do this.

01:44:31 ◼   ► And also you have to trust that it doesn't have bugs, which sometimes it does.

01:44:33 ◼   ► It's supposed to do this, but it doesn't actually do this as a bug in it somewhere.

01:44:38 ◼   ► But, yeah, it seems like, I mean, the question is, those Apple Silicon servers with the M2 Ultras in them, how long will Apple keep doing that?

01:44:46 ◼   ► Have they already given up on that effort and said, well, it's a cool thing.

01:44:49 ◼   ► We tried it, but that was under the old regime.

01:44:50 ◼   ► And now what we're going to do is what everybody else does, which is we're going to run NVIDIA GPUs, which is interesting because Apple hates NVIDIA's guts and hasn't done anything with NVIDIA for ages.

01:44:59 ◼   ► But apparently they'll use their GPUs and servers.

01:45:01 ◼   ► And it's also interesting because Google, which is running these servers, Google has their own like TPUs.

01:45:07 ◼   ► There's tensor processing units that are really good at running Google's models.

01:45:11 ◼   ► Google makes its own silicon, these TPUs, we've talked about a long time ago on the show, that are different than NVIDIA GPUs.

01:45:20 ◼   ► Why isn't Apple using those?

01:45:21 ◼   ► Well, if Apple uses NVIDIA GPUs, they're not tied to Google.

01:45:26 ◼   ► So if and when the deal ends with Google or they don't like Google hosting their stuff, they could go to anybody and say, hey, do you have a data center filled with NVIDIA GPUs that we can run our stuff on?

01:45:34 ◼   ► They probably do.

01:45:36 ◼   ► Only Google has TPUs, I think.

01:45:38 ◼   ► I don't think there's any sort of third party thing for that.

01:45:41 ◼   ► Now, I'm not sure how much stake to put in that because if I was sending these things to Google, I would just run on their TPUs.

01:45:46 ◼   ► And if we have to go elsewhere, then just like port it or whatever.

01:45:48 ◼   ► But maybe that's more difficult.

01:45:49 ◼   ► But anyway, seems like Apple might be giving up on the idea that they are going to run Apple silicon servers in their own data centers and build their own hardware and whatever that factory was in Arizona or in Texas.

01:46:03 ◼   ► And instead, they're just going to do what everyone else does and just pay some hosting provider like AWS or Google Cloud to rent racks full of NVIDIA GPUs.

01:46:11 ◼   ► And at this point, Google did a deal with XAI, Grok, whatever.

01:46:17 ◼   ► They built out data centers that they aren't using because no one wants to use Grok because it sucks and everyone hates Elon Musk.

01:46:24 ◼   ► And so XAI is renting whole data centers to Google.

01:46:29 ◼   ► So it could be that when you talk to Siri AI, it ends up going to an XAI data center that is being rented by Google where it's running NVIDIA GPUs.

01:46:37 ◼   ► Gracious.

01:46:38 ◼   ► What a world that we live in.

01:46:41 ◼   ► But yeah, this seems like a change in Apple's stance.

01:46:46 ◼   ► And again, notably the first time that I'm aware of since back when NVIDIA had a bum GPU in like an iBook or something that like the relationship between Apple and NVIDIA soured.

01:46:56 ◼   ► I believe it was a MacBook Pro.

01:46:58 ◼   ► Was it?

01:46:59 ◼   ► Yeah.

01:46:59 ◼   ► And like that was so long ago, like you would say so long ago, surely none of those people still there, but Phil Schiller is still there.

01:47:04 ◼   ► But yeah, they're still sore about that.

01:47:07 ◼   ► But this is one degree separated, I guess.

01:47:08 ◼   ► Like the deal is that Google will run servers for them because Google is good at running servers.

01:47:13 ◼   ► And I'm not sure if Apple dictated that.

01:47:15 ◼   ► And by the way, the servers you run for us, they should use NVIDIA GPUs and not your Tensor things, not your TPUs.

01:47:20 ◼   ► But that's the shape of it.

01:47:22 ◼   ► Apple and NVIDIA together again, sort of.

01:47:25 ◼   ► Sitting in a tree.

01:47:28 ◼   ► C-O-M-P-U-T-I-N-G.

01:47:29 ◼   ► Sitting in one of Elon Musk's data centers.

01:47:31 ◼   ► You're going to spell out inference?

01:47:33 ◼   ► I don't know.

01:47:33 ◼   ► Yeah.

01:47:35 ◼   ► So again, this is not, you know, getting back to video podcast.

01:47:38 ◼   ► It's not a visual medium and we don't have good images.

01:47:40 ◼   ► The tech radar images are blurry because it's like taken from a phone and an audience or whatever.

01:47:44 ◼   ► But please do look at the block diagrams.

01:47:46 ◼   ► And I think they did at one point, they like highlighted things in a particular color.

01:47:51 ◼   ► I think it was all shades of blue.

01:47:53 ◼   ► So it was tough to see on the screen.

01:47:54 ◼   ► Not a great choice.

01:47:55 ◼   ► But basically they said all the things in blue are Apple and all the things in this different shade of blue are Google.

01:48:02 ◼   ► And like everything was Apple.

01:48:04 ◼   ► The only things that were Google were like, you know, tinted shaded stuff in the cloud thing.

01:48:09 ◼   ► So they weren't forthcoming with exactly like is it a Google model that we change as an Apple model that we just still like they didn't go into that level of detail.

01:48:16 ◼   ► But their big emphasis here was most of the stuff you see in Siri AI is software Apple wrote.

01:48:23 ◼   ► And they didn't say this, but like it's clear that you kind of need the part that Google provided or otherwise none of this works.

01:48:30 ◼   ► But the Apple was really emphasizing we did a lot of work for this.

01:48:34 ◼   ► And I mean, I guess it's like also when it falls on its face and doesn't work right or you aren't happy with how it goes.

01:48:39 ◼   ► Don't blame Google because that wasn't up to them.

01:48:40 ◼   ► That was up to us.

01:48:42 ◼   ► All right.

01:48:43 ◼   ► Thank you to our sponsors this episode, Quince and Lisa.

01:48:46 ◼   ► And thanks to our members who support us directly.

01:48:49 ◼   ► You can join us at ATP.fm slash join.

01:48:52 ◼   ► One of the many perks of membership is ATP Overtime, our weekly bonus topic.

01:48:57 ◼   ► This week on Overtime, we're going to be talking about, by popular request, the WBDC State of the Union session.

01:49:03 ◼   ► There was a whole bunch of stuff covered there.

01:49:05 ◼   ► We're going to go over that in Overtime because it just couldn't fit in the show.

01:49:08 ◼   ► You can join and listen to ATP.fm slash join.

01:49:10 ◼   ► Thanks, everybody, and we'll talk to you next week.

01:49:39 ◼   ► We'll be right back.

01:50:09 ◼   ► By the way, if you're not an ATP member and you're like, oh, you're going to put State of the Union in Overtime?

01:50:20 ◼   ► Non-members hate when we put stuff in Overtime because they don't get to hear Overtime.

01:50:23 ◼   ► And like, I get it, you know, but like, here's the thing.

01:50:26 ◼   ► I'm pretty sure in multiple past years, possibly also including last year, we didn't cover State of the Union at all.

01:50:33 ◼   ► Because what happens after WWC is there's tons of follow-up and tons of news and things happen and State of the Union just gets pushed off and pushed off.

01:50:39 ◼   ► And then we look up and it's a month later and it's like, I guess we'll just delete State of the Union from the notes because, like, it's old news now.

01:50:45 ◼   ► Like, we can't really, like, go back to it.

01:50:47 ◼   ► It's just, it's already done and gone.

01:50:49 ◼   ► That's what Overtime is for.

01:50:50 ◼   ► That's why State of the Union is it.

01:50:52 ◼   ► So don't be sorry about it.

01:50:53 ◼   ► I mean, we're sorry.

01:50:54 ◼   ► Like, you know, and it is developer-y type stuff.

01:50:56 ◼   ► And it sounds like, how can State of the Union be in Overtime?

01:50:57 ◼   ► That should be part of a regular episode.

01:50:59 ◼   ► When we didn't cover it at all, people didn't notice.

01:51:02 ◼   ► But now it's in overtime.

01:51:03 ◼   ► Somebody will complain.

01:51:04 ◼   ► So anyway, I will say that the solution to this is to become a member.

01:51:06 ◼   ► Can't confirm.

01:51:09 ◼   ► All right.

01:51:10 ◼   ► I wanted to spend just a couple of minutes talking about trip results.

01:51:14 ◼   ► So I just went to Cape Charles for the last week.

01:51:16 ◼   ► This is our happy place on the eastern shore of Virginia.

01:51:20 ◼   ► And I brought a truly asinine amount of equipment and computing-related things.

01:51:27 ◼   ► In no small part because I had to record this very show, but also because I'm me.

01:51:32 ◼   ► And I wanted to talk about the Unified Travel Router, which we talked about at some point in the past.

01:51:39 ◼   ► But to refresh your memory, the standard operating procedure for travel routers is a GLINET,

01:51:45 ◼   ► which I don't know how to verbally describe how big a GLINET router is.

01:51:50 ◼   ► But they're small but not tiny by any stretch.

01:51:53 ◼   ► And a few months ago, Ubiquiti came out with a Unified Travel Router, which is exceedingly small.

01:52:00 ◼   ► It's imagine like five or six credit cards stacked on top of each other.

01:52:03 ◼   ► That's probably not exactly right, but that's kind of what I'm talking about.

01:52:06 ◼   ► And what a travel router does is if you have a single internet connection, which presumably you do,

01:52:13 ◼   ► but you would like to broadcast that to your entire family's constellation of devices,

01:52:19 ◼   ► then what you can do is you can use a travel router, be that a GLINET or the Unified Travel Router,

01:52:24 ◼   ► to log into or connect to whatever the internet source is.

01:52:28 ◼   ► So in a best-case scenario, you plug Ethernet right into this little baby travel router.

01:52:31 ◼   ► But more realistically, you are in a hotel or something like that,

01:52:35 ◼   ► And you have the travel router log into the hotel Wi-Fi and go through that whole painful dance.

01:52:43 ◼   ► But then the travel router broadcasts its own Wi-Fi that your phone and iPad and Mac and your spouse's phone and Mac and iPad

01:52:54 ◼   ► and your children's iPads and switches and so on and so forth, they all connect to the UTR, the Unified Travel Router.

01:53:00 ◼   ► And that is figuring out how to get you to the internet.

01:53:03 ◼   ► Now, the pro move, in my personal opinion, which you can do either by hand or the Unified Travel Router does automatically,

01:53:09 ◼   ► is to set your Wi-Fi that this portable router is broadcasting to be the exact same SSID and password as your home Wi-Fi.

01:53:19 ◼   ► So this way, everything just jumps on the nearby Wi-Fi, thinking effectively that it's at home.

01:53:23 ◼   ► It's particularly critical if you're a dork like me and bring one or maybe two Sonos speakers with you when you are on a long trip like this,

01:53:31 ◼   ► because then the Sonos speakers say,

01:53:33 ◼   ► Ah, yes, I'm at home.

01:53:34 ◼   ► And even though some of my friends are not here, I'm at home, so I will work just the way I always do, which is great.

01:53:40 ◼   ► This kind of reminds me of, I don't know, there's lots of animated shows that have done this,

01:53:45 ◼   ► but like some animated character or person who has to be in water, like maybe it's a mermaid or something,

01:53:49 ◼   ► or they dry out, that wherever they go, they bring like a fishbowl or a bowl of water or whatever with them.

01:53:55 ◼   ► And it just occurs to me that, Casey, wherever you go, you have to bring speakers with you,

01:53:58 ◼   ► because you wither and die without constant music playing.

01:54:00 ◼   ► Correct.

01:54:01 ◼   ► That is pretty much correct.

01:54:03 ◼   ► I really want to argue with you, but I can't.

01:54:06 ◼   ► You're the guy in Ponyo with the sprinkler thing, and he's pumping the water as he walks around because he's got a lot of water.

01:54:11 ◼   ► There you go.