PodSearch

Upgrade

517: Maybe It's Catchier in Japanese

 

00:00:00   [Music]

00:00:09   From Relay FM, this is Upgrade, Episode 517 for June 17, 2024. Today's show is brought to you by DeleteMe, Tailscale, and Chaos.es. My name is Mike Hurley and I'm joined by Surfin' Snell, Jason Snell.

00:00:26   Hey, Luce, dude. It's good to be here on the Summer of Fun!

00:00:33   Summer of Fun! I had a thought in the shower last night. People might not know what this is!

00:00:38   Oh, no.

00:00:39   I just thought to myself, I'm sure we pick up new listeners all the time. Welcome if you're a new listener.

00:00:44   Of course, hello.

00:00:45   What is the Summer of Fun and why do we scream so much during this time of the year? Well, it's called I don't know.

00:00:50   Many years ago, we realized that between WWDC and the iPhone event in September, there is a strong possibility for news to be slow.

00:01:00   This hasn't happened every year, especially the 2020. There was no Summer of Fun. Not because of COVID, like it wasn't fun, but there was just so much news during that time, which is because everybody was not taking vacations.

00:01:15   I always describe WWDC as being the start of Apple's calendar for the year, right? Because it's the updates for the next year.

00:01:21   And that also is the start of sort of our cycle. And then what I say is there's this burst of activity for WWDC.

00:01:27   And then there is the, I usually say long, slow cruise to the iPhone event.

00:01:34   And it's not that things aren't happening, right? Because there are betas.

00:01:37   It's going to be the summer of betas. As it always is. But there's not a lot that's new going on because everything is happening kind of under the surface.

00:01:50   And we're kind of just cruising through summer until the big iPhone event in terms of the Apple calendar.

00:01:56   So it's often very quiet. And so we've created a concept for our summer episodes where we try to do fun things and weird episodes and off-brand things and at least segments that are different to keep all of us entertained.

00:02:12   And also in the Northern Hemisphere, at least, it is the summer. I hope people in the Southern Hemisphere enjoy a winter of fun while listening.

00:02:19   But here, we're in the summer of fun.

00:02:22   So over the next few months, we'll have various things going on, topics and full episodes and all kinds of stuff, whatever tickles our fancy.

00:02:30   This will run all the way up and the summer fun runs from the week after WWDC to the week before the iPhone draft.

00:02:39   That's the summer fun period.

00:02:41   Right.

00:02:42   So it ends with the draft for the iPhone event.

00:02:45   Today, we are going to be doing a bumper Ask upgrade because last week I asked for feedback and feelings that people were having about Apple intelligence.

00:02:55   We got lots of great feedback and we're going to go through it today.

00:02:59   But before we do, I've got loads of other stuff that I want to talk to you about, Jason.

00:03:03   And we'll start with a Snell Talk question.

00:03:05   Matt wants to know, in a hypothetical world where you have to choose to either have an Android phone and a Mac or an iPhone and a Windows PC, which pair would you choose?

00:03:17   Well, this is easy.

00:03:18   I've made it clear here that in my personal hierarchy of devices, the iPad and the Mac are on top and the iPhone is down below.

00:03:24   I just don't use my iPhone as much. I work at home.

00:03:27   I use my iPad around the house. I work all day on the Mac.

00:03:30   I've used the Mac since 1990, basically.

00:03:34   So the answer is clear. I'm never, ever, ever using a Windows PC.

00:03:41   So Android phone is the answer.

00:03:43   I really am not sure about this one for me. I keep going backwards and forwards on it the more I think about it.

00:03:48   Because I think an Android phone is closer to an iPhone than a PC is closer to a Mac in its capability and feeling.

00:03:57   Yep.

00:03:58   But I would really struggle without an iPhone, just like in my daily life.

00:04:05   Because I love just using an iPhone and all the features that an iPhone has.

00:04:08   But then I think, oh, I could get by with a PC because most of the stuff that I do is in a browser.

00:04:13   But then I'm like, oh, but what about recording and editing podcasts and stuff?

00:04:16   So I think I would have to go Android and Mac.

00:04:19   And you have a PC right there.

00:04:21   I do.

00:04:22   You have it right there.

00:04:23   Right over there.

00:04:24   So yeah, I would also go Android phone on the Mac.

00:04:28   Yeah.

00:04:29   I think the iPhone, most of the features the iPhone has are easier to be replaced than a lot of the features that the Mac has, I think.

00:04:38   The apps are not great on Android.

00:04:40   But it's close enough.

00:04:42   And again, like for me, the Mac is the Mac.

00:04:46   The Mac is the Mac. I'm not.

00:04:48   That's, no.

00:04:49   Nope.

00:04:50   No, but if they stopped doing the Mac entirely, I would probably just buy the last of that and use it forever.

00:04:55   Until I use it into the ground.

00:04:57   Uh-huh.

00:04:58   I'm sorry.

00:04:59   I'd buy two of those.

00:05:00   Put one away.

00:05:01   Right.

00:05:02   Use the other one until it died and then use the backup.

00:05:04   Good idea.

00:05:05   If you would like to send in a question to help us open a future episode of the show, I'd love some summer themed Snail Talk questions.

00:05:12   Yeah, please.

00:05:13   Just go to upgradefeedback.com.

00:05:15   Send us in your summer themed Snail Talks.

00:05:18   I have some follow-ups. Sam wrote in and said, "I was thinking about the watchOS 11 announcement and the vitals app.

00:05:24   One of the metrics that they use is blood oxygen.

00:05:27   Currently the watches in the USA that are new have the feature disabled.

00:05:31   However, considering they are highlighting this as part of the new vitals app, they must be able to get around the current band somehow or the new sensor possibly,

00:05:39   there's a new sensor possibly for the next series watch.

00:05:41   Are they using new algorithms?

00:05:43   I went to Apple's page, the watchOS 11 page, and looking for what was there, which is a note by the vitals app, which says,

00:05:51   like, you know, they put the little number and you go down to the footnotes.

00:05:54   It says, "The ability to measure blood oxygen is no longer available on Apple Watch units sold by Apple in the United States beginning January 2024."

00:06:02   Yep.

00:06:03   So the answer here is not, I mean, we'll see what happens with what they're doing with that.

00:06:09   But there is not a gate happening here because every Apple Watch that was sold before January in the US and every Apple Watch sold everywhere else that has an oxygen sensor, it works.

00:06:21   So that's where they are.

00:06:22   And I think Apple is not, is not going to limit a feature because some devices don't have it, because some devices don't have the blood oxygen sensor anyway, right?

00:06:34   Some of them, the older ones.

00:06:36   So I don't, I'm not, I just don't think this is evidence of anything.

00:06:40   It is. It's a funny thing though, right?

00:06:42   Because it's a new feature, which is using a sensor that they're not able, it's just like a funny set of circumstances, especially because watchOS 11 cuts off the series four and five, right?

00:06:55   So there may be an incentive for more people to buy a new Apple Watch, but then they won't have access to what is, I think, a great new feature.

00:07:05   So it's just like a funny, you know, I don't think Sam's trying to suggest that there is like something crazy going on here, but it is just like a funny thing that like the blood oxygen sensor has had nothing.

00:07:16   Like they announced it and then that was the end of it.

00:07:18   And now they finally have something new for it, but doesn't work on new Apple Watches in the USA.

00:07:23   Right, right.

00:07:24   I just, I don't think, because what Sam's trying to do is be like, uh, well, this must mean something, right?

00:07:31   Do they have a plan?

00:07:32   And I'm not sure they have a plan.

00:07:33   I mean, they're there, I'm sure there are things going on in the background, but like you can launch this feature with just all the existing watches sold before January 18th and the rest of the world.

00:07:42   And it's fine.

00:07:43   In fact, American users should be, should be used to the fact that sometimes, uh, every everybody else in the world has learned this fact.

00:07:50   Sometimes there are features on Apple products that are only available in certain parts of the world.

00:07:54   Yep.

00:07:55   Imagine that.

00:07:56   And I think what Apple is trying to do, which is where we left it last time, is they are trying to fight this in courts.

00:08:03   That's what they're trying to do with Massimo to get the, the, the, the bads overturned.

00:08:08   That's what they're trying to do.

00:08:09   Sure.

00:08:10   D. Griffin Jones wrote and said, Mike mentioned that his iPad was in dark mode, but the home screen widgets were in light mode because he had light mode icons.

00:08:20   And that was true.

00:08:21   So when it made, I had to switch to dark mode icons to get my widgets to be in dark appearance.

00:08:28   But D. Griffin says, according to the developer release notes, this is apparently a bug, not the intended behavior.

00:08:34   And I went and looked and there is a known issue, which is widgets on the home screen might match icon appearance rather than in system appearance.

00:08:41   So I think that it's great that it's actually supposed to be system appearance so I can keep my icons light if, because I'm used to that.

00:08:48   I like the way that they look, but I want my device in dark mode, which would mean that I would want my widgets in dark mode because that is a larger portion of the UI.

00:08:57   But that is an issue right now with one of the beta.

00:09:02   It's a beta.

00:09:03   Developer beta one.

00:09:04   Yep.

00:09:05   Surprisingly, there are issues.

00:09:07   And there is a two part episode of upgradey award winning podcast and my favorite podcast, The Town with Matt Bellamy with Jon Stewart as the guest.

00:09:18   And it's fantastic.

00:09:19   I recommend people go listen to it.

00:09:21   But Stewart is talking about in more detail than he has before of what happened with him and Apple with the problem with Jon Stewart.

00:09:30   And essentially the way that Stewart puts it is as the show was becoming more successful and more viral, it seemed to raise questions at Apple about whether they wanted to pay for a show which could be focusing on things that could hurt them essentially, like content wise.

00:09:48   And Stewart is incredibly fair in a way that makes perfect sense to me in basically saying that this is always the deal when you produce content for a corporation, no matter who that corporation is.

00:10:03   They might not like what you're doing and if they don't, that's the end.

00:10:06   And Stewart says the thing that was different for him was the situation changed over time.

00:10:12   And I think it changed for everyone and I still stand by what we said initially where if this show was I think more successful, like more successful for Apple, that may have changed the calculus a little bit.

00:10:23   But I don't know.

00:10:24   But whatever it was, like over time it just became less of a thing that Apple wanted from Stewart and Stewart wasn't willing to understandably have to like work so closely with Apple and all of the content that he wanted to make as to whether they would accept it or not.

00:10:40   Really good to listen to.

00:10:41   I still believe firmly that if this had been a bigger hit critically and/or, you know, not ratings wise, but like statistics wise for Apple, they would have kept it.

00:10:54   But the and I think they probably went in thinking, yeah, he's going to touch on some sensitive subjects, but it's okay.

00:10:59   It's Jon Stewart.

00:11:00   We won the race.

00:11:01   And then after a couple of years, they're like, this is not doing it for us.

00:11:04   And it's worse than we thought on the content side.

00:11:08   To Apple's credit, Stewart says they paid his staff for season three, even though they didn't make season three and that they didn't have to do that.

00:11:16   So he, I think Jon Stewart's a realist, right?

00:11:20   He understands what that he's working for big corporations and that he understands very clearly that big corporations are going to big corporation.

00:11:27   So, yeah, yeah, he gets it.

00:11:31   He gets it.

00:11:32   I also really respected that he said he didn't think the first season was very good.

00:11:36   I was like, oh good. So it wasn't just me then.

00:11:39   You know, um, but yeah, it was, it's a, it's an excellent two part episode.

00:11:44   Like it's just worth listening to the show.

00:11:46   So good anyway, the town, it's just a fantastic podcast that I think a lot of people listening to the show would enjoy.

00:11:52   Um, and you should make, could maybe start here.

00:11:54   And if you like this, you're going to like the other shows.

00:11:56   Time for the details.

00:11:59   Woo hoo.

00:12:01   So we've had, I'm assuming a little bit more time with the betas.

00:12:05   I know that you, uh, I know you've, you've been using vision OS too, not just because you wrote about it, but because me and you, uh, hung out in Bora Bora together, uh, last week.

00:12:14   We did. We hung out on the beach in Bora Bora.

00:12:17   We did.

00:12:18   For a while I was staring at the mountains instead of the sea, which seems weird, but, uh, we tried to do a share play of a photo that didn't work because it's beta.

00:12:27   Um, but we did, we did a little spatial persona hang out on the beach, had a nice chat, talked about vision OS, talked about other stuff.

00:12:34   It was, uh, it was good.

00:12:36   It's that same old thing, which is, I feel like I want to start scheduling spatial persona hangouts with my long distance friends because I can't just ring you up and say, Hey, Mike, are you on the vision pro?

00:12:47   Because you're not right.

00:12:49   You're not, I'm not, you got to go get it and put it on.

00:12:52   But it's so good though.

00:12:54   I still would love an alert, like in the people view where it's like, this person is using the vision pro right now.

00:12:59   I would really love that as an optional like status indicator, but, um, yeah, it's really good.

00:13:05   We should, we should set that up a little group.

00:13:07   I heard, uh, Steven, uh, on connected, I think said, Oh my God, I'm going to run that update.

00:13:12   And then I need to find somebody to talk to.

00:13:15   Right.

00:13:15   It's like, yeah, don't get left out of the spatial persona campfire.

00:13:18   It's uh, it's cool.

00:13:20   It's happening.

00:13:21   So yeah, vision OS 2.

00:13:22   It's good.

00:13:23   It's good.

00:13:23   I can, when we talked last week, you would use it and I hadn't.

00:13:26   I can just say, yeah, you're right.

00:13:27   It's good.

00:13:29   Um, there are some bugs, but I would say it's better than this one and I'm happy to be on it.

00:13:35   Also.

00:13:35   Why, uh, why do we have this thing?

00:13:38   If not to be on the cutting edge and try new stuff, like that's the whole purpose of buying this thing and using it now is that it's got all of this and Bora Bora is really nice.

00:13:47   Although really interesting detail about our conversation, which is I told you I was going to write an article.

00:13:53   Which is literally the next thing we're going to talk about on this podcast and I decided to write it in vision OS and I was going to write it in Bora Bora.

00:14:00   And you're like, I don't want to do work in Bora Bora.

00:14:02   No, I don't want my brain to think that the beach is a place I do work.

00:14:06   And when we were done, I was like, well, I'll show Mike.

00:14:08   I'll write my article in Bora Bora.

00:14:10   And I wrote about one sentence and I thought, I don't want to do this.

00:14:14   And I went to Joshua tree and I worked there too.

00:14:16   Yeah.

00:14:17   That, that environment is very, very effective.

00:14:20   Like I've, I've been like sitting there, lowered my blood pressure and it feels relaxing.

00:14:26   Like it's genuinely very relaxing.

00:14:28   And so, yeah, I recommend people use the Bora Bora beach environment, not to be doing their email, like use that.

00:14:35   You know, maybe you're chatting with friends, maybe you're playing games.

00:14:38   Like it's just, that should be a chill out space and chill space for sure.

00:14:42   For sure.

00:14:43   There's many other environments use, use one of those.

00:14:45   Yeah.

00:14:46   But I like it.

00:14:47   It's a, and that's, that's the thing.

00:14:48   Oh, also I wrote that.

00:14:50   So I wrote this next article we're about to talk about entirely on vision pro and with the apple Bluetooth keyboard.

00:14:57   And what a wonderful thing to see my keyboard through the environment of Joshua tree, not Bora Bora, because that's a feature.

00:15:06   And it's a little weird.

00:15:08   It's sort of hazy around the edges and depending on where you turn it, it kind of disappears or reappears.

00:15:12   But bottom line is, and I'm not a hunt and peck typist, but like, I do need to like,

00:15:17   I do need to like know where the keyboard is and it helps to look down and be like, are my fingers to like,

00:15:24   are my fingers on the right keys to start?

00:15:26   And then after that, I'm okay.

00:15:27   But I do need to orient.

00:15:28   So it's really nice to have it there.

00:15:30   It's still, you know, a keyboard floating in space with hands that are floating in space.

00:15:34   I feel like there's more work they could probably do in the long run on this, but, uh, that was, that was a good thing out of vision OS two as well.

00:15:42   You wrote an article, uh, about Apple intelligence, uh, with a great, just a great headline, which is okay.

00:15:49   Fine.

00:15:50   Here's Apple intelligence.

00:15:51   Maybe you can dig into that in the second, but I wanted to read a quote from you.

00:15:55   Um, the truth is that Apple right now is like a duck serene on the surface, but paddling furiously underneath.

00:16:01   It was clearly complacent about the pace of AI innovation and allowed itself to get a bit too comfortable.

00:16:07   And now it's hiring to keep up.

00:16:08   I think it was a really effective way of describing kind of their spot right now, which is they're doing a lot, but it's very clear to me.

00:16:17   You people listen to this show that last week's WWDC looked like a regular WWDC, like on the surface of things.

00:16:25   When you hear about what they're saying about when we're going to be getting these features, that's not a normal WWDC, right?

00:16:32   But like, where straight up, they're saying some of these features are coming next year, not later this year, not in the fall next year, right?

00:16:42   Which may mean July next year or whatever, you know?

00:16:46   That aggressively future tense Siri block is the one that sounded the least like Apple of anything I think maybe I've ever heard in a WWDC where it was all just,

00:16:54   because Apple is the company that says we're doing this now or soon.

00:16:57   You know, Microsoft is a company that traditionally has been like, you announce it and then say it's coming and then maybe it happens, but much later.

00:17:05   But that Siri thing was really just all like, Siri will be able to do this and one day Siri will be able to do this.

00:17:11   And you're like, wow, okay.

00:17:13   And that, yeah, it is them coming off as, they want to come off as confident in AI, but they're definitely furiously trying to catch up in certain areas.

00:17:23   I said in that article, there's the three, I'm going to mention this last week too.

00:17:26   It was a blur last week.

00:17:28   There's three kinds of AI announcements at WWDC this year.

00:17:31   And it was the ones that they were going to make anyway, the ones they probably were going to make, but maybe not this year.

00:17:36   And the ones that they absolutely wouldn't normally make that they're making because they feel like they need to send a message to the world that Apple is on it, which seems to have been received.

00:17:46   It's one of the reasons that they did a full court press with like YouTubers talking up the AI features.

00:17:51   Like there was a real marketing push happening here because Apple wanted people to get the sense, even just generally that Apple's on it.

00:18:01   That's kind of a message they want.

00:18:03   It's like, oh, there was all this talk about if Apple was going to miss the AI boat.

00:18:06   And I think we are going to dig into all the details of where they are and aren't solid on AI stuff.

00:18:14   But I think the global broad view is, oh, they're on it.

00:18:18   Okay.

00:18:19   Like now I'm not worried.

00:18:20   And like the stock market reacted to that.

00:18:23   I mean, the tone of the analysts has changed.

00:18:27   Like a lot of this stuff has been altered by this, which was their goal all along.

00:18:32   But as we dig into it, the specialists in Apple, you can see the cracks, right?

00:18:39   You can see like the stuff that was already coming because that narrative that they weren't on it at all is wrong.

00:18:45   But there's also the stuff that they rushed probably and the stuff that they are like really out of their element trying to get because they don't want to be seen as behind and they don't want to be left behind.

00:18:57   So there's a lot going on there.

00:18:59   I tried with the headline.

00:19:01   So, you know, I write an article, sometimes I write the headline first.

00:19:03   Usually I don't.

00:19:04   I get to the end.

00:19:06   And what I wanted to get across is, so back in the day, I went to the, what was it called?

00:19:16   Which gate was it?

00:19:17   It was the one where the, Antennagate.

00:19:19   It was Antennagate.

00:19:20   Ah, yes, Antennagate press conference, which was like Steve and Phil and Tim all sitting up on stage at town hall in Cupertino at the Loop.

00:19:34   And one of the things that Steve Jobs said, which was so dismissive, was like, you can have a free bumper case if you want.

00:19:45   It was like, we don't think this is a problem, but it's annoying to us that you think it's a problem.

00:19:51   So fine, we did something for you.

00:19:53   Here it is.

00:19:54   That was the attitude.

00:19:55   And I felt a little of that at WWDC this year.

00:19:58   So I wanted something that was a similar kind of tone.

00:20:01   And so it was, okay, fine.

00:20:04   Here's Apple intelligence.

00:20:05   Right?

00:20:06   Like, all right, we did it.

00:20:07   Are you happy now?

00:20:08   And I think the answer is yes, that, that, you know, there's nothing, some of this stuff is just like nowhere near shipping, but like they did the job of sort of saying, we're on it.

00:20:16   Hold on.

00:20:17   We're going to get there.

00:20:18   We got it.

00:20:19   And that is not, I think it's not a place Apple's comfortable being.

00:20:22   I think Apple would much rather announce it and ship it.

00:20:26   Right.

00:20:27   But the realities of this situation, they couldn't do that.

00:20:29   There was no way.

00:20:31   I think if it's, I'm like, I was Googling around to see if I could find the, uh, the quote.

00:20:37   And I think he said like, okay, fine.

00:20:40   We'll give everybody a case.

00:20:42   Yeah, see?

00:20:44   Yeah, exactly.

00:20:45   Okay, fine.

00:20:46   Okay.

00:20:47   All right.

00:20:48   Fine.

00:20:49   Whatever.

00:20:50   Like it's really, that was the attitude.

00:20:51   So I, I felt like a little of that with this, we're just, okay, fine.

00:20:53   Apple intelligence.

00:20:54   All right.

00:20:55   Like that they were reluctant, that they weren't, their heart.

00:20:58   Not that they aren't aligned on it now, because I think organizationally they are aligned on it now, but I did get the sense that this was one of those things where they resisted for a while.

00:21:10   And then they're like, all right, you want it.

00:21:13   So we're going to do it here.

00:21:14   It is right.

00:21:15   And not that they're not committed to doing it, but that I think a normal.

00:21:19   Also, I think both of these things can be true.

00:21:21   I think that they were spurred into doing a whole bunch of things that they wouldn't normally do.

00:21:24   I also think that they were too conservative and limited before, and they needed that kick in the pants that I've been talking about.

00:21:29   And that's why I think that there are the things that they would have announced anyway.

00:21:33   And then there's those ones in the middle where it's like, you know, I bet they've been working on this for a while saying we'll ship this eventually.

00:21:39   And it was only because of this new focus on AI that those things got out now, instead of, you know, some of them, I was like, you know, old Apple before they got that kick in the pants would be like, yeah, well, let's just kick it down the road another year.

00:21:53   We'll get back to it.

00:21:54   And now there's no kicking AI features down the road, right?

00:21:58   Like that is not happening.

00:22:00   And then there's the ones where they're out on their skis a little too much where they're like, yeah, we're going to do it.

00:22:06   We're going to do it.

00:22:07   And like, it's not even there, but it doesn't matter if they're going to announce it and they're going to figure it out.

00:22:13   And, you know, it's a weird time.

00:22:17   It's a weird time for Apple.

00:22:18   They're doing some things that I think that they're not comfortable with, but you know what?

00:22:23   It's not the worst thing for Apple to be a little uncomfortable and a little hungry and a little motivated.

00:22:29   A lot of people told us the bumper solves the signal strength problem.

00:22:34   Consumer Reports said that this week we've heard it from a lot of people.

00:22:37   Why don't you just give everybody a case?

00:22:39   Okay, great.

00:22:40   Let's give everybody a case.

00:22:41   That's the quote.

00:22:43   I found it.

00:22:44   I found a YouTube video and that's the transcript of it.

00:22:46   Incredible.

00:22:48   Okay, great.

00:22:49   Let's give everybody Apple intelligence.

00:22:51   Yeah.

00:22:52   Yep.

00:22:53   Do you have any bigger thoughts on the beta at this point or are you mostly in the getting to know you kind of phase?

00:23:00   I mean, Vision OS is the only one I've used a lot.

00:23:03   I've got the Mac and iPad ones installed on devices and this week I'm going to install it on an iPhone and an Apple watch.

00:23:11   But I haven't had a chance to dig in.

00:23:14   I mean, part of it is that it's beta one and a huge amount of it is not going to be there for a while.

00:23:20   And there are other issues about things at WWDC that I want to write about, right?

00:23:26   So I don't have a lot more to say right now.

00:23:31   I mean, we've got a summer of betas for that and we'll talk about it more.

00:23:38   Anything you've noticed?

00:23:39   I have three small things that I think are really cool.

00:23:43   Two of them I'm very excited about for myself.

00:23:46   One is Vision OS 2 now features a setting for apps persist on restart.

00:23:51   Wonderful.

00:23:52   So previously when your Vision Pro shut down, turned it back on again, you had to open all your apps again.

00:24:00   Now it will keep them, which is wonderful because the battery dies on that thing way faster than I want it to.

00:24:06   Exactly.

00:24:07   Or to save the battery so you don't have to wait five minutes for it to charge back up.

00:24:11   You detach it and it's dead and then it restarts at which point again, you really wanted to open up back where you were.

00:24:19   Zack also was a good one that I should have put in here too in the Discord.

00:24:23   Compatible apps, so iPhones and iPads, you can now choose to have them set to dark mode if you want.

00:24:29   I would still like more dark mode options in general for even for Vision OS apps.

00:24:35   Like for example, Notes, like even I can't not see a white page.

00:24:41   I would like a black page with white text, but hey, it's better.

00:24:45   You can select preferred home hub now in the home app.

00:24:48   So this can be good.

00:24:50   Sometimes I have in my studio, I have a couple of home pods here and an Apple TV.

00:24:54   And just sometimes the Apple TV won't see some of my home kit items and I don't know why.

00:24:59   And I have to have multiple days to be plugging and unplugging and plugging and unplugging.

00:25:03   And if I could just change it to the home pod mini, which always works, it'd be great.

00:25:07   So you can now make the choice of your preferred home hub.

00:25:10   And there is going to be an express mode for unlocking doors.

00:25:15   Like there is express mode for transit, but even better using ultra wide band on an Apple Watch or an iPhone.

00:25:23   If you have an ultra wide band compatible door lock on your home, you can just walk up to the door and it will open.

00:25:29   No tapping needed.

00:25:31   Yeah. I saw this and thought, oh God, I'm going to be buying another.

00:25:36   You're going to be buying another lock.

00:25:38   I bought an NFC lock to replace my Bluetooth lock because the NFC lock is way better.

00:25:44   But for those who are wondering like why all these different technologies, the problem with smart locks is that device proximity is not enough.

00:25:53   So the Bluetooth door lock, just think about it this way.

00:25:58   If you're in your house and you walk up to the inside of your front door to go somewhere else, but you walk past the front door.

00:26:06   If the door lock sees that you're close to them with your watch or your iPhone, the door lock doesn't know that you're inside.

00:26:14   It just knows you're close.

00:26:16   So if you said to unlock when you're close, your door just unlocked randomly as you're passing by.

00:26:21   And that's no good.

00:26:22   So the Bluetooth door lock I had, the way it worked, it was very clever, clever hack, but it was a hack is it wouldn't auto unlock with Bluetooth until you left.

00:26:32   And it would use location services and it would literally draw a circle around your house.

00:26:36   And so if you went down the street, you know, far away, whatever it was, far enough away to be outside that circle, and then you came back and it saw you, it would unlock the door.

00:26:45   And that, so it was a proxy for you must be outside because you left the house.

00:26:50   But that's, but it didn't really know.

00:26:53   The NFC locks are less convenient in one way, which is when it worked and it didn't work that often, but when it worked, the Bluetooth auto unlock meant that my door unlocked itself as I was walking up to it, which is really awesome.

00:27:08   And the NFC, it all is like home key, it's express pass, it's in the watch, it's in the phone.

00:27:15   I can hold my watch when I come home, I just hold my watch up to the lock and it unlocks.

00:27:20   It's not bad.

00:27:22   NFC though, or UWB though, we've talked about UWB technology a lot.

00:27:28   It isn't used in a lot of things. It's starting to be in car locks.

00:27:30   It's going to start being in door locks.

00:27:32   And this is why it's absolute positioning.

00:27:35   The UWB sensor knows exactly where you are in 3D space, which means it knows if you're outside the door or inside the door and can be set to detect you approaching the door from the outside and unlocking automatically.

00:27:51   It will be, or your car door, it's the same idea.

00:27:55   It isn't going to unlock your car because you're in your car, right?

00:27:58   Like it needs to know you're outside your car and now it's going to unlock that door.

00:28:03   So, you know, yes, again, somebody will announce they're making a UWB lock in the next six months and then they won't ship it for another year and then they'll be hard to come by.

00:28:15   But eventually, tune back into a far future episode of Upgrade.

00:28:19   I will be undoubtedly buying another one of these and writing about it because it is, this is the dream.

00:28:25   This is how all this stuff should work, right?

00:28:27   It's the absolute positioning, knowing the context so it can literally know I'm walking up to the front door so you should probably unlock the door.

00:28:35   But not yet.

00:28:37   This episode is brought to you by Delete.me.

00:28:40   Do you ever wonder how much of your personal data is out there on the internet for anyone to see?

00:28:45   Privacy is an important thing. It's important to me, I'm sure it's important to you.

00:28:49   And it can be an uncomfortable thought when you consider that it could be information out there that could lead to identity theft attempts, phishing, harassment, unwanted spam calls and more.

00:28:59   But now you can protect your privacy with Delete.me.

00:29:02   Having your personal information on the internet can feel a bit like leaving doors open.

00:29:06   Delete.me not only helps you close those doors, it locks them, keeps your information safe and then makes sure nobody else will find a key.

00:29:12   So you don't have to worry about waking up one day to find out your information has been compromised.

00:29:16   And it's why I use Delete.me.

00:29:18   They remove my personal information that I don't want online and it makes sure that it stays off.

00:29:24   Delete.me is a subscription service. It removes your personal information from the largest people search databases on the web and in the process helps prevent potential ID theft, doxxing and phishing scams.

00:29:35   Now being a subscription service is actually a pretty important thing because Delete.me isn't just a one-time thing.

00:29:41   They are monitoring regularly on your behalf, doing the hard work for you.

00:29:45   You get regular personalized privacy reports showing what information they found, where they found and what they removed.

00:29:51   I get these every time I get one in my email. I love to look through and I'm always surprised at how many services there are and how many new services there are since the last time that I've got some piece of information that Delete.me is removing.

00:30:04   They are like playing whack-a-mole with my information on the internet which I'm very happy about.

00:30:08   Making sure that the stuff that I want and I tell them by the way, this is one of the key parts of it that I like.

00:30:14   When you sign up for Delete.me, you tell them this is the information I want you to remove because I don't want all of my information removed.

00:30:21   I don't want to be a ghost on the internet but I don't want my personal information to be out there.

00:30:26   Take control of your data and keep your private life private by signing up for Delete.me today.

00:30:31   I have a special offer, a special discount for listeners of this show.

00:30:35   Get 20% off your Delete.me plan when you go to joindeleteme.com/upgrade20 and use the promo code UPGRADE20 at checkout.

00:30:44   The only way to get that 20% off is to go to joindeleteeme.com/upgrade20 and enter the code UPGRADE20 at checkout.

00:30:53   That is joindeleteme.com/upgrade20 and the promo code UPGRADE20.

00:30:59   A thanks to Delete.me for their support of this show and Relay FM.

00:31:04   Room around up time.

00:31:06   Yeehaw.

00:31:07   Mark Gurman is reporting that no money is changing hands between Apple and OpenAI for their integration into iOS 18.

00:31:15   A quote from Mark Gurman, "Apple believes pushing OpenAI's brand and technology to hundreds of millions of its devices is of equal or greater value than monetary payments."

00:31:25   And I'm sure Eddy Cue said that very succinctly to Sam Altman.

00:31:30   That's my, Mark Gurman did not say that part, I said that part.

00:31:34   The expectation that Apple has here and that OpenAI have here is that it will cement OpenAI's leadership position in the industry, let more people see the benefits of the service, and then in turn push more people to sign up for a pay plan.

00:31:47   More people know about it, if more people value it, maybe they'll sign up.

00:31:50   If they sign up on the iPhone, Apple will get 30% of that, of course.

00:31:54   And keep in mind, like why would you want to cement OpenAI as leadership position if you're Apple?

00:32:00   Well, I mean, we're going to say in a second, they're going to open this to other things too, but OpenAI is not Google.

00:32:05   And I think that Apple strategically is happier having OpenAI and yes, Microsoft be stronger in this area because the alternative is vertically integrated Google, which means Android, which means it really puts the iPhone at a disadvantage.

00:32:23   And if you're OpenAI, well, your biggest competitor owns one of the two platforms for phones, and that's the most important platform.

00:32:33   So that's not great, right? So if you're OpenAI, Apple is also a logical partner. And so here we are.

00:32:40   Apple has been talking to Google and Anthropic.

00:32:44   It's expected that Gemini will be an option later on in the year.

00:32:48   I still believe that if they could have announced it, they would have.

00:32:53   Because they were very clear, right? Like outside the keynote, Craig could stop talking about Google Gemini.

00:33:00   Everywhere he popped up, he was mentioning it.

00:33:03   We love it.

00:33:04   And I still stand by the fact that I think Google wants money for it and Apple doesn't want to give it to them, as we can see from this.

00:33:09   And then further from Mark Gurman, "Eventually, Apple aims to make money from AI by striking revenue sharing agreements whereby it gets a cut from AI partners that monetize results in chatbots and Apple's platforms.

00:33:22   The company believes that AI could chip away at the billions of dollars it gets from Google search, from its Google search deal, because users will favor chatbots and other tools over search engines.

00:33:32   Apple will need to craft new arrangements that make up for the shortfall."

00:33:36   Yeah, it's going to be interesting because it's possible that will happen, but we don't know for sure.

00:33:42   But it's a risk at Apple's search deal.

00:33:47   And from Apple's perspective, Apple's got this great platform and it can point the fire hose at whoever it wants.

00:33:55   And so the argument there is, if this is going to be valuable for your business, and I think that in the short run, maybe this stuff is so expensive that it's going to cost OpenAI.

00:34:05   But in the long run, if this is truly going to be a foundation for your business, then Apple choosing to have you be on its platform is a big win for you because you're building your business.

00:34:22   And Apple basically, you should pay Apple.

00:34:26   And I think they're right.

00:34:28   You should pay Apple for the user base of Apple's platforms.

00:34:34   That only makes sense in the long run if this is truly a foundation of your business.

00:34:39   If it's not, then it's not.

00:34:41   The way I phrased it in my piece about "Okay, fine," was just think about all the things that Apple did.

00:34:49   The hottest company in tech, literally the hottest company in tech, OpenAI, just gave its crown jewels to Apple for free.

00:34:55   Apple responded to that by introducing its integration as optional, turned off by default, and tagged with warning labels.

00:35:02   To which I say, "Who has the upper hand here?" Exactly.

00:35:05   Because the whole idea was, "Oh, Apple is going to have to big OpenAI, and they're so behind, and it's so sad."

00:35:11   And then when it comes, Apple's like, "Yeah, we got OpenAI. Sure we do. I mean, it's off by default. You can use it if you want, I guess. We'll put a warning label on it."

00:35:20   That was their attitude. It's just wild. Just absolutely wild.

00:35:24   It's definitely different to the story that we thought it was, which was that they were going to be using OpenAI as the underpinnings of everything, which is very different to where it's been.

00:35:35   Very much not the case. Ben Thompson mentioned this, and you put it in our show notes, but this is the point, which is, "Apple," as he said, "Apple is positioning itself as an AI aggregator."

00:35:47   "This is the power of its platforms," he says, "The company owns users, and by extension, generative AI demand by virtue of owning its platforms, and is deepening its moat through Apple intelligence, which only Apple can do."

00:35:58   The on-device OS integrated. "That demand," Ben writes, "is then being brought to bear on suppliers who probably have to eat the cost of getting privileged access to Apple's user base."

00:36:07   This is what I was saying before. It's this idea that the reason Google pays for Google Search on iPhones is that there's a whole lot of users who just use Safari on iPhones.

00:36:19   They use it on all of Apple's platforms. Apple's user base is incredibly valuable. You could say to yourself, "Why would Google give money to its competitor?"

00:36:28   It's like, it's its competitor on the phone space, but in terms of search, it is a great source. They wouldn't be spending that money if it wasn't an enormous source of traffic.

00:36:38   So if you're an AI company, especially if you're competing with Google, but also if you're Google, you look at the Apple user base and say, "Wow, those are really valuable customers that we want to reach, and we don't want our competitors to reach."

00:36:53   If you're Apple, you say, "Well, yes, that's why you're going to pay us for access to our platform because they're our users, not yours, and we can choose where they go, at least by default."

00:37:03   Ben's absolutely right. That means that Apple is aggregating the AI services. It'll be really interesting to see how they all respond because we are potentially in a very weird place right now where AI is more like an idea of the future, but it's not quite the future yet.

00:37:21   If AI turns into search engines, where it's essentially a huge monetizable blob, where it turns out OpenAI and Google are just making money hand over fist on their AI stuff, then they should absolutely pay Apple for access.

00:37:38   The problem is right now it's very expensive. They're probably losing money on every query, and so now is an awkward time for Apple to demand that ransom, but at least OpenAI ended up doing the math and saying, "How about free? How about we just go in on free for now?" And that's really interesting.

00:37:59   Although I don't like, as we're talking about this, I'm hearing some of the things that I don't like, which is Apple making the decisions again, right? And it feels like the App Store all over again, right? They are going to be the gatekeeper for access to their users, and Apple will then financially benefit as a company based on the decisions that they make.

00:38:19   Yeah, although what I would say is if what we're hearing is true and Apple is going to integrate all the AI providers, right? Well, okay, many of the AI providers.

00:38:32   They will decide.

00:38:57   It'll be interesting to see how it plays out, but yes, this is acting as a gatekeeper but in a limited way. I think what would happen is if I had an AI chatbot that Apple decided for whatever reason it refused to have as an option in its integration, I would have a real beef with them.

00:39:14   But I think Apple could allow everybody in and then also take a cut of the queries that go to all of them, and that may be the direction it's going. We'll see.

00:39:27   Moving on, in his Power On newsletter, Mark Gurman reports that Apple is working on making thinner devices across the board after the introduction of the new iPad Pro. This would include a slimmer iPhone that we spoke about before rumored for 2025, as well as work underway for the Apple Watch and the MacBook Pro.

00:39:46   A quote from Mark, "The plan is for the latest iPad Pro to be the beginning of a new class of Apple devices that should be the thinnest and lightest products in the categories across the whole tech industry." Mike Hurley says, "Yes, please. Go for it. Make them good, make them capable, make them thin, make them light."

00:40:03   That's what I love about Apple in general. We've both spoken about our love for the M2 MacBook Air. This is part of the reason. It's incredibly powerful. It's incredibly thin and light. I love the new iPad Pro. It's incredibly powerful for what an iPad can do, and it's incredibly thin and light.

00:40:20   So yes, do this. I want the balance. If you can continue giving me that balance, great.

00:40:27   The danger is always you make it too thin and light and it's not powerful enough and all that, but I think the truth is that it's balance.

00:40:33   I don't think we're in that world anymore. With Apple Silicon, I think we're at the point where it's capable. They tried to do this previously before they were actually able to do it.

00:40:41   Ming-Chi Kuo is reporting that the Apple Watch Series 10 will not only be thinner, it will have a larger display. Kuo says that the screens are going to go from 41 and 45mm to 45 and 49mm respectively.

00:40:58   That would mean that the watches would also be getting bigger. As Benjamin Mayo at 9to5Mac points out, unlike other Apple devices, these sizes measure case height of the watch, not the diagonal screen size.

00:41:11   Naturally, a taller case also means a bigger display. So those numbers usually say like the Apple are saying, "Oh, it's a 13-inch iPad Pro." That's the display.

00:41:22   But on the watch, they're actually measuring the size of the case because that's kind of how you measure watches. That's like how you talk about watches, is like the physical size of the watch.

00:41:32   This means that the display on the larger watch could be as big or maybe a little bit bigger than the Ultra, depending on the way that they measure it all out.

00:41:45   I mean, look, this is a thing where we'll wait and see on that, right? Like, yeah, okay, bigger screen is one thing, but if it's thinner, maybe it's overall less bulky and it's not so bad.

00:41:56   Like, I think this is one of these things that just taking the numbers and making a decision about them, I would just caution.

00:42:02   But for me, if they made a watch with as big a screen as the Ultra that is a little bit more delicate in design and still has a great battery life, that would be my ideal.

00:42:15   I use the Apple Watch Ultra because I like the big screen and the battery. I don't necessarily need or want the Ultra for what the Ultra does. So we'll see.

00:42:23   Right. I mean, we'll see. And of course, they've got to be very aware of who is buying the smaller watch and feel confident that they're not going to get turned off by this, right?

00:42:32   That's the danger here is that they say, "Oh, congratulations. Your small watch is the size of the old big watch." That's like, that's not enough right there. So we'll have to see.

00:42:42   Lawyer up, Jason Snell.

00:42:46   Order, order.

00:42:49   There was a pretty huge news last week that got hidden by WWDC.

00:42:55   Japan has passed legislation that will require Apple and Google to allow third-party app stores and payment systems and to make it easier for users to change default apps and browsers.

00:43:06   Have you heard this one before? This will be called the Act on Promotion of Competition for Specified Smartphone Software. Not as good a name as DMA.

00:43:17   Maybe it's catchier in Japanese.

00:43:20   I mean, yeah, maybe. Yeah. Even if this is like the way they specify in English. Never mind. This will come into force by the end of 2025 and penalties for not following the rules would start at fines of 20% of domestic revenue.

00:43:36   So less aggressive than the European Union when it comes to fines. But we were waiting for it to happen. Like who was going to follow Europe and Japan is following Europe.

00:43:46   Yeah. Not surprising given that Japan has already been active in this area, but this is also that example of once the seal is broken, once the DMA goes into effect and Apple has to build all those features, it's very easy for other regions to say, I want that. Let's do that.

00:44:01   And Apple can't say they like, this is supposedly coming into effect at the end of next year, but like Apple can't say they don't have this built because they've got it built.

00:44:11   So it's just going to be other regions saying, yeah, Apple, all that. I think we are headed for a world where there are two iPhone experiences and it's in the legislatively opened world and the classic closed Apple world. That feels like where we're headed.

00:44:30   Genuinely, I think we may be getting closer and closer to an America and rest of world. I think is probably what's going to happen, but.

00:44:38   Could be.

00:44:39   Well, I guess, well, that's me forgetting about the DOJ case. So who knows?

00:44:44   Who knows?

00:44:45   The Financial Times is reporting that the European Commission is planning to charge Apple over DMA violations in the coming weeks.

00:44:53   It is suggested that they feel that Apple continues to steer customers away from using third party marketplaces in the EU and there is continued dissatisfaction with the makeup of the core technology fee.

00:45:06   The report from the Financial Times does say that the EU's reports would be preliminary, meaning that there could be adjustments Apple could make before any fines are levied.

00:45:16   So they may say again, like, hey, we're unhappy with this. You're going to change it? And then Apple will go, okay.

00:45:23   And they'll go back and change it again before they get levied many billions of dollars.

00:45:26   Or they'll get a fine and agree to change it, right?

00:45:30   Like that's the other way that this can go is they've gotten some fines and they'll probably continue to get some fines. But, um, because the, one of the reasons you do the fine is you're like the fine is for your past behavior.

00:45:41   And you agree like, okay, we'll pay this fine and then make this change so we don't get fined in this way again.

00:45:49   This episode is brought to you by Tailscale. Tailscale is a programmable networking software that is private and secure by default and is the easiest way to connect devices and services to each other wherever they are.

00:46:03   It's super secure, providing remote access to production databases, servers, kubernetes and more. And it's fast, like really fast.

00:46:11   Plus it grants privacy for everyone and every organization. It is an intuitive programmable way to manage a private network, including a zero trust network access that every organization can use.

00:46:23   With Tailscale, you can use device posture management to granularly restrict access to resources based on a wide range of attributes like OS version, user location and more.

00:46:34   You can save time of a trusted and proven networking solution that just works and you can code from your iPad with Tailscale and VS Code Server.

00:46:43   Try Tailscale out for free for up to 100 devices and 3 users for free at tailscale.com/relayfm, no credit card required.

00:46:52   That is tailscale.com/relayfm to try Tailscale for free for up to 100 devices and 3 users, no credit card required.

00:47:01   Our thanks to Tailscale for their support of this show and Relay FM.

00:47:05   So on the last episode, I encouraged the Upgradients to reach out to us. Tell us how they're feeling about Apple Intelligence, how they're feeling about how we feel about Apple Intelligence.

00:47:19   And we got lots of good feedback. And what I've tried to do is bring some of it in today. We're going to talk about it.

00:47:25   There was obviously a lot of overlapping stuff so I'm going to try and pick out things that are relatively unique or have different takes on the same point.

00:47:35   And we can talk about it. So I'm going to read these two. Oh, thank you. Sorry, yes. Extra lasers today. I'll give you another attempt at lasers in a minute.

00:47:46   So this first two pieces of feedback I'm going to read together. So Dave writes in and says, "Criticism for Apple using public web content to train its models prompts the question, 'What is the realistic alternative?'

00:48:01   Apple Intelligence's performance will be measured against ChatGPT and others. Years of Siri criticism likely made Apple weigh a likely worse performing Apple Intelligence against a less pure version.

00:48:13   The term 'table stakes' that you've used is to evaluate Apple. Using the same training data as their competition was probably considered table stakes by Apple in making Apple Intelligence."

00:48:24   And Stephen says, "It's rare I'm in disagreement with either of you, but I see no issue with the way Apple have gone about it.

00:48:31   They were between a rock and a hard place. If they didn't trawl the open web, they'd have been even further behind and have got hammered for it.

00:48:39   I think it was a calculated decision that annoying the techies versus impressing the populace only had one possible win."

00:48:47   I'll just say that what Dave said, I said last week.

00:48:51   Sure, I know. But we said a lot last week.

00:48:54   We did say a lot. But just to restate it, yes. Apple did what it did because it's trying to catch up with the others in this category and this is how they did it.

00:49:04   And Apple, my guess is Apple was like, "I don't love it." But they have no chance.

00:49:09   They view this as potentially an existential threat and everybody's doing it.

00:49:16   And I think what I said last week is if it turns out that some court ruling makes it illegal to use the open web without permission to build models,

00:49:25   all of them will get hit and be reset to zero.

00:49:28   But the alternative is that it's not illegal and if Apple stands up for this very specific issue that is debatable,

00:49:38   that Apple could potentially lose everything. And so of course they didn't choose that.

00:49:43   It's possible that I'm projecting, but I've heard Apple in a couple of places talking about this.

00:49:51   I get the impression that they're a little embarrassed about the fact that they did this.

00:49:57   I don't think that they're particularly proud of the fact that they have scraped what they call the open web to build their model.

00:50:08   I would expect there was a lot internally about this because they know that this is going to upset people and it is upsetting people.

00:50:17   I agree with these statements that Dave and Steven have made that if you were going to do this, this is probably the only way to do it.

00:50:26   I think that there are some ways Apple could have gone about this a little bit differently, which would have been better.

00:50:32   I think a scenario where they would have given people the opportunity up front to opt out,

00:50:38   and/or a system where instead of just what they're saying now where you can say, "Hey, don't include us,"

00:50:44   that there was actually a simple system to be like, "Pull me out of your model," even though it's a bit late for that,

00:50:49   but that may have gone a little bit further.

00:50:52   I genuinely think, though, in the ways that I've heard them talk about it, they're saying very specific language and trying to move on.

00:51:00   I don't think that Apple is proud of the fact that they have done this, and I'm not proud of them for having done it.

00:51:06   I'm completely of the belief that this is probably right now the only way to build a large language model,

00:51:13   because you need such a huge amount of text, and I don't think any amount of licensed data would probably get you there right now

00:51:22   if you want to be competitive at least, but I'm still disappointed about it.

00:51:26   And again, I believe it is possible for humans to hold two opposing views in their minds, and that be okay.

00:51:36   I am disappointed about the fact that this happened, but I am also looking forward to using the features that it will enable.

00:51:42   And corporations can also hold opposing views because they're made up of people who hold opposing views.

00:51:47   I would say the other tell here is that Apple is licensing content for AI.

00:51:51   They're licensing text content. They're licensing image content.

00:51:56   They're making some original image content to train their machine learning models.

00:52:00   That suggests to me that Apple actually does have some opinions about that all the content doesn't belong to them.

00:52:07   And yet for this particular foundation model, they couldn't see their way to do it.

00:52:14   They couldn't do it. Maybe in the long run that will change, but I think this is the case where Apple,

00:52:20   this was such a high priority that Apple said, "No, we're going to do this. You may not like it, person within Apple, but we have to do it."

00:52:28   And so they did it.

00:52:29   There is a thing though about the licensing thing which shines a light on it and makes it worse in some ways.

00:52:37   So, okay, Apple, which content do you think is worthy of a deal then?

00:52:42   You know what I mean?

00:52:43   So is it that text is okay to steal or scrape or however you might want to say it, gobble up, but imagery you should only license?

00:52:54   Why is imagery better than text? You know what I mean?

00:52:56   There are a million things about this and I think it's too complicated.

00:53:00   Let's get to the next bit of feedback because I think this will help us address that point actually.

00:53:06   Ben says, "On the topic of Apple training their LLM on the open web, I think it's useful to consider how their text-generative AI is being applied.

00:53:14   That is, so far, Apple Intelligence is only applied to augment or transform your personal data, like adjusting the tone or grammar of your emails or summarizing a PDF.

00:53:23   It's not being applied to prompts regarding world knowledge, which could replace the website that generated the source content.

00:53:30   It is an important distinction regarding the ethics here."

00:53:33   So everybody gets to have their own opinion about this, but I generally agree with Ben.

00:53:38   I listened to your conversation. Federico Fatici is very upset about this and I understand why.

00:53:42   There is a scale. Federico is at the top of that scale.

00:53:45   He's at the top of it.

00:53:46   Yeah.

00:53:47   So here's what I would say. I do think that for some uses, likening the AI training to search engine bots indexing the internet is a good parallel.

00:54:01   The fact is, it's not a copyright violation for Google to scrape or search the entire internet content and index it so that they can do a search of the internet.

00:54:15   It's not.

00:54:16   Some people probably disagree with that too, but I would say it's not.

00:54:19   The thing is, though, the thing is that that search is, at least up to now, has generally had a benefit for the publisher, which is it sends traffic to the people who find those pages.

00:54:36   And they're able to make money on that traffic.

00:54:38   And so there's a good, clear relationship there.

00:54:41   And the danger of using an LLM to find out information is that if you're generating the answer based on the contents of someone else's work and preempting the user from ever going to the generator, the generator is not being compensated.

00:55:00   And they've lost everything.

00:55:02   The content has been generated from, at least in part, their content.

00:55:06   They get nothing.

00:55:07   That's an issue.

00:55:09   So I think that's part of what's going on here.

00:55:12   And then the other part of it, I would say, which Ben touches on, is using all of the internet to generate a model that knows how people write and how people speak and can do things to correct your grammar or change the tone of your article or whatever you're doing.

00:55:34   I would argue different from using it to generate material.

00:55:42   And this is what some of the lawsuits, this is what the New York Times lawsuit is, is you can ask ChatGPT a question and basically a New York Times article comes out of it.

00:55:53   And that is, I would say, that is probably a copyright violation.

00:56:00   You're literally parroting my words back without giving me credit or without sending me traffic.

00:56:06   I'm less inclined to be bothered by the idea that you're scanning the entire internet because you want your LLM to understand what humans in general think, like a search index, than if you're using it to make new things, to be generative.

00:56:26   And I will say this for Apple, it seems to me like this is where Apple has drawn the line right now.

00:56:33   It seems to me that Apple is doing things that are transformative and not generative.

00:56:39   And where they are doing generative things, they are talking about licensed content.

00:56:44   So the image generation, they said they have stock libraries, not photos, like illustrations and stuff, and their own stuff that they trained it on.

00:56:51   And I don't think they said that the image playgrounds is trained on the open web.

00:56:55   I think they said it was trained on some specific things. I think it was part of what Craig Federighi was talking about.

00:56:59   I wonder if that's where they've drawn the line.

00:57:02   It's like, look, we're going to train the LLM using the web, and we don't love it, but all we're using it for is not world knowledge, this phrase that they keep using.

00:57:09   It's for the sum total of how words are put together by humanity.

00:57:14   And we're using it in ways that are helpful, but that are not going to generate content that takes the place of the content that we scanned.

00:57:23   And I think people can disagree about where that line gets drawn, but I think that's where Apple is drawing it.

00:57:32   And actually, I think that's probably where I would draw it, because I don't think it's fundamentally evil that Google built a giant business on indexing the internet.

00:57:42   Because I don't think they stole my stuff to index the internet.

00:57:46   I think the whole product was built to let people find my content on the internet and then read the content on my site, not read it in Google.

00:57:54   And the more Google has diverged from that ethos, I think the less ethical it has been.

00:58:01   And the AI generation is like an extreme lack of ethics on that point, I think.

00:58:08   But everybody can draw that line in a different place.

00:58:10   Anyway, I like Ben's comment here because I think there's something there about the idea of generative versus using the model to do some transformative things on your own content.

00:58:21   But again, reasonable people can absolutely differ on this point.

00:58:26   So I think the thing about Apple drawing the line on the text content, I think the thing that we don't know is did they do this because of philosophical grounds or technical grounds?

00:58:38   Maybe their model is not good enough to generate from zero. And so that's why it doesn't do that.

00:58:44   I'm sure their model can do generative and that it's terrible. And that's why they've walled it off. I'm sure of that.

00:58:51   And so the question then is if it gets better later on, does this line get redrawn? And I don't see why it wouldn't.

00:58:59   Are they drawing the line for ethics or expediency? Or both? Why can't it be both?

00:59:05   They clearly have not drawn the line against all generative AI from zero because image playgrounds exist.

00:59:12   So it's not like they have a fundamental issue with the entire idea.

00:59:17   But if that's being generated from licensed content, that's a different story, right?

00:59:21   I would agree, but I don't know if they do.

00:59:24   I don't believe that Apple have made specific statements because I think they're being purposefully cagey about what their belief of the ethics of open web content.

00:59:39   And that's why I think that's what… And the other thing for me is I agree with the point of it's maybe better if they're just using these models to create things that will do things for my device.

00:59:53   But ultimately the result is the same, which is like every company that is scraping the open web is doing it for their financial gain.

01:00:04   Apple is still doing this for their financial gain.

01:00:07   Oh totally.

01:00:08   But their financial gain is make the iPhone an even stronger product so that they continue to sell more iPhones.

01:00:19   And I believe that that is different to the search engine thing because the search engine deal, which is amended over time I admit, but the search engine deal was you write on the web and if it's good, people will find it on Google and come to you.

01:00:38   And then you will be able to do whatever you want from that point on, whether it's pay wall content to get people to give you money, put advertisements on it so you can make advertising revenue or build an audience.

01:00:50   The LLM stuff, however it's used, takes that content that was created under that arrangement that people had with the internet and with search engines and how content was found and completely removes it from the end user who has zero understanding of where it came from and any attempts that I have seen to try and point people towards that content is bad or pointless because once you've already gotten everything you need, why would you continue?

01:01:19   I agree, although that's not what Apple's doing with their stuff today.

01:01:23   No, but Apple today has taken that content, I went a little bit further, but Apple is taking the content from the web, they're using it to build a product that they are able to benefit from with their customers and the person who created the content that helped generate that model get zero for it. Nothing.

01:01:45   I see your point, I just think that there's no harm being done in there at all, so it's a different transaction because it's taking the sum total of knowledge.

01:01:54   I don't believe there is harm, I agree with you. But this isn't about somebody losing out, it's just the ethics of it.

01:02:04   I'm not saying the ethics are right or wrong, but they're there, and that's just the question that everybody has to ask themselves. Do you think this is okay?

01:02:16   Honestly, for me, I don't know the full answer to this. I have not gotten to my answer as to whether I think the ethics on this are okay, but I know that this is more of an ethical question for me than a search engine.

01:02:31   You know what I mean? I come down on the side of if it's indexing and not generating, that it's probably essentially a search engine index, and I'm not as concerned about it.

01:02:43   But like I said, reasonable people can differ about it, and there are different lines to be drawn. I think the generative stuff, especially when it's literally generating, for example, art in the style of an artist,

01:02:53   or using stock photography from an unlicensed stock photography source and generating, for example, the watermark. You can see how, and they can engineer that out, but you can see what that means,

01:03:07   is that they're literally taking pixels and relationships of pixels from copyrighted art and using it to generate new things. That's where I think it becomes completely unethical.

01:03:17   I don't feel that way about, and my human learning argument would probably come in here too, which is every human being learns by observing things in the world that don't belong to them, that are copyrighted by other people, and they take them and they learn them.

01:03:32   Every musician listens to music. Every artist looks at paintings or drawings. Every writer reads, and they are building a model in their head that is their brain,

01:03:43   and then they use it to generate new material based on that. I think there is a limit. I'm not saying that AIs are people, because they're not, but I'm saying there is some degree to which using things that are out there to inform you in a way that doesn't,

01:04:00   again, cause harm or generate copyrighted content is probably okay, but like I said before, if courts decide that this is too far, then all the AI companies will suddenly have to license content.

01:04:21   I would say that what's going to happen is they're not going to license Federico's content. That's not going to happen. Also, I would not put it past Google to say, "If you want to be in Google search, you have to agree to this."

01:04:32   I think they already do that, but if it came down to it, they'd be like, "You either disappear for the web or you're trained in our model. Those are your choices and you would have to decide."

01:04:40   We'll see how that goes, but from an Apple perspective, I feel like, yeah, if you think that this is fundamentally unethical to train models on the open web, even for general use and not for generating content,

01:04:54   then Apple is being unethical here, because I think Apple feels like they have no other choice but to do this due to everybody else.

01:05:03   Literally, it's like, "If your friends would jump off a bridge, would you jump off a bridge?" It's kind of like that.

01:05:08   Depends how many friends. Maybe if I have no friends left, maybe I would. I don't know.

01:05:15   Sirijit writes in and says, "I live in India. I use a mid-range Android phone that is seven times cheaper than an iPhone Pro. It absolutely does the job for me.

01:05:24   I am seriously thinking of getting an iPhone just for making my mother or my late grandmother into superheroes. It's too soon to tell what it might turn out to be, but to be dismissive is telling. Maybe you guys are just too jaded."

01:05:37   And an anonymous listener writes in and says, "Both of you sound overwrought about Apple including image playground and the Apple intelligence suite.

01:05:47   Sure, it is childlike, but not all iPhone users are serious professional types. Sure, you can doctor up an image using a photo of someone other than you, but nothing stops you from using markup now on a grandma's photo and giving her a mustache or a goatee."

01:06:01   Okay, I appreciate how everybody inserts thoughts about who we actually are. Yes, we're very serious professional types. That's what it's all about here. And super jaded as well.

01:06:13   Somewhere a phone.

01:06:14   Yeah, I mean, you got us.

01:06:15   People hear how, like, look, this is how people take it. I mean, I'm not going to argue with this. Like, if people hear me complaining about this and think that I'm jaded, like, I understand that.

01:06:24   Alright, well, I don't. So Sridhar's answer is it's kind of a bogus argument because you can generate stuff on Android phones just as easily as on iPhones. There are third party apps that do it. They already exist.

01:06:36   I don't think you need to buy an iPhone just to do this. Although I guess it shows you why Apple did it, right? And that's fine. Again, part of my problem is that I feel like mid journey and stable diffusion have ruined this art style for me.

01:06:52   Because now I look at it and I think it's just kind of gross because it's this weird AI art that's not. I think there's an uncanny valley. I think it's kind of gross. As we said last week, Apple is approaching this as being, it's fun. I'm having fun on my phone.

01:07:06   They've said numerous times, and I think this is really interesting. I guess this goes into our next comments too. Likening it to clip art, right? Where the idea is, you're not, and I think we mentioned this last week as well. We did mention it.

01:07:18   You're not really costing anybody a job here. It's like clip art. You're putting it in a message or putting it at the top of a newsletter that would otherwise have no art.

01:07:29   I think that's what Apple is approaching this as. It's like, come on, it's fun. Especially if it's not, in my opinion, if it's not being generated based on people's unlicensed artwork, then I'm more okay with it. But I don't love it, and I don't love the examples.

01:07:48   Remember, these are the best examples Apple could come up with because they use them in their keynote and their marketing materials. I didn't think they were great. In absence of any other evidence, I would say, is this really going to be good or is it going to be kind of weird and unpleasant?

01:08:06   But I agree with our anonymous commenter and our non-anonymous commenter here that I think that's Apple's intent is basically, come on, it's fun. People like making this stuff. It's true. Also, you can make that stuff with other apps. You don't actually need image playgrounds to do it.

01:08:21   But Apple is figuring again, if it's on the device and it's a default, makes it really easy to do it.

01:08:29   I'll just restate my points from last time. I think Genmoji is good. I think it's a great idea. I think people are going to have fun with that. I'm looking forward to that. That's going to be fun.

01:08:38   I don't like the idea of making it easy to create imagery of other people. I just don't like that. Especially based on photos of them. I think I find that gross.

01:08:51   But my biggest issue is that I just don't think what they produced in the keynote was good enough quality. I just felt like if you're going to do this, at least make it look good. And I don't think it looked good. So what was the point in doing it?

01:09:07   I think the number one rule is whatever thing Apple is showing in the keynote or in its marketing materials, this goes for the colored icons on the home screens too, that is literally the best Apple could do. And so you take that as the highest bar.

01:09:22   Because realistically, the stuff that they show is never as good as the stuff that you actually generate when you get your hands on it, of whatever it is. And so we looked at that stuff and we're like, "I don't know about that."

01:09:34   Also, I have a severe revulsion to the idea of people taking images of me and making generative art out of them. And so, you know, grandma's photo is great and I'm like, "But I don't like it." And so having Apple make that easy is unpleasant to me. I don't like it.

01:09:52   Eric writes in and says, "Is image generation just the next phase of clipart? I think if it's viewed that way, it's less impactful than trying to hurt creators. If I'm just looking for a basic image for a slide or a document, something that I made that's basic seems better than just grabbing a random image from Google."

01:10:08   And Brandon says, "One positive generative images might be the reduction of image copyright violations that happen when somebody wants an image for a PowerPoint, a flyer, a brochure, and the most common solution is a Google search, copy and paste. This probably happens millions of times a day and is likely never even considered by most to be an issue."

01:10:27   I think these are good points, interesting ideas. And I would just say, again, I'd like the images to be based on licensed material. And there's a whole other conversation about, like, did the artists who put them in the stock library and assign that agreement know that they were agreeing to this?

01:10:45   But for me, that would be the only thing because I totally see this point. I mentioned this, I think last week, that I have a friend who does a newsletter and like a Patreon and didn't used to have art and now has little cute cartoon art that's generated by an AI model at the top.

01:11:02   And like, he would never have been paying an artist and he gets to express himself and do a little bit of branding in a way that he couldn't have done before. And like, there are two ways that artists are harmed by this stuff. One is their content is stolen, unlicensed and used in the model, and then it looks like their art, but it's not their art.

01:11:20   And two, the harm is that they lose their jobs because nobody wants to commission art anymore. And those are the two things that I'm kind of looking at here, but I think this is a good point. There is truth to the fact, and I know people might not like this argument, but I'm going to make it.

01:11:35   There is truth to the fact that there are enormous copyright violations going on all the time. Just casual theft of images on the web. It happens all the time.

01:11:49   Let's be honest, listeners. We've all done it. Alright? Like, you cannot tell me you've never gone to Google and just copied an image and used it somewhere, anywhere, even in an iMessage. We've all done it, right?

01:12:04   That was why hearing this, I was like, "Oh, this makes a lot of sense." For that presentation you have, to have an easier way or a different way of creating that thing you're looking for, rather than going to Google and typing in, I don't know what, a photo of a building, or whatever.

01:12:24   Yeah, a baboon using a computer, right? Which is a great series of stock images. But those then get licensed, put on websites, and then other people come around because it's the open web and just rip it off. Like, just pull it out and use it themselves.

01:12:37   So, when that happened, there was a Wikipedia page. I paid for a licensed image. This is a long time ago now. A licensed image, and put it on my website. And I found that Wikipedia, there was a Wikipedia section that was using it as an icon. And it was very clearly pixel for pixel.

01:12:55   The size and everything about it was the licensed image from a stock library. And I said, "Do you have a license for this image?" Because Wikipedia wouldn't allow it, right? And the guy was like, "No, it's a picture I took of my grandma's TV." And I thought, "Well, that's a lie, but what else can you do?"

01:13:11   He found it, probably on my website, but if not, someone else's website that used that stock photo, and just took it. That happens all the time, very casually.

01:13:20   So, the idea of you needing something for the presentation, I think that this is a very valid way to use tools like this in a way that I had not considered.

01:13:30   The thing that I don't like, and I think the thing that I would assume most people could agree on, is the scenario in which somebody goes from paying an artist to not paying an artist. That's the thing that I find concerning about these tools, and is the thing that I don't want to happen, and that I'm sure will happen.

01:13:52   And that's what is disappointing to me about this, but I actually don't believe Apple's tools are of that quality.

01:14:00   Oh, certainly not. That's why they've constrained them to certain cartoony styles, and they've got very limited... I want to say again, one of the things I love about image playgrounds is they're trying to put a proper user interface on something that otherwise is a command line, where you have to kind of exhort the image generator to do exactly what you've got in your mind's eye, and then do it again and again and again, because it just doesn't get it right.

01:14:23   Apple is trying to give you a bunch of little things to click on, and even when you do the text, it sort of turns it into tokens that you can edit, and it's trying to do what Apple should be doing with all this AI-driven technology, which is build a good UI on top of it, because command line ain't it.

01:14:40   Right, like we learned this lesson before. Anybody who thinks the command line for a chatbot is enough, I'll introduce you to the TRS-80, I guess. We've gone past that. 1984 happened, the Mac happened, good UI on top of that stuff is better. So they are trying that.

01:15:00   This episode is brought to you by KRCS, one of the oldest Apple premium resellers and your one stop shop for all things Apple in my home of the United Kingdom.

01:15:11   I cannot tell you the joy that it gives me to be reading an ad from a UK based company. I feel like this is such a rarity and it makes me very happy.

01:15:21   At KRCS you can shop for the latest Apple products online and in store, book an Apple certified repair, and even trade in your preloved devices.

01:15:30   For all of our listeners in the UK, right now at KRCS you can spread the cost with 12 months interest free credit on any Mac over £1000 in value.

01:15:40   This means you can buy the latest MacBook Air M3 from just £78 per month over 12 months completely interest free when you shop online at KRCS.co.uk or in store at their 3 high street stores.

01:15:54   KRCS are one of the few places in the UK where you can customise the specs of your new Mac. Whether it's a faster processor, adding more storage or extra memory, or all three, you can do it at KRCS.

01:16:06   You can have the exact specification you want with pricing you won't find anywhere else.

01:16:12   Many of the popular models and configurations are kept in stock for next working day delivery when you order before 3pm Monday to Friday, so why not make your Mac yours and build your own?

01:16:22   Or choose one from stock and as long as it's over £1000 it qualifies for 12 months interest free credit.

01:16:28   With free next working day delivery on in stock items ordered before 3pm there's never been a better time to get a Mac at KRCS your local Apple experts.

01:16:37   Get more information today at KRCS.co.uk/podcast or click the link in the show notes.

01:16:43   These are available for 18 years and over, subject to approval in terms of conditions apply. Please spend responsibly, borrowing beyond your means could seriously affect your financial status.

01:16:54   Ensure you can afford to make your repayments on time by the due date.

01:16:58   One last time go to KRCS.co.uk/podcast to learn more. Our thanks to KRCS for their support of this show and Relay FM.

01:17:07   I have more, we have more Apple intelligence ask upgrade questions. There you go, we get some more lasers going on.

01:17:17   So like your half time lasers right there.

01:17:19   Yeah. So Steve writes in and says I'm surprised at your reaction to Apple's new AI features.

01:17:26   Apple has a long track record of borrowing, copying and even outright stealing from other developers and companies.

01:17:32   Apple got its start from borrowing technology that Xerox and HP already had.

01:17:36   AI is brand new, it cannot feel, cannot reason and cannot learn on its own.

01:17:40   AI only knows the information that's fed into it by people, whether that information is right or wrong or should or shouldn't be shared. Apple had to start somewhere.

01:17:49   Okay. I mean, he's not wrong in the sense that Apple often strikes at what it thinks is the right time. I think that probably is the thing.

01:17:59   Hence Sherlocking as an entire thing.

01:18:01   They were a little late and also integrating things into the operating system is the thing only the platform owner can do and that's where Sherlocking often comes from and that's fine too, I think.

01:18:10   Because Sherlocking gets this treatment as like a death sentence and it's generally not. Except in the case of Sherlock. Yeah, maybe then.

01:18:20   Anyway, there's some truth in that and yeah, so Apple's getting started and building its own thing and I think it did have to start somewhere with this.

01:18:29   I'm not sure I entirely buy the long track record of borrowing and copying but I mean, okay, you can view the world that way if you want to. I get it.

01:18:40   So yeah, I don't think I have… Given all of this, I'm not quite sure why Steve is surprised but that's fine. Great.

01:18:49   I hear what Steve is saying. I don't buy the idea that because you borrowed or sold them once that you should be able to do it forever and that I shouldn't be surprised if you keep doing it in new and interesting ways.

01:19:01   Oh no, that's… don't you know that's how theft works. If you steal and they don't catch you, then stealing is legal.

01:19:09   Raphael writes in and says "What worried me the most about Apple's announcements is the pure lack of taste. The generated images in the keynote looked so bad, like meme material. Was this really necessary? I feel they could have not shown this feature and no one would have missed it. Gemmoji would have been enough."

01:19:27   I think I wanted to include this one. I know we spoke about this kind of stuff already. I just like the idea of using taste as part of the argument because I do think that that was a key component here. I think that there was some taste that was lost in doing this stuff.

01:19:43   I don't work for Apple. I don't own Apple stock. I can't. So I'm going to put myself for a moment in the position of what if I worked for Apple and I don't know all the details of everything. If I were making the decision, I would not do image playgrounds and I would have just done Gemmoji.

01:20:03   And I would have been able to point at it and say, see, this is how constrained image generation trained on a limited set of emoji concepts makes a really nice generative model that is going to be fun and understandable. And maybe we'll do more in the future, but this is what we're doing right now.

01:20:23   And yet they even could have gone as far as to generatively made Memoji, which they didn't do. There is generatively created emoji that looked like your friends. Then there is this image playground stuff, which clearly builds on Memoji, but is trying to go a bit further than Memoji, I think.

01:20:43   So what I don't say, what I can't say is did this help build Apple's case with investors and the general tech industry about perception, fighting the perception that Apple's behind?

01:20:59   And my guess is that that's why the image playgrounds is in there. Everybody can have an opinion about all of this stuff, obviously, but for me, I didn't think it looked very good. Lack of taste is a great way to put it. But is it serving another model here?

01:21:18   Is it serving the other thing that they're trying to do, which is just get across that we get it, that we have AI stuff? Because we're not, the upgrade podcasters and listeners are not the primary audience for this keynote.

01:21:34   And we often are, but we're not. The big message here was to the rest of the world just saying, "Apple gets it. We're on the AI thing. We got it. Don't worry about us. We're fine. We haven't said anything about it, but here it is. We got a roadmap."

01:21:48   Apple doesn't do roadmaps, but they basically did a roadmap. And they're like, "Yeah, we got it." And is the image playgrounds part of that? It probably is. And that's probably why it's in there.

01:21:57   Then again, I also think, like I said last week, there's a group inside of Apple that is sort of like, "Come on, it's fun. We're having fun." And I think there's some truth in that.

01:22:07   And I also think that some of it is a little delusional, and which this is, I can't say. But I think that that's part of it too. Is they're like, "Look, we're being whimsical with our AI art." And your mileage may vary about how you take that.

01:22:22   Brad says, "In my opinion, the image playground feature is likely to be tucked away and rarely used. But how Apple reframed AI as an OS personal assistant was huge to me.

01:22:32   AI as a text box is so limited, but AI as an OS feature deeply integrated into every app changes everything. Seeing their plan and imagining the next steps is fascinating to me."

01:22:44   The reason I'm willing to accept any of this is because I want this. There's incredible power here. That's what's the most encouraging about this whole thing, right? Is all the places where Apple is not saying, "Here's a text box, figure it out."

01:22:59   Which is what AI has sort of been up to now for a lot of these companies. It's like, "Yeah, just here's a box, whatever." And Apple's like, "No, we're going to have very specific things that our models do, and we're going to build UI around them.

01:23:10   And we're going to have them solve problems that people have with their stuff." I used... Did I mention this last week? Again, it was a blur. I used... There's a new version of Grammarly that's got an upgraded AI model.

01:23:22   And I use that because I don't have a copy editor. And the new version, the first time I used it, I was like, "Oh my God." The old version was like, you know, it was basically a grammar checker. It was okay.

01:23:33   The new version suggestions were like suggestions from a copy editor. They were really good. And it's like, what I like about that is, yeah, there's an AI model back there.

01:23:44   You know, right? There's totally an AI model back there, but it is very limited in like, we are focusing on fixing mistakes in your writing and building a UI to show you, here's the change we're suggesting.

01:23:57   Do you want to make it or not? Do you want to tell us it was wrong? And then we'll learn from that. Like that kind of stuff, the potential here is enormous.

01:24:07   The potential for making our phones more usable is enormous. So there's a lot going on here that is interesting. I also understand, as we've covered a lot, the trepidation about it.

01:24:19   Sam says, "Do you think Apple intelligence, this base model, plus all of the data it builds based on the user will take up enough storage space that will start to need devices of larger storage?"

01:24:32   I doubt it.

01:24:34   You think so? I think, I mean, the model is multiple gigabytes.

01:24:39   I think memory is more important than storage.

01:24:41   Yeah, but it's got to be stored somewhere, right?

01:24:44   Yeah, I mean, they've gotten way better at compressing models down in terms of size.

01:24:50   Yeah, I mean, if the model is huge, then sure, I think they will. I think the RAM is more important because you've got to have that model in memory.

01:24:58   And they're getting better at compressing models down. Just following Whisper, I think the Whisper models keep finding, they find new ways of making them smaller and smaller.

01:25:06   And the semantic index, I don't know how big that is either. I mean, anything is possible. I think Apple is probably more concerned with RAM than storage at this point.

01:25:14   But, you know, we'll see. I don't think Apple wants to get caught in a case where any Apple intelligence capable device can't do Apple intelligence because its storage is too small.

01:25:25   I think they don't want to end up there. So if that means that they will raise the storage limits in addition to the RAM, I'll just say, though, that Apple loves its margins.

01:25:34   And if they have to choose, I think they're choosing the RAM over the storage because they'll just try to get you to pay for more storage.

01:25:40   Yeah, it might mean that for at least a little while as a user, you may want to decide to jump up one tier. But we don't know yet.

01:25:49   But the thing to say is it is storage that wasn't used before, right? However much it is, right? So if you were going up to the limit, maybe you will want to bump up.

01:25:59   I realize I forgot what I wanted to say in regards to Brad's question, so I should jump back a little bit.

01:26:05   So this is in the idea of AI as an OS assistant. We don't know if this is going to work, right?

01:26:14   I feel like me and you over the years have seen many instances of Apple saying, "Hey, look at this thing that we've done."

01:26:21   And then when you get it, you realize it's not as powerful, right? And shortcuts is such a great example of this over the years.

01:26:27   Or people don't adopt it and it doesn't get developer support, so it doesn't move on, right?

01:26:32   But if they are able to pull it off, this is exactly what I want, right?

01:26:37   And we spoke about this leading up to this. The ability for me to be able to issue commands to my computer and my computer can action those and use third-party apps to do so is incredible.

01:26:51   If I'm able to look at a photo and say, "Can you make this look the way that I want it to?" and it understands that what it wants is to open darkroom on my iPhone and apply a standard filter that I've created.

01:27:04   You know what I mean? That kind of understanding and action would be incredible, but we're a long way away from that to knowing if they can actually do that or not.

01:27:16   John wrote in and said, "You expressed disappointment in Apple for profiting off the platforms it created by taking 30% from developers.

01:27:25   You also expressed disappointment in Apple for profiting off others' online works by using them to model their AI engine.

01:27:32   Why should online content creators make money when Apple uses their creations for AI modeling, but Apple not be allowed to make money when developers use their platform and APIs? Is this not a contradiction?"

01:27:44   I'll sum this up just in case anybody didn't follow because there's a lot going on here.

01:27:51   So John is saying that we criticize Apple for taking 30% from developers, but we don't.

01:28:01   I think I will say maybe he's issuing this more to me because I am the more heated I think about this sometimes, that I say that I believe Apple make enough from selling their iPhones and all of their other things to not need a full 30% and I don't think developers get enough from that.

01:28:18   So he's saying that why should content creators not make money then? Why are we saying that it's okay for Apple to take it?

01:28:28   For Apple to steal, but yeah, it isn't a contradiction, John.

01:28:34   And the answer is because what happens, first off, Apple makes a lot of money selling iPhones. Developers make a lot of money selling apps.

01:28:42   The complaint is really about the cut and also many of the rules.

01:28:47   I mean, the cut is less relevant for a lot because for most developers it's now 15%.

01:28:53   It's a very limited number where it's still at the full 30 and also that's a partnership where you sign an agreement with Apple as a developer and we think that the partnership is sometimes unfair, but it's a partnership where both parties are making money and building their businesses.

01:29:09   Apple, and I don't feel as strongly about this as Mike does, but I'll just say it here.

01:29:14   Apple going out on the internet and taking all the content that everybody does and taking it back and building features that are going to sell iPhones is not a partnership.

01:29:22   There's no partnership there at all. It is Apple strip mining the open web for money.

01:29:28   And so no, John, it's not a contradiction. They're not comparable at all.

01:29:33   Thank you. I mean, look, the thing is that in both of these scenarios, the way that they are comparable is Apple is exerting its power to make its products more desirable to sell more.

01:29:46   Yes, exactly.

01:29:47   That happens in both instances, but developers at least have an agreement.

01:29:51   Right. Never forget that. And I'm not saying that it's bad, right? Apple, we live in a capitalist economy.

01:29:56   Apple is a profit-seeking corporation that's a public company. Apple, I've said this for 30 years now, Apple is not a charity.

01:30:05   Apple is not a religion. Apple is a profit-seeking company. They are allowed to make money.

01:30:11   But to say that, how dare you criticize Apple for taking money and doing other things to developers, and then criticize Apple for stealing from people on the open web, because aren't those contradictory?

01:30:27   No, they're not remotely contradictory. It's a difference. There's no consent happening. There's no quid pro quo at all.

01:30:35   It's literally all of the AI companies are just going out and taking your stuff.

01:30:39   And again, I have limited issues with it that I delineated earlier, but it's not the same.

01:30:45   It's not the same because we don't have a relationship with Apple. People who put web pages on the internet do not have a relationship with Apple.

01:30:51   That would be great. What if Apple came to me as somebody who is in their search engine index and said, "You know what?

01:30:57   We found a way that we're going to start paying you for all of the work that you did that feeds our model."

01:31:03   That would be an interesting thing they could do. They're never going to do it, but they could do that, and so could open AI, and so could Google, but they're not.

01:31:10   And that's the difference.

01:31:12   Yeah, 0.000001 cents for if your iPhone's sold. I don't think Jason would turn that one down.

01:31:18   You know what I mean?

01:31:19   I'll take it.

01:31:20   That's some good money. I don't even know how much that is, but I bet it's good money.

01:31:23   Ava says, "Do you think there will be any kind of opt-out for the generative AI features of Apple Intelligence?

01:31:32   Personally, I hope there is. I don't like it in general, and I don't want it taking up battery, CPU, or storage space on my machines."

01:31:39   I think "Don't use it" was probably the... right?

01:31:43   Yeah, I think you'll probably be able to turn off some of it.

01:31:46   Yeah.

01:31:47   And yeah, I don't know what form the model is going to take. Are they going to download it? It may take up some storage space on your device regardless.

01:31:54   That may be the case.

01:31:56   There's stuff that's taken up storage space now. Your spotlight index takes up space, right?

01:32:02   There's stuff in there, and the semantic index is going to be kind of like that.

01:32:05   But I don't think it will take up battery and CPU if you don't use it, and probably some of it you'll be able to turn off.

01:32:15   And Justin says, "While I understand the technical limitations of Apple Intelligence being limited to devices it's limited to,

01:32:22   don't you think it will be a poor user experience for those who are a mixture of some older devices?

01:32:27   Siri will behave completely differently depending on what device you're using."

01:32:31   Yeah.

01:32:32   Yep.

01:32:33   This is going to be awkward for a while for a lot of people, right?

01:32:36   But this is one of those things where, in part, it's Apple pushing this forward faster than it might have otherwise,

01:32:42   where there'd be a larger portion of the install base that was on the latest.

01:32:49   If they had waited a year, right, like John Syracuse would probably have an Apple Silicon Mac, but he doesn't right now,

01:32:53   and so this stuff's not going to work on his Intel Mac Pro.

01:32:56   But also, I'll just go to -- I'll refer to my previous statements about Apple being a profit-seeking company.

01:33:01   I think this is a great opportunity for Apple to sell new hardware, saying you need the new hardware for these features.

01:33:08   And they're happy to do that.

01:33:10   Yeah, I do believe, though, that the cutoff being where it is on the iPhone, I think, speaks to the fact that this came quicker than they were expecting.

01:33:21   I know this is not an original thought, but I agree.

01:33:23   My feeling on it is, look, they know they would have sold devices regardless, I think, right, for this.

01:33:32   I believe that they would have preferred to have had at least a couple of years of iPhones that could have run this system.

01:33:41   I think -- and yeah, this definitely came up on a podcast, it might have been your podcast, but --

01:33:46   and I agree with the idea completely that one of the pieces of evidence that this happened faster than Apple expected

01:33:52   is actually the lack of support for Apple intelligence in the regular iPhone 15s from last fall, and probably from the iPhone 14s and 14 Pros as well.

01:34:06   Even if you just went to the iPhone 14 Pro, right, like even if you just only went that far, like, I am convinced that they would have wanted to go that far.

01:34:14   Totally.

01:34:15   I think this may have -- I think I may have heard an Apple executive say this, I don't remember at this point, but they didn't limit it to the M4 iPad, right?

01:34:24   It's any M series iPad will do this.

01:34:26   Any M series Mac.

01:34:28   I think it just shows that they had a technical limitation with the iPhone specifically, but I am --

01:34:35   I think it's a shame that this is the case and it's disappointing, but I do not believe that this is a ploy to sell iPhones.

01:34:43   I don't believe that's the case.

01:34:44   No.

01:34:44   Because they will sell iPhones.

01:34:46   iPhones will be sold.

01:34:49   It's fine.

01:34:51   And maybe they will sell more iPhones, but this is just going to be one of those things where all they'll do is pull forward demand, which I think Apple doesn't want to do.

01:35:00   They know the problems that are inherent for them in doing this.

01:35:04   That if now they sell more iPhone 16s this year because of Apple intelligence, that only hurts them in two years.

01:35:13   It's not like you will be able to sell a new iPhone every year to these same people.

01:35:19   You may just be pulling forward a bit more demand because of it, right?

01:35:22   You know what I'm saying?

01:35:24   Like every time they had this happen, where they have this spike for some reason, they end up having to answer to it and earn these calls for two years.

01:35:32   Yeah.

01:35:33   I think they're doing it for lots of reasons and I think that if they could, they would --

01:35:38   it shows you the length of time that goes into an iPhone cycle because I think they would make different decisions about the iPhone 15 now.

01:35:47   Now they were kind of stuck because of the three nanometer process thing, but I think they would have made different choices if they had known that this was coming in this timing back when they were developing those phones.

01:35:59   And they didn't.

01:36:00   I feel certain that they will all support it this fall, right?

01:36:04   Yes.

01:36:05   But it didn't happen last time.

01:36:07   Yeah.

01:36:08   It's going to be interesting to see how they handle it.

01:36:12   I also wanted to mention a thing that came up last week, again, which is one thing Apple's not doing is taking stuff that some slower devices can't do and putting them in the cloud.

01:36:29   Right now, or at least when this ships, what Apple has told me is if it happens on device, it happens on device everywhere.

01:36:39   If it happens in the cloud, it happens in the cloud everywhere.

01:36:43   It's possible that if they had wanted to relocate more stuff to the cloud, they could have gotten some older devices on there, but I don't think so because I think the model that runs that determines whether it'll go to the cloud or not itself needs that power to do that.

01:37:03   But it's possible.

01:37:04   They're like, "Couldn't they have put some of those models in the cloud?"

01:37:06   It's like, "I don't know, maybe."

01:37:07   But they decided to draw the line here.

01:37:09   Now, going forward, I would not be surprised if there are things that happen on device.

01:37:14   And this will maintain, to Justin's point, this might maintain more device continuity going forward across the platform.

01:37:20   In the future, they may introduce Apple intelligence features that can run on device for newer devices and run in the cloud for older devices.

01:37:30   And that is not happening yet, but it could happen in the future.

01:37:34   And that would mean they don't work as well on the older devices, but they do work, and then you get to create more platform consistency.

01:37:42   But you really want the M6 MacBook Air because it can do magical whizzy thing locally.

01:37:50   And the M5 MacBook Air can only do it in the cloud.

01:37:55   Maybe. We'll see.

01:37:58   Thank you to everybody that took the time to write in.

01:38:02   If you would like to write in to talk about this with us or anything else, go to upgradefeedback.com.

01:38:08   You can leave your questions, your follow-up, your Snell talk, anything you like in there.

01:38:13   You can check out Jason's work over at sixcolors.com.

01:38:16   You can hear his podcasts at the incomparable.com and here on Relay FM.

01:38:20   You can listen to my shows here on Relay FM and check out my work at cortexbrand.com.

01:38:25   Jason is online. He is @jsnell, J-S-N-E-L-L-L.

01:38:29   I am @imike, I-M-Y-K-E.

01:38:31   If you'd like to see video clips of this show, you can go to TikTok, Instagram and YouTube where we are @upgraderelay.

01:38:37   Thank you to our members who support us every week with Upgrade Plus.

01:38:40   This week we're going to talk about a horrific internet mishap at Jason's house.

01:38:45   You can go to getupgradeplus.com if you would like to sign up and support the show.

01:38:50   Thank you to our sponsors this week, KRCS, Tailscale and DeleteMe.

01:38:54   But most of all, thank you for listening.

01:38:56   Until next time, say goodbye, Jason Snell.

01:38:59   Goodbye, Mike Hurley.

01:39:02   [Music]

01:39:13   [Music]

01:39:14   Thank you.

01:39:16   You