Upgrade - 629: Jony Ive Loves a Handle

Episode Date: July 20, 2026

From Relay, this is Upgrade, episode 629. It is July 20th, 2026. Today's show is brought to you by FitBod, Decagon, and Squarespace. My name is Mike Hurley, and I'm joined by Jason Snell. Hi, Jason. Hi, Mike. It's good to be here. This is a good time.

We've been doing a lot of things. But you know what? As we discussed before the show, we don't have time for throat clearing and small talk. We do not. here at the upgrade program, we get right down to it. We get down to the Snell Talk, and I have a question from Sims who wants to know, Jason, do you have plans to see The Odyssey? And have you read the book slash poem?

I see people referring to it as a book. Some people refer to it as a poem. I think people that are referring to it as a poem are trying to show they've read it. I think that's what you say if you've read the Odyssey. Jason, what are your feelings about all of this? It's a poem in book form that I read in college. Jason gave it away. One.

Ravell College, Humanities One. I think humanities one might have been humanities two five quarter sequence about Western civilization nice um probably more tough probably not like that anymore but in the in the late 80s and early 90s that's what it was like anyway yes we did read the Odyssey and uh or at least all of it some of it I don't know I don't remember that far back I remember that there's stuff in it that I was also on the screen at the Odyssey the film by Christopher Nolan that I saw on with my family on Friday night. Okay. So, uh, the answer is yes. Uh, or I guess the answer is no. I have no plans to

see the Odyssey because I saw the already. Yeah. Although actually I do have some plans that are forming to go see it in an IMAX 70 millimeter because I would like to see it again. So you do. In a gigantic format in which it was shot. Uh, however, I will, I will reveal here that in San Francisco, they're really good IMAX theater that we've got, uh, currently only has show dates out to August 12th. And they're all. all sold out. And over the weekend, they were showing it at 6 a.m. and 3 a.m. in order to pack in more people to see the Odyssey. So Odyssey, mania has struck. But I'm hoping to maybe see it in mid-August if I can get a get in there at a reasonable time. Yeah, I haven't seen it. I was actually

sharing this if you on Friday when we were recording, designed in California, that I felt like I had to make a decision as to whether I would see the Odyssey or Spider-Man, because those movies come out very close to each other. And right now, things like going to the movies requires a level of negotiation in my life. It's very, it's very difficult. I'm excited about the Odyssey. I want to see the Odyssey. I need to see Spider-Man on close to day one as I can make it, because that's just what it's like if you really care about Marvel movies, which I do. The best thing to do is to try and see them as soon as possible so that they don't get spoiled for you. I will, I mean, I hope that maybe I can do what you're doing and in a few weeks' time go

and see The Odyssey, I reckon it will be in IMAX for a long time. Yeah, I think so. So that's a good thought. I might take a look at that. The first movie that we saw as parents was the original Toby McGuire Spider-Man. And it was literally, can we do this? And we took the baby. Yeah. Yeah.

And we made it work. It was very much like the baby may need to nurse during the runtime of Spider-Man to remain silent and fall asleep and it totally worked. And we saw it and it was very successful. And that baby is now our video editor. So hi, Jamie. Hi, Jamie. Adina took Sophia to a few like, they do baby showings, right, where it's like you can just bring your baby, but they're not kids movies. So when I saw, I think I said this in the show, but when I saw F1, I saw it in IMAX with underscore David Smith, Sophia and Edina were watching F1 in a more like, local cinema at the same time.

The same time. That's nice. That baby slept unbelievably through what I think is the loudest movie I've ever seen. Obviously they turn the volume down a bit. But if you were Adina even whispered your voice, she'd wake up immediately. Yes. Because that's how baby sounds work is that they don't care about that car sound. They care about their parents.

Yep. If you would like to send in a Snell Talk to help us open a future episode of the show, just go to UpgradeFeedback.com and send in your Snelltalk. I have some follow-up. Henry wrote in and said, the 2mm increasing thickness for the iPhone 18 pro rumor that you spoke about last week has been debunked based on the dimensions of the 18 pro dummy models. The current 17 pros are 8.75 millimeters thick and the iPhone 18 pros are rumored to be around 9mm. I think the person who reported this wanted to communicate that it was 0.2 millimeters increase

and instead said two millimeters. This is one of these things that when I read this, I was like, oh, yeah, this was something that went into my head and I just never fully articulated it. And now here we are. And this makes a lot more sense. That thicker, but not like 25% thicker.

No, you know, we bunked that and they debunked it. It's good. It's good. It's good. The bunking and the debunking is an important part of the process. Get out of my bunk. Get out of the bunk. Get out.

Yeah. You take time. top bunk and then we're going to push you out. Reuters is reporting that the Chinese government has finally approved Apple Intelligence to be permitted for use in the country on Chinese iPhones. There is no version of Apple

intelligence, kind of original or current that is running. Apple is going to be working with models made by Baidu and Alibaba to be compliance that are using kind of like different models for image stuff and different models for kind of the general

Siri AI. And so So the Chinese intelligence authorities have apparently said this is going ahead now. I don't know if it's Siri AI. I think it may be for the Apple intelligence as we think of it from two years ago. That's a good point. I think about that. Yeah. But I would assume that Siri AI will eventually follow probably once they've got this all established. They're going to want to use official sanction models in China and get that up to speed. So I think this is literally a two year later Apple intelligence thing. But I do think that. think over time, it's probably going to be the case that Apple will have an AI framework with China where they'll know what models they use in China. Yeah. There's a line in this writes a story that I find funny, which says the development could help

bolster Apple's position in China where consumers have been waiting for the rollout of Apple intelligence. Now, I have multiple issues with this sentence. Yeah. You know, there are a lot of business articles. Yeah. that you can see the writer or the writer's editor

in a lot of cases is like, please put this in the wider context of Apple's, for example, Apple's business in China. And so they're like, uh, uh, this might make Apple's business better in China

because people might want it. And like, is that, but you got to justify your story. Like, why is the story relevant at all? Um, and so they're,

they're like, oh, well, it's relevant because people, waiting for Apple intelligence on the iPhone in China. It's like, A-Bey. I mean, they keep selling iPhones in China.

The 17thor line seemed to have done really well in China, right? So it's not like they're having a China problem. They're actually doing great in China right now. And I don't think anyone is waiting for this version of Apple intelligence because it essentially has nothing. Yeah, this is the Reuters equivalent of when Bloomberg inserts

that line three hours after they run a story that says what the stock price did after they posted the story, implying that the story somehow impacted the stock price, which it almost certainly didn't. My favorite one of these, we may at some point in today's episode to talk about the OpenAI smart speaker rumor.

In that article, they referenced Sonos' stock price dropping. That's what I was thinking. It's like, on this report, Sonos price went down and Apple's price. And then the best line is, not only do they quote Sonos, and I roll my eyes so hard, like,

oh yes, I'm sure you, you're the ones who did that. Then they quote Apple stock price, and they're like, Apple stock price basically didn't do anything. I'm like, whatever. Whatever. The company that this article was about did nothing to them and a completely unrelated smart speaker company that don't have really much in the way of an assistant, their stock went down. In the wake of publishing this article, gold went up slightly. It's a great good for gold. A butterfly flapped its wings and this article was produced.

Apple Music is going up in price. It's now a dollar per, a dollar more per month, and it's pulled the prices of the Apple One bundles accordingly. Yeah, I'm going to be paying more. I'm an Apple One bundler. I'm going to be paying more. It's more. It's $3 for family.

So it's $1 for student and individual $3 for family. And since family has rolled into the bundle, the bundle is then going up, what, $2? So, yeah, for family and premier. Oh, yeah, because they're like, maybe there's three people in your family. Pay up. There's an article, I'll put in the show notes, that kind of has all the prices and stuff in by Chance Miller at 9 to 5 Mac.

And Chance has a line in here, which it kind of enforces to me the never-tweet kind of aspect of modern life. It says here, the price increases for Apple Music come after Spotify announced its own round of price hikes earlier this year, which Apple poked fun at on social media, which they did, right? So this was when Apple, when this happened a few months ago, Apple music tweeted BTW, we're still the same price,

as is in February. And it's like, I get that that social media manager, like they're just doing their job or whatever, but it's like now you just look silly, right? When you post things like that and then a few months later,

you're just doing the same thing. Like, this kind of like dunk culture is just, it just never works for you in the way you think it will. I don't disagree that that's probably not the way to approach these kind of announcements is to mock your competitors because you'll probably be there too eventually. I would also say, though, nobody remembers what happened four months ago and nobody cares. So it's also that one of those things is the gotcha. The dunk culture is dumb, so is the gotcha culture.

Exactly. They actually go hand in hand of each other where ultimately it doesn't do anything for anyone. Every dunker gets got. Every gotcha gets dunked. D dunking and debunking. That's what we Yeah, debunking and dunking. That's what this episode is all about. And,

and de-dunking, which is very important to. So, yeah, if you don't get, if you don't get got, you're getting dunked on. Yeah. And it's like you're punching your own,

yourself in the face. So just don't even, just don't even do it. Just don't do it. And I have some lawyer up follow-up as well. Oh, wow. Yeah, I know. It's a new vertical here on the show.

There's so many levels to this. Do we need new show art? Do we need follow-up show art that then, contains what we're following up inside. I don't think we do. No, I don't think we do. Move on. Let's move on. The Financial Times is reporting that Apple has now issued, quote, legal preservation letters to around 40x Apple employees who have joined Open AI. Essentially, these letters are telling these people don't delete anything because this will all become part of legal discovery.

Can you imagine, Jason, how it will feel to get one of those letters? Like, oh, no, I'm in it now. Not to be confused with that weird NBC story that was like, hey, Apple sent an email to the wrong guy. It was like, what, what? Why is that anything resembling? And then they didn't. And then they called, then they emailed the right guy. Oh, well, this is, how did the markets move on the release of that NBC news story? I wonder. How did it affect NBC's stock as well as open AIs and Apple's? We will never know. No, we'll never know. And in unrelated Apple legal news, great. I know. Mark Gohmann is reporting that Apple is in the early stages of a discussion with the U.S. Department of Justice about a settlement to the antitrust case brought against them in

2024. I will read from Montgomery. Apple has made multiple offers this year to the Justice Department to bring the case to a close. The activity has picked up in recent weeks with the two-size exchanging draft settlement materials. Yeah, this is, first off, it's the Trump administration, and it's a thing brought during the Biden administration. So my default is that maybe the Justice Department won't, you know, will.

not, we'll take it lightly, actually, because that seems to be what they do with everything. A lot of this stuff is, I mean, again, not to keep saying this, but I rolled my eyes when I was reading what was covered in this suit, because to remind you, this is like the Apple harmed competition by not having super apps. Yeah. That was one of the big points in there, which is first off the super apps thing is kind of dumb. And then they said, no, we actually, we do off, we do allow that now. And they made changes to the app store policies to kind of reinforce the mini apps thing. Yeah. And they support RCS now, which was part of it.

There's the Apple Watch availability. But like, I mean, this was not the most interesting case anyway. And now I suspect you've got a Justice Department that cares even less about it. So if I were Apple, I would also be making some offers to say, can we wrap this up? I think it's also a Justice Department that, this point is well known for loving a deal. It's true. It seems like the current Trump Justice Department will take a deal.

So if you've got a deal to give, they will accept that deal. And it seems like they're going to deal with the deal. This episode is brought to you by our friends over at FitBod. When you're looking to change your fitness level, it can be hard to know where to get started. And that's why I want to tell you about FitBod, because it is an easy and affordable way to build a fitness plan that is made just for you. because you have your own path to personal fitness.

So FitBod will use data to make sure they customize things to suit you perfectly. FitBub will adapt as you improve, so every workout remains challenging, pushing you to make the progress that you're looking for, because it's important to get pushed but in the right ways, because your muscles improve when they work together. If you overwork some or underwork others, it can negatively impact your results. So FitBod tracks your muscle fatigue and recovery to design a well-balanced workout routine for you. What I also love about this is you get new exercises often, so you're not going to get bored,

things are being mixed up. I find myself demotivated when my exercises are very similar. So FitBod's really great in that it gives you that. It kind of mixes stuff up, new exercises, new circuits, that kind of stuff. FitBod tracks your muscle fatigue and recovery, so you can avoid burnout and keep up your momentum, and it builds your best possible workout with the use of exercise science. They have analyzed billions of data points that have been fine-tuned by certified personal. trainers. The app is great. It has one in a thousand demonstration videos, so you can see exactly

how to perform every exercise. They have progress tracking charts, weekly reports, and sharing cards, integrates of your Apple Watch, you're aware of a smart watch, and apps like Strava and Apple Health. Personalised training of this quality can be expensive, but FitBod is only $15.99 a month or $9599 a year, and you can get a huge 25% of your membership by signing up at FitBod.me slash upgrade. So go now and get your customized fitness plan at F-I-T-B-O-D. dot me-sash upgrade. That is FitBod. com for 25% off your membership.

Thanks to FitBard for their support of this show and Reli. So last week, we were... Oh, by the way, it's time for the details. Woo-hoo! You almost missed a segment there. I know. Can you imagine? Last week, we were...

We were ready in case the public bases dropped. We were. We were so ready. and they dropped like two hours afterward. And then they came out before we got the episode published. But we're going to talk about them. You have focused your beta impressions on Mac OS Golden Gate, right? Like that's kind of where you've focused your attention.

But what's great now, I think, in this, I would say in the last few years, but also in this year's 27 releases, the biggest features are available on all platforms, right? so like Siri AI, some of the photo stuff. But what has been your impression with Golden Gate so far? Well, I think it's good. I

had to, of course, preface my whole thing by saying, look, I liked Tahoe. I know so many people are like, I'm not even running Tahoe. I don't even own a television. And I whatever, I found Tahoe to be a really good update in the sense that

Apple actually added productivity Mac features that like Mac, the Mac is a place where we want to be productive. And for years, I felt like Apple didn't really focus on that. And they did some Mac specific productivity features that made me really like that aspect of Tahoe. Yeah, the design regressed, although I think it was not as bad as it could have been, which is I chose to look at that glass as half full, which is if they had devoted more time to liquid glass on the Mac, I think it would have just made it worse, not better. So I'm kind of glad that they only gave it a light touch and then ran away. That all said, Golds and Gate rolls back. You know, I would say,

most of the egregious changes that were in Tahoe to the point where I think that, you know, somebody could go straight from previous OSs to jump right over Tahoe and land in Golden Gate and you would be fine. The, like the sidebars metaphorically were right in Tahoe where it was like, well, this is a paint of glass on top of a paint of glass. But the net result was a mess. And the sidebars now make sense again. Toolbars are toolbars again instead of it being sort of like an air.

where individual buttons can appear with content sliding under them that changes based on what contents and all that is gone and it's just a toolbar now. So like the interface stuff is a lot better. And so I think that's good. I mean, all the windows feel a little square,

which I think was the intent to get the roundedness of it to be something that makes a lot more sense. So I thought that was all, I thought, you know, so I guess if you're, somebody who has held out because you're like, oh, I just don't like how it looks. I think you could go into Golden Gate even now and certainly in the fall and not worry about it

because the look has been cleaned up a lot. And you're going to get the benefits of all those productivity features in Tahoe that you haven't had for the last year if you said, I'm not going to use Tahoe. So that's good. Did they change Spotlight much? I mean, that was one of the biggest productivity changes, right? With Spotlight and the different features that it brought. Has it changed much with Golden Gate?

It looks totally different because they're picking up the Siri AI style. Okay. And it's got one of the engines in it is that it's detecting if you're asking a Siri question. And then it flips over to ask Siri. But otherwise, it's the same in that it's got the multiple categories. It's got apps, actions, files, and the clipboard history. And that all seems pretty much the same as last year. So again, if you're coming from two versions ago, it's a pretty dramatically changed spot.

because they did so much with it. Obviously, the underlying spotlight index is different now because they're using the same index that they're using elsewhere. And so how that will play out, who knows? Like, it's a long process to index. And, you know, I've found it to be pretty successful. In fact, my biggest complaint is at this point, if I do use it as an interface for Siri, I feel like I almost have to give it a minute to realize that I'm asking Siri something. I'll type something in really fast and then I'll just sort of sit there until it's,

turns to ask Siri. Sometimes that can be a little bit sluggish. But it is something. Joe Rosenstiel wrote a piece on Sixth Colors like, I don't know, last year, that I'd never really thought of it, but he made a really great point, which is that Apple has too many places in its interface where you can type things in. And it's like, why do I have a place that I type to Siri and a place that I typed to Spotlight?

Why do I not have a place to type? And that's one of the things I really like about the Siri AI Spotlight integration. is command space type of thing. If it's a search, it'll search. If it's a Siri query, Siri query, it will, it doesn't quite rhyme. It will, it'll do that. Like, that's how it should be. I should not need to think about which box I want to bring up

in order to type a thing in to find out an answer, right? And then, as you talked about recently in your beta experiences, as a question engine, Siri AI seems fine. It seems like they did. We tend to downplay it because we've been surrounded by all these other LLMs and that so much of the LLM conversation now is about these heightened things like writing code and agents that are doing work for you and all of that. But so much I think of what AI is actually used for is essentially what used to go in search engine queries, which is, can you give me the answer to this question or can you give me a list or suggest this or something like that? And in my experience, all that stuff works pretty well. Like, they did that.

They solved that issue, which is why I think it may be as smart that companies like OpenAI are deemphasizing their simple chatbots and focusing on their complex agents and code building and things like that. Because it really feels like if Apple and obviously Google with Gemini have nailed the table stakes of, I just want to ask a simple question, that's not, I mean, people still use chat GPT and stuff. but like it's going to defang a lot of that functionality because your phone's just going to... You're not going to need to take an extra step to get there, right? Yeah, I have... I mean, I don't know if you want me to do this now. I have a bunch of thoughts about Siri AI.

Like I could share them now while we're kind of on the subject. Sure, we're in a beta train. Let's do it. So I am very impressed by Siri AI. Like I've been using it for a bunch of tasks, I guess some small, some kind of medium. and on the whole it performs well. I would say tasks that I consider to be large tasks,

I'm still defaulting to something like Claude, right? It's like... Of course. Vibe coding. Or if I'm like even doing things on my Mac, like file-based stuff, I'm still using Clore for it. And maybe some of that stuff will change

when I install Golden Gate. I'm not using Golden Gate on any devices yet. I don't really like to play around with a Macbator. They scare me. Honestly, like I... there's too much going on there. But in general, my hope was that Siri AI would kind of take the place of general knowledge or simple research stuff, which is what I was using chat GPT for.

And it's absolutely done that for me in my experiences over the last kind of four to six weeks or whatever. I want to give some examples for the types of things I've been asking that I think are interesting. because it still requires a little bit of massaging in places. I've mentioned this one on the show before, but I have a repeating reminder for putting my laundry away. And I've been quite impressed by just kind of like out of the blue, I can say to my iPhone, I put my laundry away. And it understands that and then just marks the reminder as complete.

Which I just think that that kind of interface is the kind of interface that you would want. I didn't need to say mark my laundry task as complete, right? Like that kind of like very specific way that we talk to computers. You don't really need to do that. And I think that's quite clever. And it's more of a natural dialogue. And I've done the typical thing that everybody's doing, right, which is like asking about a trip, that kind of thing.

But you could just say like, how is the weather looking for my trip? And then it knows where I'm going. It gives me the temperature. I was actually quite impressed with this. It gave me what the temperature was today. I asked this like a week or so ago and it was like, I can't tell you what it is going to be in that time. It's too far away.

And I think it's a clever result, right? Like it could have just said, like, no, I can't do that. But it's like, no, here's the temperature in New York today. But I can't tell you what it's going to be like in two weeks or whatever. And then I've been doing some kind of like the typical trip planning stuff with it. And I've been asking like, oh, give me some suggestions for restaurants. and what I said was, within a 30-minute walk of where I'm staying, I didn't tell it where I was staying, but it knows, right, it knows where I am, and so it just does it.

I don't need to say, within a 30-minute of this hotel or this street, it just knows the answer to that question. And, you know, like every AI tool, there are the things that you need to give it prompt-wise. Like, for example, when I've asked any question around food, it gives me answers from Apple Maps, which is not what I want. I want it to go out online and find me recommendations based on articles that are talking about the best places and such and such place. And so what I'd say is, look online and it then gives me exactly what I want. Because I just don't really think that reviews on Apple Maps or Google Maps help you find the right. places to eat.

Right. So it's like a thing of just say, oh, and check this on the web, and it will go away and do that. I asked, when will Toy Story 5 be coming to Disney Plus? I feel like that's a pretty good question. And it said it, and I love the answer is that I don't have an answer yet, but based on the way that Disney releases their films, I expected to be a 45-day window, so probably sometime in August. Like, I thought that I found this to be a really great answer where it's using knowledge.

the answer doesn't exist because Disney haven't said but it's using context to try and give me that answer. This is a question I never would have asked the old Siri because it's like there's no way you're going to give me

the answer that I want. And I've done stuff like at the moment we're like coordinating stuff for designing California and I'm working with the people that bought our sponsor tier and I said for it to list me

the sponsors information I have in my inbox for designing California. And it then pulled me a text list of all of the emails of all of the people, then create at a table with their names, the company names, and the sponsorship details. I was like, wow, that was, I would not have expected this kind of stuff from Siri. But yeah, and then in general, the on-screen awareness stuff is great. Who is that? Where is this? What is that? You know, things you see online, like people, places people are eating on social media. Like I've had this.

I'm like, where is this place? And it takes a screenshot, goes away and finds it. And I thought it was pretty good, right? Like all of this stuff is really good. So all of this to say, just reset your expectations of series performances in the past. Like, forget all of that.

And go into it and just try it out. And I think you'll be surprised. Like I've found for me that my kind of typical usage of Chad Chubit has been replaced with Syria AI. And it's also just integrated so well into the system. It has the personal context. I think it's doing a good job. Like I've been very, very pleased of it.

Yeah, I find the personal context to be much more hit and miss. But again, I'm willing to give it the benefit of the doubt at this point in the summer. Yeah. I think some of the personal context stuff is difficult. Like I'm having okay or fine reactions to it so far. But I think we've got to wait until third-party apps are in there as well. And I think, or then what happens, you know, it's going to be interesting to see how that kind of stuff rolls out. But in general, you, I got from your review that you just haven't really been that impressed with Syria AI yet.

Yeah. I mean, again, the world knowledge stuff is fine. Like, I think that it's, I think it does the job. I think my challenge is it's really hit and miss with the stuff. that is based on a personal context. And I'll keep using it and I'll keep trying it. Sometimes it does it. When it works right, it's like magic.

But then there are times when I ask it things and it just doesn't seem to understand what I'm saying. And is that because of the system? Is it because the index on the particular device that I'm asking is not complete? I don't know. I don't know the reasons why. So I'll keep an eye on.

Yeah. what do you think about the photo editing stuff how much how much time have you spent with those tools oh quite a bit actually um the you know all of them have a generative element right like extend it's very cleverly done extend is literally the crop tool but the crop tool goes the other way which is just a very simple concept what if you could uncrop an original image and expand the horizon of it to the point where I think in a very clever move not only did they do that but like all the aspect ratio controls remain so like if you've got an image and you want to make it 16 by 9 you can do that and then it'll extend in exactly the space that it needs to I think that's smart. However, every pixel you add is generative. Every pixel you add it's going to upload that data to private cloud compute and it's going to crank on it and it's going to turn imaginary pixels to you. And that's why I advise people to be judicious with it. I think that it actually is a great answer to the problem I've certainly had, which is I want to print this and put it in a frame or I want it to fit in this particular shape in a calendar. And it's just not the right shape. If I put it in that space, it cuts off the sides and it cuts people off. And so I can't use

that photo. And so with extend, you can say, I'm going to make that photo a little taller. And it does it. And generally, it does a great job. It really does. You know, the warning is you got to look. You got to, you got to look because it may do a bad job. So you got to, and Apple's tools actually do a pretty good job here. When the detail comes back from the server and you see the new generated result, it's not, it hasn't saved it yet. There's like a button to say, okay, I'll take. it or redo it. Let's clear that. And then maybe you can redo it or you can give up. And that's good, because sometimes it's very bad. I had a photo of Julian at his graduation, and it was taken literally on the floor where they did the graduation. There's a place that they stand and they take your

photo and then families get that photo. And so it was a little too narrow. And I wanted to just be a little wider. And I widened it with extend. And so he's standing there and it's just him and his gown and his cap with the tassel hanging down because he's a graduate. Yeah. And it generated a metal table next to him. And hanging from above dangling down like a lamp or something was an enormous tassel. Oh, like a tassel the size of a Muppet. It was like a horror movie about graduation where the tassels descend from the ceiling and kill you. And I thought, well, I understand why it did this. It saw a tassel and thought, oh, I guess this is

tassel land that this person is in. Maybe there's a giant tassel over here. And I was like, nope, that's, nope, not it. But I did it again. It did it right. So again, it's like, it happens, but you've got to be careful. There's still going to be mistakes. But I think extend, used judiciously, is really smart

because you're taking a photo that might have been the perfect photo, but unusable for the application and making it just a little bit more. and with good, usually the results are amazing. So, very impressed with Extend. I think it's going to help people. The combo of using Extend and then recropping the image

is a pretty good way of doing things. Like I've had it, you know, for me, the best use of this is wallpapers. I already, on my iPhone now, my luxury wallpaper was a wallpaper I made by using the extend feature. And I, you know, basically I extended it and then recrruct it to cut off

a lot of the stuff that it would have generated about the people in the picture. And it's just like behind them, they've just added more wall. Right? And it's like, that's perfect. It's exactly what I need for a luxcreen image. And so I've been really impressed with that.

And also the cleanup feature is much, much, much better now. Yeah. As far as I can tell, there's two versions of cleanup. There's the old version of cleanup and there's the new one that uses private cloud compute. Which is the one I've used.

And there's actually in the interface, it describes them as there's auto where it chooses and then there's sort of like fine and high quality and a fine seems to be the on device relatively fast

one that was there before and then the high quality one because you see the progress bar when you try to do it that's it's sending data out to private cloud compute and getting the result back but I did a test because there's a photo in my book

about photos that is cleaning up, it's Jamie standing on the beach in Hawaii, and there's two guys standing behind her. And they're like overlapping right behind her, and she's got her long hair, and so they're overlapping. And clean up tried, right? And it's okay, but you can see it's kind of rough around the edges,

and you can see that there's something back there. And the private cloud compute, the high-quality version in the current betas, did a vastly superior job to it. I had to put it in high quality mode to do that, but it did a great job. It's not always perfect. But you can also, that cleanup is also a thing that can be progressive. So I did a cleanup where I took, I was at a baseball game a couple weeks ago,

and we got to do this thing where we were out in the outfield before the game on the warning track. And I took a selfie. And then there were people next to us, and I cleaned them out, right? It said, let's get rid of these people. It's great. But it made an assumption that there were these, like, barriers, like blocking off. the rest of the compartment we were in that weren't there. But then I just cleaned up the barriers and they went away and the result was then good.

So it was like, you know, that's how it's supposed to work. As you say, no, no, no, not this. And then it gets rid of it. And again, sometimes that doesn't work very well, but sometimes it works really well. The reframe feature that uses the spatial. I guess what I should say about that is don't move it very much. because I'm sure there are applications where moving it a little bit.

I think that their suggestion that it's for things like, I didn't like this thing that was right behind the head of one of the people in my shot. And I moved it a little and now it's gone or now it's not behind their head.

I think that's a good use of this feature. But don't do too much because it has to invent everything. And I was taking pictures of people where you pivot it. Like, there's a picture of me and I was like, let's just move it over

to the right to the left a little bit or to the right and what it does is it reveals the left side of my face that wasn't in the shot and it's like my face has a head growing out of it like a second head and it's like it's grotesque and awful and it's like that's that's a problem especially with faces every pixel that is being revealed of that face is being generated and that means you're going to get synthetic face and our brains are really good at noticing that that's not that person's face Now, will there be a future version of this in a few years that looks at who that person is, recognizes other source photos of them

and your photo library is able to send those as reference and generate something that is completely realistic? Yeah, probably, but not this one. This one doesn't do that. This one is entirely out of context. And so, like, be really careful. But I find extend to be a much less useful feature than, or, I mean, spatial reframe to be much less useful than extend.

I think extend and clean up are the winners here. Yeah, I think I've, the more I thought about it and the more I've played with it, I'm struggling to find a good use of reframe that can't just be taken care of with cleanup. That's true. Just clean up that thing that's behind that person. Yeah. Otherwise, I don't know what it's really for. Refram is a cool magic trick, but it feels very much like what we said at the time, which is, oh, I see, they've assembled this feature out of other features that they have. Yeah.

Because that's exactly what it is. They're using the spatial. Photos feature, which allows them to build a depth map. And then you can pivot it around in real time because that's what they do on Vision Pro. And then you can send the result to a server and do generative on it because it's going to be able to tag the stuff that changed and regenerate the image. So I get why they were able to build this feature. It feels a little more forced. Like this is a feature we could build so we built it.

Yeah. It's got the feeling of image playgrounds to me. But just not as bad. to be like, I don't feel as strongly about this as I do about image playgrounds, where it's kind of just like, you know, what's the phrase I'm thinking of? Like, you know, just because we could

didn't mean we should. Yeah. Kind of, every pixel in the reframe, it's not like extend or cleanup. Every pixel in the reframe is regenerated by an AI model. Yes. Because the perspective has changed. Yes.

And I think in some cases when, we definitely saw this, there was somebody sitting with us who was doing some of this at WWDC, and they were pointing out how uncanny it could be. And I think it's because, in part, the 3D scanning depth model

is not always that sophisticated. I think that they, it's kind of like, oh, this is like a mannequin head, basically. Like the head goes here and then the body goes here and then the background is back there. And the problem is, people's faces aren't mannequin heads. Like, the nose is a different shape. The chin is a different shape.

The hair is a different shape. All of those things are true at different depths. And I think maybe that model is not quite that sophisticated. And as a result, when you rotate it, it's rotating and calculating a new face based on probably some pretty simplified geometry. And the result, if you know a face really well and you look at a face and reframe, you will often go, it's not quite right. That's not quite right. And I think that that makes a much harder feature to use. Yep. What have been your experiences of shortcuts so far? Obviously,

the Mac gets the kind of generate a shortcut, but also has new tools. that other platforms don't get. Describe a shortcut is very good, right? I mean, that's, that's, it's simple, but it's iterative. I'm very impressed with that. I think they did a good job. I hope for growth there, but it's a good place to start. Yes.

Yes, very much so. Because, you know, it's even people that know what they're doing. It's like a good way to get started, and then you can very easily go in and change things and see how it did it and stuff like that. I think that it is a rare feature that benefits both regular and power users. I think everybody can get some benefit from this, which I think is nice. You don't need to know more than to say, I'd like this thing to do this thing and have it do it.

And then you test it and you go, yep, that did it. And if it doesn't do it, you type in the box and you say, no, no, no, I want it to do work like this. And then it changes it, it rebuilds it and then does it again. Yeah, there's otherwise if now for people who are using that. it's got storage now so you can save storage and it's persistent over time and it can update it, which is really great. That makes it much more functional. And then there's a new three-paned interface on the Mac that is totally useless because the third pane is just that giant shortcut summary that is pointless. And then when you edit it...

Well, at least it means that you don't have to swap between the views, right? So you just got that benefit, I guess. Yeah, you can't. Instead, you have to double click and then it opens the shortcut separately. So, but there's no... I don't know what you're thinking of. I don't think there's anything that is Mac only. Oh, yeah, no, you're right. You're very right. I guess the three-paint interface is the only thing. Yeah, it's just kind of pointless because it's, yeah.

I mean, I get why you, it'll work on any shortcut. So it's using its AI model to try to describe shortcuts, even using doing things that it can't actually generate itself. And that's kind of fun. But in reality, it takes up in a huge amount of space and doesn't do anything. So I would like to be able to close it, please. Yeah, but otherwise it's good. Like, again, it's a very simple feature that is, it shows the way to the future more than anything else, I think.

Speaking of Way of the Future, you know, obviously you've played around with Sidecar. It's very clearly the harbinger of touchscreen Macs, right? Yeah, and the story about Sidecar is that it's boring because it just works. Like, you attach an iPad via Sidecar and you start using the Mac interface with Touch, and it just works like you'd expect. it to work. You can scroll and you can select and you can move windows around and it's fine. Like that's for all of the hype about like, oh no, touchscreen, Mac will have to be redesigned for touchscreen. Not really. It just seems to work fine. Like, it just seems to work fine. So I wish I had more exciting things to say about, about a sidecar. That is the most exciting thing in my opinion.

that it just works perfectly fine. Like, I don't want macOS to start, like, shooting fireworks, right? Just because there's a touchscreen. Like, I don't want it to do anything weird. Like, just do what iPad does. If I'm using a MacBook Pro with a touchscreen, I expected to do exactly what it does on my iPad, which is I'm sitting there,

my hands aren't on the keyboard. I'm looking at something I want to scroll, and I just idly reach out and flick and scroll the web page. And I do that on my iPad all the time when it's in the magic keyboard, and it works fine, and it's just not a big deal. That's what I think this experience should be like.

And as far as I can tell from sidecar, that's exactly what it is. Like, you're not, like, this is a secondary input method. This is not, for people who are not, you know, and I know several people who are like this, who have like a mouse on one side and a trackpad on the other side. Jason, in front of me right now, I have three input methods. I have a webcam tablet, a magic track pad, and a mouse.

Exactly. So for people who are not used to that concept, it might seem weird. But like having an additional input method, method is not a big deal and is sometimes nice ergonomically. I do not anticipate that people are going to be navigating their Macs entirely by touch. It's just not going to happen. It's just not going to happen. Maybe in a pinch you could do it. But like it's just not going to happen. And that's okay. It doesn't need to happen. It's fine. It just works.

Okay. Doing exactly what we've been doing with iPads with keyboards attached for like five years now. Yeah. Six years. Something like that. I would assume it's similar with Golden Gate to iOS 27, which is that it's all performing very well for me. It's all very stable. And there are just tons of little improvements throughout the operating system. Like I think people are going to be happy with these releases just from that perspective. I think especially when they launch to the world later on this year, I can only imagine them feeling even better than they currently feel, which is very good. Is Golden Gate the same?

Yeah, they're bugs. But I'm, I mean, that's, my standard public beta advice applies here, which is, if you want to see what the future is and you want to mess around with it, I think you can do it and it's pretty safe. But just know that those of us who use betas have come to accept the fact that it's a beta. That means some of your apps might crash, or they might not work right. Like all the menu bar utilities are broken. Yeah, actually, the Mac kind of has a different. different struggle area, right? Because you've got the app store and non-app store stuff.

Yeah, my iPad is, I mean, my iPad is draining battery more than it did before the beta. But like, yeah, on the Mac, you just have, like, there's third-party apps that get weird. And there's bugs. And, like, I restart my Mac more when I'm in a beta because things are working weird. And I'm like, okay, let's just restart. And then usually that solves it. I've restarted my iPad a couple of times in the last week. because things got weird.

And I was like, all right, let's restart it. So it's not, you know, I'm not giving you a free pass here. But if you're willing to run into those issues and say, well, this is what I get for running a beta, I think you run the public beta and you're going to be fine. And you get to try out all this new stuff. If that turns you off, then don't. And you can wait until September or longer if you want to. I will give you a freer pass for iOS 27.

I think iOS 27 is feeling really good to me. And it doesn't have a lot of the. edge cases that macOS 27 will have for what will and what work let me just give a quick aside on watchOS 27 jason which i have now recently installed because i was having some issues uh with notifications on my apple watch so i upgraded and it fixed them i think that there was something was getting lost a couple of things so syria ai is on the apple watch um it's not on all apple watches actually no it is on the ones but WatchOS 27 is not available on that many Apple Watchers.

I like having the app there. I think you can ask complex questions to get answers. I like that you can read the full text of previous conversations. I just think that's a nice thing you can have on your Apple Watch. But I've noticed some of the more basic stuff

like setting a timer. Yes. Takes a little bit longer than it used to. I have seen it take longer and I have also seen it fail. Yeah. And it also fails at adding things to my reminders list. Okay. I do one of the things I use my Apple Watch for is add sour cream to

the shopping list. Yep. And it and it says, I'm sorry, there's no list called shopping list, which there is. There is. And then it says, would you like me to make a list? And if you say yes, it fails because there's already one there. And if you say no, then suddenly my default list has sour cream on it, which is not a shopping list list and it's very baffling. So I mean, this is this is why I have not had I think watchOS is the worst of the betas because I think that they did break

large parts of Siri and it I can't recommend it to people No I wouldn't I wouldn't they've got to tighten it up it's busted because I think

say with the timer stuff I think it's waiting for more information like it's like is this a big request they're asking me and like it takes a like there's a delay

yeah got to address it. And also, if I set a timer, you don't need to wait for another, like, it's still waiting. Like, you know, it sets the timer. And then it's like, I'm still listening. It's like, you don't, it's fine. You don't need to keep listening. So I think they've got to tighten that up. Actually, you mentioned the shopping list thing was just another great experience I had where Adina texts me a bunch of things that she wanted. And I just held the button down. It was like, add all of these to my shopping list. And then it just did it. And it also did it in all the, the pre-made shopping

this category too. This is on my iPhone. I thought it was great. The new app layout thing,

so when you press the digital crown and it gives you the apps, I like this UI actually because it

is surfacing the stuff that I want. And so I think as a, as a maybe fourth or fifth try

of trying to work out how to show you apps on the Apple watch, I like this one. And also,

I felt like in all of liquid glass Ui, the Apple Watch was the worst. Like it looked the worse than anything else from my taste. And I think they've done a lot of UI refinement in 27. And I think it looks a lot more considered now where it felt like to me with WatchOS 26, they were like, let's just make everything clear.

And there wasn't really any consideration about it. And I feel like it's looking much better. But my battery life on my Apple Watch Ultra is destroyed. It's so bad. So, but it fixed my main problem, which was I wasn't getting notifications on my Apple Watch, and it's one of the main reasons I wear my Apple Watch, so it fixed it. But this one will get better. I think once it does get better, I think this will also be a pretty solid release.

Right. And I believe the way Siri AI works is that it's talking to your pair of iPhone. So it's going back to the iPhone for that because I can't do that itself. It's too tiny, too dumb. But I feel like if I tell you set a timer and like do it. Well, actually, I don't know if it's. The way I set timers, I don't know if you do this or anybody else does this, I just say five minutes. That's it. I don't give it anything else. I do this because you showed it to me. This is the power user tip you gave me when I was in London.

And I don't know, I just say an amount of time. It just sets a timer. And I don't know if maybe that is a problem, right? But that has worked for me forever. But here's the thing with all of this. And this is true for the iPhone, but it's especially true of the Apple Watch. If we get Siri AI on the Apple Watch, but it comes at the, expense of all of these simple commands that make life easier because I'm just adding sour cream to my shopping list or I'm just starting a five minute timer. If those get bad but you get Siri AI, they blew it because you got to do that stuff right. All that simple stuff has to work and it has to work well and it needs to not spin for 15 seconds before it starts a timer. Just do you can't do it. If you can't get that to work, revisit what you're doing on the Apple Watch because that stuff has to work. Yes, I agree. Especially something like that, like this, the very, very simple table-stakes

stuff for, like, why you'd wear an Apple watch, you quickly setting timers while you're cooking, right? And just being able to speak into the watch, I get, that needs to be tightened up. Don't, don't over-complicate this, right? And this is, it's interesting, I've not had this experience with the iPhone. Like, doing things like asking it to set timers, change lights, you know, all that kind of stuff. That is as quick as it ever was. But, but, But it feels like in trying to shove this technology into the Apple Watch, they've maybe added some kind of latency that wasn't there before. And that's not great.

This episode is brought to you by Decagon. The best customer experiences happen when customers feel like they're the only one, almost like they have their own concierge. But once more or more requests start to come in, a lot of companies turn to AI and lose that personal touch. But Decagon helps your support feel personal, no matter how many requests. you're getting. Decagon helps companies create personalized concierge-style customer experiences with AI agents across chat, email, voice, and SMS. They're available 24-7, feel natural to

talk to, and can resolve customer requests on their own, so businesses can keep up with requests without losing the personal touch that's important. Decagon plugs directly into your existing systems so agents can get live quickly, and because workflows can be updated using natural language, teams can make changes themselves about long engineering cycles. This means less time spent managing workflows, more time focused on customers and ROI in weeks. Once live, agents can pull up to-date information, take actions on requests, and understand the context behind any conversation. Plus, Decagon gives your team full visibility into why agents make the decisions that they make and what's happening across every conversation so it's easier to spot trends and improve experiences

over time. Decagon helps power millions of customer conversations every day for brands you know and love, including Avis, Affirm, Fanatics, and Ura. Are you ready to transform your customer support? Go to Decagon.a-i-slagrade to get a personalized demo and see what Decagon can do for your team. Check out Decagon at D-E-C-A-G-O-N-A-I-S-Ugrade. That is Decagon.com.A-I-S-Ugrade. Thanks to Decagon for their support of this show and all. of Relay. Rumor roundup time, Jason Snell.

Yeah. So this one is a slight aside to the typical rumor roundup fair, but considering all of our discussion about Apple and OpenAI, let's talk about a report from Mark German at Bloomberg about what OpenAI's first hardware device is likely

to be. So Mark is reporting that they are working on a smart speaker without a screen. It is intended to be a, quote, new type of home computer for the AI era that is designed to be a human-like AI companion that lives in your home.

It is, I mean, I guess it's probably I don't know what smart speaker means in this context. Is it going to be canister-like? Is it going to be a little ball, like a little pebble? You know, what is it going to be like? It will have a rechargeable battery so you can move it around, but it's not designed to leave

your home. So that tells me it's a Wi-Fi product, right? You'd guess from that. Because if you take a Wi-Fi product out, it's dead. It's got to be on the internet all the time, so it's going to be on your Wi-Fi. And it's presumably not via a Bluetooth connection or whatever, you know. Yes.

To a helper app on your phone, no, it can't be. It's a Wi-Fi device, for sure. As well as being something that you talk to chat GPT with, this product will also play media and control smart home appliances. I mean, sure, right? What else would it do? Sure. OpenAI expects that this device will become increasingly personal.

to a user and their preferences over time, even offering proactive suggestions. This is a quote, Open AI envisions the device anticipating needs, surfacing information proactively, and serving as an expert on its user, they said. If you remember, Jason, a long time ago, we spoke about one of Mark Gohmann's report

for the robotic arm product. The tabletop robot? Tabletop robot. And in that, he was saying that, like, over time, it may start just like saying things. You know, like, hey, you, you're wrong. You know, like, I don't know what it's going to say. What do you think about this idea?

Like, imagine that the technology works, right? Like, and they start making proactive suggestions. If they are useful, proactive suggestions, would you want a piece of technology to just start talking? Well, I think having the option is okay. I have lamented for all the time I was using an Amazon Echo, that there's no way for me to, like, on a scripted basis, sent an alert to have my...

I echo say, you know, you're exporting solar, you ought to do some laundry or something like that. Yeah. Like that would, it's like a push notification. And it would never do it. It's like, okay, there are scenarios where having your in-home assistant ping you, uh, in a place other than your phone and say, here's a thing that you might want to know. That's different than I'm listening to your conversation and, and budding in. And it's different than what Amazon poisoned all their.

products with, which is like, by the way, I have a sale on laundry detergent. And it's like, go away. Leave me a lot. I'm going to rip you out of the wall and throw you in the garbage can right now. Our relationship is over. That's how my relationship with Alexa ended. Same. Me too.

By the way, I have things to sell you. And it's like, nope, gone. But goodbye. Goodbye. So, but I mean, I think there are scenarios, but the question is just like how, look, so much of this description is like people spitballing in a somewhere in Silicon Valley saying, yeah, yeah, yeah, and it could do this and it could do this. It's like, okay, but what do people actually want it to do?

But yes, like, I built that whole thing a couple weeks ago just to alert me on my phone or my iPad that I left the EV plugged in and it was going to charge overnight because this time of year, I do the EV charging during the day using overage from solar and it's free. And if I do it overnight, I will fill up the battery, but it will be off the grid and it will cost me. So I have an alert to say, it checks to see if the car is plugged in and says, there's a car plugged in. You might want to unplug it. And it runs that at 9 p.m. But I have to have that on my iPad or whatever.

And I have to be paying attention to it. That's the kind of thing where I might say, why don't you ping me at 9 p.m. and say out loud whether if this thing is true. Like there are scenarios in which you might want somebody to interrupt you like that. Yeah.

But I look at this and I think, is that what they're going to do? Or are they going to do a whole bunch of dumb stuff because they don't understand what people actually want? Which is probably the answer here. But like, I'm not opposed to it.

I think the idea that you can't send or at least when I last used it, you couldn't send a push notification to an echo because there was no API for that. It's like, no, no, you can't do that. It's like, that seems really dumb. I should be able to alert myself with an automation I built somewhere.

So I could see it or using AI to say like, you know, if you say, why don't you warn me when so, if so-and-so ever emails me? And it's like, oh, did you know you just got an email from so-and-so? You're like, oh, no, I better go check that. Okay. If I choose to do that, having my little roaming speaker person thing tell me, about it, we'd be fine. So like, best imagine, you know, you're here in mega studio with me and we're talking about the rest of history again. And we're like, oh, such and such thing happened in 1874.

And then this device was like, oh, actually that was 1875. How would you feel about that? I wouldn't like that, right? Like, I don't want a thing kind of correcting me. Never. Never ever do I want, never ever do I want somebody, uh, some AI model budding in on my conversations. Yeah. Because I worry about that kind of stuff too. Or like me and your say, what do you want to go for lunch? It's like, oh, why do you try this do pizza place? No, you need to leave me alone. By the way, there's a new pizza place, Mike.

Because these are also the kinds of things that I imagine a proactive assistant could be doing. And I just don't know about that. No. Mark says that it could also include a mechanical element that will move to show emotion

and response to a user. I like the thought of this. I I think that could be fun. Yeah. Like, that it, if this thing, this was a conversation again

from the tables hot robot, like, if it wants to be proactive, kind of raise your hand, right? You show me you have something that you want to say. And then maybe I'll ask you

to give the information. Like I like that kind of thing. Or if a device, if I start speaking to a computer, it alerting to me that it knows that it's to receive this input, I like. We have that visually, right? Like, if I start talking to the assistant on my phone, it lights up to let me know. But I have just kind of like this little cylinder on my desk, you know, it kind of like raising an eyebrow essentially to let me know it's listening. I like that as an element for a device like this, I think. Sure. It's just ambient alerts, whether it's color, a little light somewhere, or whether it's movement or a different position that it usually is in. I'm not. I mean, that could be fine.

Sure. I mean, that's the, I think that's the idea of the tabletop robot and what we saw in a bunch of those demos to those, those demo videos that that Apple AI engineer was working on where, you know, you're basically, uh, you've got a little thing that can move. And so it can have a personality and it can be like a little, again, tabletop robot. We, we, we laugh about that. But that's the idea is to have it have personality and feel like it's, uh, alive on your countertop or, you know, wherever you put this thing, the same idea. I'm not opposed to it as an idea. It feels very gimmicky and it might be dumb, but I see the impulse there and it could potentially be interesting.

And, you know, I feel like there is this whole vibe of like trying to bring humanity to technology. And giving us a sense of life physically to these products is a way to inject that feeling, right? In the way that like every time anybody gets a robot vacuum, they follow around the first time they think it's hilarious. They laugh at it bumping into things. because it does more pet-like stuff, right? It's like it's moving around your house,

and it's like a fun novelty. Apparently, this device will also have a camera and other sensors to gain an understanding of the environment that it's in. Yeah, I mean, sure, right? That feels like it would have that. This is one of my favorite things in this article is this quote. OpenAI believes that this device veers significantly from anything Apple has on the market today

and that it is unlikely that it violates trade secrets belonging to the iPhone maker. I love the significantly. And so carrying on from this is, though Apple sells the HomePod and HomePod mini smart speakers, OpenAI doesn't see them as comparable to what it is building. The new speaker's audio system and other hardware differ substantially from Apple's designs,

and OpenAI believes its first product doesn't violate any trade secrets, according to people with knowledge of the matter. Good news, everybody. We didn't violate trade secrets on this one. Come on. I mean, all right, sure. You didn't steal anything related to this. But also, really, how different is it? You know, like, how different can you be? No, the real counter argument is what they don't want to probably say, but was my reaction to this entire report, which is, get ready, everybody. Open AI has finally made some hardware. And what have they made? They made an echo. They made a home pod.

This product is not interesting. This product is a battery-operated Wi-Fi speaker like you can get from all sorts of different companies attached to an AI subsystem, although directly, so more like an Echo or a HomePod than something that would be paired with a phone. Okay. And it might have a thingy that moves on it or something,

but like everything else is like, oh, it's a new type of home computer for the AI era. Come on! In what scenario is this a new type of anything? It's not a new type of anything. The only new thing in it is it'll be attached to chat GPT. And, you know, yes, chat GPT is very interesting. And they've got that new GPT live where it is like even more kind of like a back and forth

and it's listening and reaction to as you talk. Like they're building this whole thing up, almost like Disney imagine, imagining style to make it a richer experience, fair enough. But like, as to what this product is, as described, it's not even remotely interesting to me. And nothing about it is new or significantly different. But I will grant you that this

stuff is so old that I'm certain it doesn't violate any trade secrets. Because I don't think there's a secret, I mean, short of the tabletop robot moving around a little bit and that maybe Apple could claim that if they do some stuff that the mechanical element on this, that that is lifted from aspects of what Apple's doing. But I believe them. I, I, I completely believe over an AI that this is probably not a product they needed to steal any trade secrets to release because it's not that interesting. I feel like they are going to have to go to great lengths to try and show me why this is better than using my phone. Right. Like I have chat GPT and GPT live on my iPhone if I want it. Right. Like, it's really,

you know, this like proactive and the physical nature of this device is going to have to be very appealing. And I'm just not convinced that it would be, honestly. This whole hardware thing that they're doing is because OpenAI has a has an AI platform and doesn't have any place to put it. Right? Because they've got Apple in the way. They got Google in the way. They, they, the most important interface is the, is the phone, the iPhone and and Android phones. So Apple and Google and Samsung and you know, You know, you can list some people, but there's like in the way. And Google and Apple are not going to turn over and hand it all to chat GPT, to open AI, right?

They're not going to do it. So what do you do? And their answer seems to be, well, we've got a lot of money because we've got a lot of investment. There's this huge AI bubble. We're just going to pour money into making our own hardware. And, you know, honestly, building your own Amazon Echo line that's based on chat GPT and substantially better than what, because let's be clear,

the smart speaker market is, the AI smart speaker market is terrible. Like, there's nothing out there that's any good. None of them do a good job. And so there's an opportunity for them to do something there where they build hardware that allows them to embody chat GPT

and get it in people's lives and get it in people's homes. And good for them. But, like, that's what is going on here is that they're stymied by hardware, being the place they want to go to get this stuff in the real world and having the companies

that are running the operating systems running the hardware don't use their stuff and they don't control it so they're going to build their own. Like I get it. I get it. But this is an uninteresting product

in any other way. I bet this product has a nice handle. I bet right? Like some way. Sure. Because I think it's going to be intended that you pick it up and take it around the house with you.

Johnny, I loves handles. And he loves a handle. handle. And I think I am intrigued, I am intrigued about this because I just want to see what, what are they designed, right? What does I.O. Johnny, I have loved from, whatever mixture it is of this company. Like, sure. What is the, what is it? Right? Like, I think that's going to be interesting. And I, I will want to try it for that reason as like a piece of hardware. Like, what did they make? It could be good. It could be good because all of these are so bad now that it could be good.

Yeah. Sure. And apparently this is the first of five products that the company is working on. They're still aiming to release this one in 2027. And as Marx is obviously one of these devices is intended to replace a smartphone, whether it's a smartphone or a lot. But we'll see. We'll see about that one if they will ever be able to get it to where they want to get it.

Yeah, I will, yeah, we'll see what they're going to do here. I understand why they're making this product. I mean, the challenge, too, is that their competition is also getting better. Now, chat GPT is going to be better, but what I said before about how I'm not going to use this thing to like, I'm not going to use it to build software. I'm just going to use it to ask questions and remind me of stuff. And that means that although OpenAI undoubtedly has a better AI subsystem than Apple does, let's say. or Google or Amazon. For a product like this,

are they really going to blow the Apple HomePod screen Siri AI thing out of the water? Which we'll probably see before this, right? And that will have a screen, which is its own massive selling point for a product. Indeed, because people do like to see stuff.

And it will be a beautiful photo frame, right, that you'll be able to have in your house. all the time of all of your favorite photos on it from your photos library. So if I were then, I'd build, this is the most sensible feature or hardware to build for them, is you look at what Amazon has done

and all the others in this category and you're like, their AI isn't very good, their ecosystem isn't any good, and we've got the best of those things. So if we could build a box that embodies this that people can carry around and put in their house,

that would be great. It's like, sure. I mean, I don't just, dispute it. It's not groundbreaking in any way or interesting in any way, but I do think it's a product they should probably make. Because again, I mean, unless you find a partner, I mean, you could you could just buy a partner or work with a partner to make this product for you, but they've decided that they're going to hire Johnny Ive and they're going to do it. So like, good luck to them. I think this requires less of an ecosystem than other things. It still would

require an ecosystem. And I think it's probably going to require things like you adding a Gmail skill and having it be able to read all your email and your calendar and stuff in the cloud so that it'll work. Whereas Apple is obviously going to just tether this thing to iPhones and stuff and be able to look at your personal context that way. But still, you know, yeah, this is,

I think far more reasonable than them building a smartphone and having to rebuild the entire concept of a smartphone interface using just open AI stuff. I think that's going to be a harder sell for them. Yeah, absolutely. All right, this episode is also brought to you by our friends over at Squarespace. They are the all-in-one website platform

that is designed to help you stand out and succeed online. Whether you're just getting started or scaling your business, Squarespace would give you everything that you need to claim your domain, showcase your offerings for a professional website, grow your brand and get paid all in one place.

Talking about getting paid, so you can set up on Squarespace a store. You can sell physical goods, digital goods. It's a really great system. We've used it. Brad Dowdy, my co-host in the past, He has a shop that he runs through Squarespace. They have everything a customer is going to need.

You know, it's very easy to make your purchases and the UI and the, you know, like the website looks great. As a seller, it has everything you're going to want. You get all the information you need and a bunch of great integrations with other platforms and services that you might use for kind of shipping the products out. But you can also offer services with Squarespace. So from consultations to events and experiences, you can showcase your offerings of a customizable website designed to attract clients and grow your business.

They have built an appointment scheduling, email marketing tools and more. You can keep everything cohesive of on-brand invoices and get paid easily with online payments. You can also upload content and you can organize that content together and showcase it on video pages, for example, but you can also sell access to this content by adding a paywall, which is perfect for online courses, exclusive tutorials and premium workshops. Go to Squarespace.com slash upgrade and you can sign up for a free trial. Then when you're ready to launch your site to the world, you just use the offer code upgrade to save 10% of your first purchase of a website or domain.

That is Squarespace.com slash upgrade with the offer code upgrade and you will get 10% of your first purchase of a website or domain. And you'll be showing your support for the show. And thanks to Squarespace for the continued support of this show and all of Relay. It is time for some ask upgrade questions. Logan says, I'm currently on the developer beta of MacOS 27, and while it's been great so far, I think Siri AI needs long-term memory.

Yes, it can access the information you have, but what if you want it to remember something that isn't in a stored, in a note, a text or an email? Do you think Apple could add memory to Siri AI in the future? Could? Sure. I think should as well, but whether they will, I don't know, but I hope so. It's nice.

Yeah, maybe, or maybe. maybe it's framed as you can ask it to store stuff and there's like a Siri category in notes. Yeah. Right? Because every device has notes and notes syncs via ICloud and that allows you to store it in a way that it will know about it and it will remember it. Because it's got access to all of that stuff. So I wonder if there might be a way to do that, that even if it's like you remember this and it'll be like, well, I added a note in the Siri category to remember this. And that's how it does it.

But I agree. I think that that's a useful thing to have is, you know, remember this thing when I ask you about it later or take it into account when I ask you about it later. That's a good feature to have. Yeah. Federico is playing around with this by like just adding a bunch of like custom instructions to a note. Yeah. And it would work, but he would sometimes have to ask it specifically to reference that note. Because the request that you're making, I guess the system doesn't necessarily know it needs to access that specific set of instructions to be able to answer.

answer it. So you can add it to a note and if it's like something complex, you can say a reference the note, blah, blah, blah, and it will do it. But that's not as automatic as you would want where if it was in a kind of custom memory, even if that memory was stored in Apple notes like by Siri, say, it could then know, let me just check all of my memories before I... There's an efficiency there though, because then every query is checking the memory and that might be, that might be unwieldy over time. Yeah, maybe. that's part of the challenge. I don't know how that stuff works

if I'm being honest, right? But I think that's what people call the harness. I don't know. People keep saying this word harness and I kind of like go like, uh-huh, a nod and like, I'm not entirely. You put it on a horse, Mike. Oh, that's what I'm missing.

Yeah, you miss the horses. Yeah. Thomas says, I upgrade my iPhone every year and I can't use the iPhone upgrade program because I am on Mint Mobile. It's not a carrier that Apple currently offers in store.

Why do you think Apple doesn't offer unlocked iPhones on the iPhone on the iPhone upgrade program? So I have a theory about this. I don't know if you have one too. I have no thoughts about this at all, so go for it. I figure they do it so they don't upset the carrier relationships.

Yeah. That they don't... In the iPhone upgrade program, they want you to get a new iPhone every year, but they don't necessarily want to circumvent the carriers. And so doing it in partnership with the carriers.

It means the carrier stays involved in the relationship with the customer, which I think they want. And Apple do have to kind of keep the carrier happy. So they don't. Makes sense to me. I mean, the answer is something to do with carriers. It's either with the existing carriers they have relationships with, or it's with the carriers that don't want to participate at that level with Apple, right? So the other way to say this is Mint Mobile offers all sorts of discounts and stuff.

And maybe one of the things they're doing is not paying Apple. or whatever, and as a result, they're not in this preferred zone, so they don't get all these other features. I don't know if this is worldwide. Maybe in other countries you can, because it's less restrictive. I don't know. The iPhone upgrade program sounds like a great thing that terrifies me if I was that person. Because I know that Dan Moran is in the upgrade program, and the reason I know this is because I think almost every year Dan talks online about problems he's having upgrading his phone. So this is why I know I don't want to be in that program myself.

Phil Wright-Sinner says, this is, I think, a great question. I was excited. This one came in. Given the impending increase in price for the iPhone 18 line, is it worth buying a 17 pro now to replace my iPhone 14 pro? The battery on my 14 pro is at 81% of its original capacity, and I was planning on getting an 18 pro in September, but I'm not keen or maybe paying 1,200 pounds for I say yes. I say you should just go buy an iPhone now. I think, and so I wanted to mention this like general advice, so maybe you've thought about this and maybe you haven't.

If you are someone who upgrades their phone every few years, you should be upgrading your phone right now. Because how, I mean, I'm sure the 18 line is going to be better than the 17 line. But if you've already waited years, don't wait years and pay extra prices. Like, especially if then maybe in a few years' time, Maybe the prices do come down again. There's no way we can know right now.

But I think if you are a person who waits, I think you should get a 17 pro. That would be my, I would recommend this to Phil. I recommend this in general. I think that it would really suck if you had to pay 20% more when you're probably not going to get 20% better fun. I mean, it's hard to know. There's rumors of big camera upgrades.

But if you're coming from a 14 to a 17, you're going to get a better camera experience anyway. How much better will it be with one more iteration? I don't know about that, for being honest. I think there's a good deal to be had here and given what's going to almost certainly happen in the fall and you're coming from a 14. I don't think you have to wait.

If you're that worried about paying more, you can lock in at this price and you're going to get a 17 pro is a really great phone. Yeah, definitely. David writes in and says, did Jamie really love it when you mention her on the show, or is that some kind of inside joke to annoy her also hi Jamie? Yes, she really does, and she pointed out she was visiting us this last week, and she pointed out that I can also use it as a way to tell where in the episode she is

as she's doing the edit, because if it happens at the end, so in this episode we mentioned her at the beginning and the end, so she's going to get me twice, but she will text me and then I'll know where she is in the episode because she does like it and she confirmed she likes it and also when you mentioned her on Connected

last week, two weeks ago and pointed out that she doesn't listen to Connected I actually clipped that from Overcast and sent it to her so she could hear her being mentioned on a different show. Hi, Jamie. Matt says, I'd like to know if designed in California

will cover other companies that sadly didn't make it in the market despite having arguably better hardware and software. Oh, please. I don't know about the arguably part, Matt, we have to say. In particular, I'm thinking... It's an Amiga person writing in. I get it. In particular, I'm thinking of Amiga, which had custom graphics and audio hardware and preemptive

multitasking as an operating system. I think Apple had some interest prior to Commodore buying them, and I feel like there must have been talent transfer given how close the offices were. You know, Designers in California is about Apple. but it's also about things related to Apple. I don't anticipate doing a series about other companies that didn't make it, although anything is possible. I think if I can find interesting Apple connections,

I'll go there. I don't really have any plans of covering something like the Amiga, but never say never. It is very likely that we will cover companies that were successful and how they interact and intersect with Apple, right? Like you've spoken about wanting to do Microsoft and Apple series. which will require some Microsoft's info. I can see down the road in, you know,

several years down the road that there could be a series where I talk about, you know, Atari and Commodore and Tandy Radio Shack and even the rise of the PC clone and what that was in framing it in the context of sort of Apple and computer history. I could see it down the road, but it is tangential enough

that I can't see doing it anytime soon. That's my answer. And also we've played around with the idea of like alternate timeline kind of stuff. Not to create fiction, but like, you know, it could be fun to,

if this had happened, where would we be? Like, and that could be like a, that feels kind of like the summer of funness of designing California. Of like, all right,

so let's play this out. Like, if such and such, what could have happened? You never, all I'm saying is you never know, right? What we might do on the show, because we don't know yet. Right, but I would say that having not released any episodes and having looked at our calendar of ideas, we're not going to do it.

I appreciate being asked to take a podcast about Apple history and not cover something related to Apple at all. Yeah. It's going to be a while before anything like that happens. Sorry to say, Matt. But listen in, Matt, because, you know, maybe in 2035 we'll do a two-part series on a media. Maybe. If we need to.

If we need to. If you'd like to send in a question for us to answer on the show, just go to Upgradefeedback.com or click the link in your show notes. You can also send in your follow-up and feedback there as well. Thank you to our members who support us of Upgrade Plus. You get longer ad-free versions of the show each and every week. Go to GetUpgradeplus.com to sign up. You can find a video version of this show on YouTube.

And I would like to thank our sponsors, Squarespace, FitBod, and Decagon for their support of this week's episode. but most of all, as always, thank you for listening. We'll be back next time with a full summer of fun out of time episode. Yeah, it's true. It's true. So we're looking forward to that. Until then, say goodbye, Jason Snow.

Bye. Bye, everybody.