00:02:29
◼►
And once I switched to reduced latency, the live activities, I no longer have the effect where my watch taps me to tell me that the opposition has just scored a run.
00:02:42
◼►
And then I look at my watch and then I look back up and the batter hits a home run.
00:03:23
◼►
And that means my live activities not only keep me informed about what the giants are doing, but they don't spoil me seconds before events occur, which is, I think, important.
00:03:34
◼►
I wish all streaming services had that.
00:03:36
◼►
Like, I think if you do sports, you should try to offer it because this is like, I don't use any live activities for, there's a, there's an app that I like that has used live activities.
00:03:46
◼►
And I've used live activities for in the past for formula one called box box, but I don't use them anymore because sometimes I get the information before I see it.
00:03:54
◼►
Uh, real time followup from Zoe, by the way, uh, Apple sports now has toggles per team for live activities.
00:04:00
◼►
So I could turn off the giants if I want to, but I, I'm very satisfied with my Apple sports live activity experience and I wish to continue.
00:04:06
◼►
So I can just let the MLB app, uh, be where it is.
00:04:11
◼►
Nice thing about MLB is you can pick any random game and just say, please track this game and it'll track that game, which is kind of cool.
00:04:18
◼►
If you're like, you know, watching a different game or something, but, uh, anyway, yeah.
00:04:23
◼►
And I do like live activities and I really like how they extend to the, to the watch and a big, uh, shout out to YouTube TV for offering that feature.
00:04:30
◼►
Uh, and it's definitely a thing that's starting to happen, uh, where like for the Superbowl, Fox was talking about how they're going to try to do the lowest possible latency live stream.
00:04:40
◼►
Um, because they know that people want that.
00:05:01
◼►
I feel like if the streaming services want people to move to these like over the top services, which they do, they need to find a way to make this better.
00:05:10
◼►
It's not great for people whose internet can support it.
00:05:12
◼►
And, uh, as my friend, Will Carroll pointed out on downstream a few episodes ago, um, well, sports stuff is motivated by gambling and, you know, I'm not a gambler.
00:05:23
◼►
I don't love gambling as a thing in sports, but the fact is a lot of money is in sports gambling and latency is the death of sports gambling.
00:05:32
◼►
Because the people who are betting are betting on events that may have already occurred and you can't do that.
00:05:39
◼►
So the more, the less latency, the better.
00:05:43
◼►
If you would like to send in a question of your own to help us open a future episode of the show, just go to upgradefeedback.com and send in a snow talk question.
00:05:50
◼►
Thank you to Brantz for that question.
00:06:42
◼►
That's interesting that they did that.
00:06:43
◼►
I think they let the, let the developers start to think about, uh, doing their, you know, uh, compiling for Apple Silicon and what the universal binary format was going to be like.
00:06:55
◼►
And, um, I believe, didn't they, they ended up with like a developer test hardware that was essentially like a Mac mini with an iPad, iPad put in inside it.
00:21:29
◼►
The image that 9to5Mac pulled out and they were doing some analysis on the kind of the metadata of it suggests that the screen resolution and size could be close to that of an iPad mini.
00:21:52
◼►
It appears to be continuing with this product because it was considered to be in kind of no man's land with the Siri changes.
00:21:59
◼►
And in the last week or so, there seems to have been a lot of talk online about a session and a new API called Interactive Snippets, which looks a lot like something the home device uses.
00:22:12
◼►
These are little compact pieces of UI that are displayed via app intents.
00:22:18
◼►
So it can show information, allows some interactions.
00:22:21
◼►
They look like widgets basically, but they're triggered from intents.
00:22:25
◼►
So you can trigger them in the new iOS and iPadOS from Spotlight, Siri or Shortcuts.
00:22:30
◼►
And it feels like maybe this kind of thing would be maybe interesting in this scenario too.
00:23:57
◼►
There have been a selection of rumors that Mac rumors have pulled together about the iPhone 17 Pro featuring a vapor chamber cooling system.
00:24:06
◼►
These kinds of cooling systems are found a lot on gaming phones, but other phones like the S25 Ultra.
00:24:11
◼►
Essentially, vapor chamber cooling provides better, more efficient cooling for an iPhone.
00:24:16
◼►
So I'm going to read from their report here.
00:24:18
◼►
It would consist of a thin sealed metal chamber containing a small amount of liquid.
00:24:22
◼►
When the iPhone heats up, the liquid would turn into vapor and dissipate across the chamber's surface area.
00:24:28
◼►
Eventually, the vapor would cool down and condense, allowing for the process to repeat.
00:24:32
◼►
The system would help to move heat away from the A19 Pro chip that is expected to power these models.
00:24:37
◼►
This could indicate a couple of things.
00:27:22
◼►
That is two things that are in absolute direct conflict when it comes to the iPhone.
00:27:27
◼►
But, yeah, I think this year's teardowns will be really interesting, right,
00:27:31
◼►
when we start seeing these devices pulled apart and seeing what on earth they're doing to make them work or attempt to work.
00:27:39
◼►
Mark Gurman is reporting that Apple executives have held talks about buying perplexity.
00:27:46
◼►
Quote, Adrian Perica, the company's head of mergers and acquisitions, has weighed the idea with services chief Eddie Q and top AI decision makers.
00:27:53
◼►
Apparently, these discussions at Apple are a very early stage and, of course, will probably lead to nothing.
00:28:01
◼►
But they would consider it as a way to build an AI-powered search engine along with bolstering their AI efforts in general.
00:28:09
◼►
Basically, this was news to perplexity, which I find very funny.
00:28:13
◼►
But apparently, the two companies have already been meeting separately to look at integrations for search and other features.
00:28:19
◼►
I have a couple of links to some posts in the show notes, one from Parker Otolani and one from Federico Fatici.
00:28:26
◼►
I will give him his full name because I gave Parker his full name, even though I think everybody knows who Federico is.
00:28:30
◼►
Just saying that perplexity is probably not the right product to buy.
00:28:36
◼►
Perplexity is a product-focused company, right, is the argument that they're making, that use other companies' AI models as a back-end to build a nice search experience on top.
00:28:47
◼►
And perplexity will tell you they have lots of really powerful models that are pulling all this information together and da-da-da-da-da-da.
00:28:55
◼►
And I'm sure they have some stuff, but they are essentially building on top of the foundations.
00:28:59
◼►
They've also have done some stuff about building that voice mode thing.
00:29:06
◼►
They have a voice mode in their app, like all AI apps do, but they've actually built Apple's APIs into their voice modes.
00:29:13
◼►
If you use the perplexity voice mode and ask it to create an event for you, it will add it to your calendar because it's using Apple's APIs, which is interesting.
00:29:23
◼►
My kind of feeling on it is like, I'm sure they're talking about it, but I don't expect it to happen.
00:29:29
◼►
Yeah, I mean, so it is suspicious that everybody's suddenly talking about perplexity because that sounds like somebody is trying to boost their valuation or get somebody to buy them and they want to turn up some conversation about it.
00:29:41
◼►
It could be all a coincidence, but I don't think it is.
00:29:43
◼►
I think something is going on here where somebody at perplexity may be looking at their valuation, looking at their burn rate, I don't know, and saying, hey, it would be great if a bunch of tech giants were vying to buy us.
00:29:57
◼►
I think their survival means somebody needs to buy them eventually.
00:30:01
◼►
Yeah, so what's interesting about them is that they have this product that they have worked hard on search.
00:30:08
◼►
I take Parker Orlitani's point that they don't have a model, right?
00:30:15
◼►
They just have a product built on top of models, which is what Apple's already doing.
00:30:19
◼►
So if you're Apple, the only reason to buy them for that is because you feel that there is a lack in your culture of building AI-focused products, which I think Apple's been able to admit a lot of things about itself in the last year regarding AI, but I am not sure that this is one of them.
00:30:39
◼►
The idea that they would really need to feel like, oh, well, we just don't have enough people who are thinking about productizing AI, and we can buy this company for many, many, many billions of dollars because it's going to cost a lot.
00:30:50
◼►
You're paying a premium for an AI company in 2025, right?
00:30:54
◼►
But maybe we need to do it because we feel a lack of that.
00:30:57
◼►
I am really skeptical that Apple has come to that point.
00:31:04
◼►
They've realized that, like, oh, actually, this is one of our cultural problems is that we're not productizing this.
00:31:09
◼►
I will say the one part of this that makes me intrigued is the fact that they have a pretty good search product, and we're in an era where Apple's relationship with Google search is a question, and the idea that Apple may or may not be able to, you know, count on the money that they get from search referrals in Safari.
00:31:30
◼►
So I would be open to the idea that the reason Apple buys something like Perplexity is not because they want, like, all AI product stuff to be better at Apple, but because they could basically pick up an AI search engine, and a bunch of people devoted to the idea of a search engine product, and boom, Apple's got its own AI search engine.
00:31:54
◼►
That it could put in all its products.
00:32:01
◼►
It's like buying Beats to start Apple Music.
00:32:04
◼►
Yeah, so that would be, that would, right.
00:32:06
◼►
And you take, the rest of it comes along with, but that, like, because I know that this is going to be, it would be like a $14 billion purchase.
00:32:20
◼►
If Eddie Q, and that, so that's how I wanted to view this, is there are lots of ways this could be viewed.
00:32:26
◼►
But if it's talking, the report says that the head of mergers and acquisitions has weighed the idea with Eddie Q and top AI decision makers.
00:32:54
◼►
If you're him and you're looking at the current search world and you look at Perplexity's search product, I can see, like, you ask yourself,
00:33:05
◼►
could we build a product like Perplexity inside Apple?
00:33:13
◼►
And I'd put all that in a bowl and I'd look at it and be like, well, what do I think?
00:33:19
◼►
I can see a scenario where if you felt confident that one of your ways forward with search was to have AI-powered search products and to have,
00:33:29
◼►
and you look at Perplexity and you say they're actually really good and we really like their product and we like their culture and we think it would be a good fit.
00:33:34
◼►
If you're Eddie Q, I feel like the money is not even a question, right?
00:33:38
◼►
Like, if the way Apple goes to a new level, shifts gears from the old kind of Google relationship, search is such a moneymaker for them that, I mean, you got to look at that.
00:33:52
◼►
I'm not saying that they should buy it.
00:33:53
◼►
I'm saying that's why you would look carefully at it, is for something like the uneasiness about search and what if you could build your own search product, even if it's on top of other people's models or eventually your models.
00:34:07
◼►
And also, you have to go back to like what is the exact analog of this discussion, Google Maps, right?
00:34:23
◼►
Apple had to build a map service from ground zero and they struggled.
00:34:29
◼►
If Google, for whatever reason, either is forced to, decides to, or Apple decides to remove them as the default or whatever, and we're going to have an Apple default, that's the decision that they decide they want to do, that search product has to be really, really good.
00:34:47
◼►
Like if you thought Maps was a problem, if you take away the, if you change the way that people use Safari, because people just open Safari, they type what they want, and they're taken to a Google search result, right?
00:35:00
◼►
Like if you're going to change that experience in some way, you have to have a darn good search product in Safari to replace it, right?
00:35:08
◼►
I would throw in the other possibility here, which is, you know, what if the pressures of the change in search from traditional search engines to AI means you can't stick with the existing Google deal, right?
00:35:22
◼►
Regardless of legality, that's the other issue here.
00:35:25
◼►
The decision might just become Apple's, right?
00:35:27
◼►
They've even given a hint to that, whether that was, you know, we're not sure exactly what Eddie's reason was to say what he did like a month or so ago, right?
00:35:38
◼►
About the Google search stuff going down.
00:35:40
◼►
There are many reasons he could have said it.
00:35:42
◼►
But if you're Apple, if you're working on that team or whatever, you've got to be paying attention to that and be like, well, we, what if all of our competitors, you know, they integrate Gemini now instead or whatever.
00:35:53
◼►
And we're stuck with, with Google and nobody likes it.
00:35:57
◼►
So it's, it's, uh, the other thing I'll throw out there, because I think that the idea of viewing this as a product versus model is useful.
00:36:08
◼►
Because there are other conversations that people can have about should Apple, how, how confident is Apple in its model versus others?
00:36:17
◼►
And I think the problem is there's no very clear product, company out there that doesn't already have its, uh, big tech partner that's got its hooks into it.
00:36:26
◼►
And therefore it's not going to be something, you know, like Microsoft's relationship with open AI and open AI really believing like they're, they're number one and they don't even need, everybody's coming to them.
00:36:35
◼►
Um, and, and there are others, you know, Anthropic has what, do they have an Amazon deal?
00:36:44
◼►
But I will say if you view AI acquisitions from Apple in the light of, um, improving their culture and, and you think again, fundamentally, you think they aren't on the right path now, not a year ago, but now with their models, right?
00:37:03
◼►
Um, and, and, and, and what I don't want to put this as is you have a choice between, do we think we can build it or should we buy it?
00:37:25
◼►
Because the truth is the choice is, do we think after failing once we now are on the right path or not?
00:37:34
◼►
It's not, it's not like build it or buy it.
00:37:37
◼►
It's like, well, we built it and we blew it, but now we have to build it again or buy it.
00:38:06
◼►
And so they built a Maps system in the background and then flip the switch and kick Google out and Google had to make its own app, which it did.
00:38:13
◼►
Beats though is about, it's not about like replacing necessarily the, uh, the, the existing default with your own default.
00:38:22
◼►
It's more about directions they wanted to go because they wanted to do Apple music and Beats had a music service that was pretty good.
00:38:33
◼►
Like it was human curated and they had it as, and like, it was like a style that Apple has maintained.
00:38:38
◼►
Like there, there was something about the DNA of Beats music that really spoke to them as like lovers of music as institutionally as a company.
00:38:45
◼►
And honestly, the head, the headphone business continues on, but it's like a side business.
00:38:50
◼►
That is just a nice thing that makes them some money and they just leave it alone.
00:38:54
◼►
But really the motivator was, was launching Apple music.
00:38:57
◼►
And that was the case where they, they had to do some cultural, you know, integration.
00:39:00
◼►
And Mark Gurman points out in his, in his piece, like his piece is weird.
00:39:04
◼►
Cause it's like, Apple's gotta buy something.
00:39:06
◼►
And then he's like, but Apple is bad at buying things and maybe it shouldn't.
00:39:09
◼►
I'm like, okay, this is all over the place.
00:39:13
◼►
I mean, he's wrong about Apple's gotta buy something or else.
00:39:16
◼►
Like, I don't know if that premise actually works, but, um, but the key is like, if you're an Apple executive,
00:39:23
◼►
if you're at a queue or, or if, if you're somebody else at that level, what's your confidence in Apple's ability to build something?
00:39:30
◼►
Because Apple has all the money in the world, essentially.
00:39:33
◼►
Apple's got so much cash that it can buy, it can buy almost anything if it really, really wants to.
00:39:40
◼►
So what's your confidence in what you can build internally, culturally?
00:39:45
◼►
Do you benefit from bringing in people who've been thinking about AI all the time from the outside,
00:39:50
◼►
who have are experts in either the product or in rapidly iterating the models, or do you feel confident?
00:39:56
◼►
I think the problem right now at Apple is, do you feel confident when you go to John, Jan, Andrea, and he says, we are great.
00:40:05
◼►
We're doing great because I wonder if they believe him, right?
00:40:09
◼►
Like that, that's a question I have, but that's, that ultimately is the decision.
00:40:14
◼►
It's like what you spend your money on, um, has in part to do with your confidence that that's a thing that you can do yourself versus importing it from the outside.
00:40:27
◼►
And that, that, that comes down to a self-confidence about Apple.
00:40:30
◼►
And that means a level of introspection that Apple executives have to have about their abilities, because there is a time for you to say, I love my team.
00:40:59
◼►
I know you love your team, but really are they up to it?
00:41:03
◼►
And that is brutal because you need, you need an honest appraisal as a senior executive, right?
00:41:10
◼►
You need an honest appraisal of like, are they up to it?
00:41:14
◼►
And happy talk gets, happy talk is a killer for business strategy, right?
00:41:20
◼►
Where, where people are like, you know, I don't, I don't want to say like John, Jan, Andrea again, but like that sort of thing of like, if those stories are true, the idea of like, no, no, no, no, no, no, no.
00:41:35
◼►
Do you believe anybody when they say, oh, my group that is doing this thing that is make or break for the company.
00:41:42
◼►
And they're like, look, we can go spend $5 billion and solve this problem or, or, or we don't have to, we can save that money and your team can do it.
00:43:56
◼►
But I could imagine if you were at Apple and you were really interested in this technology, it is likely that you may have gone to work somewhere else.
00:44:05
◼►
If it didn't seem like your company was going to do it, right?
00:44:22
◼►
Like, maybe you would go and work for OpenAI.
00:44:24
◼►
Maybe you would go and work at a startup.
00:44:26
◼►
Maybe you'd go work for Claude or whatever, right?
00:44:29
◼►
And so there is something to be said for Apple buying a company that has 100 excellent engineers focused on AI, where they may have a smaller team.
00:44:44
◼►
Or maybe they're having to get people interested in it who maybe aren't.
00:44:49
◼►
Like, I don't know what the internal feeling is amongst all developers on it, right?
00:44:53
◼►
But like, you know, you're like, okay, iOS team, you really need to get good at this now.
00:45:04
◼►
But do they have enough people who care about it, still left at the company when they decided they wanted to start down this road?
00:45:09
◼►
And it's a fun exercise, a very fun exercise, to talk about spending Apple's money.
00:45:14
◼►
Maybe we will have fun with that later.
00:45:15
◼►
But the problem is, and this is like, this is not fun for us or for people like Mark Gurman or anybody else who's like a pundit or an observer of Apple.
00:45:30
◼►
The truth is, the right company for Apple to buy is a company you've never heard of.
00:45:34
◼►
Maybe somebody at TechCrunch or, you know, somebody who's following very closely AI startups knows who that company is.
00:45:43
◼►
But like, that's probably the company Apple needs to buy.
00:46:52
◼►
That's the problem with a lot of this is those little companies, most people don't even know are there.
00:46:58
◼►
And they're, they're probably the ones you want to buy.
00:47:01
◼►
If you think that you could get a really nice infusion of talent and yeah, it's kind of aqua hire, but also like, if they've got a product, that's like the product you've been trying to build and failing.
00:47:28
◼►
Because everybody wants to be there when the, uh, when music stops.
00:47:32
◼►
And, uh, if you're one of the ones left standing, then you're going to be worth a fortune theoretically, or they want to be there when they get sold off to a tech channel.
00:47:40
◼►
This episode is brought to you by express VPN.
00:47:46
◼►
How did you choose which internet service provider to use?
00:47:50
◼►
The unfortunate thing is that many people have little choice.
00:47:53
◼►
Like you just have the one that is available to you.
00:47:56
◼►
ISPs have a lot of control in some of the regions that they serve in.
00:47:59
◼►
They can use this control to take advantage of their customers.
00:48:02
◼►
Things like data caps, streaming, throttling, the list goes on.
00:48:05
◼►
Stuff you don't want to have you think about because it annoys you.
00:48:15
◼►
Without express VPN, third parties could see the websites you visit, even in incognito mode.
00:48:20
◼►
That could be your ISP, your mobile network provider, or the admins of your Wi-Fi network.
00:48:24
◼►
The good news is express VPN reroutes 100% of your traffic through secure, encrypted servers, so third parties can't see your browsing history.
00:48:32
◼►
This could be used for advertising, all sorts.
00:48:36
◼►
And it's because especially when express VPN is so easy to use.
00:48:39
◼►
You just fire up the app, click a button, and you're protected.
00:48:41
◼►
It works on all devices, including phones, laptops, tablets, and more, so you can stay private on the go.
00:48:46
◼►
It's easy to see why it was rated number one by top tech reviewers like CNET and The Verge.
00:48:51
◼►
I use express VPN when I'm on networks that I don't control.
00:48:54
◼►
Like if I can't trust them, why would I just blindly connect to it?
00:48:58
◼►
And express VPN is so easy for me to enable on my Mac.
00:49:01
◼►
I use it a lot when I'm traveling, and I'm able to get around geode restrictions and stuff and continue watching TV shows when I'm at home.
00:49:08
◼►
Like as if I'm at home, so you can carry on with that Netflix show or whatever because it's available at home but not in the US when I'm traveling there.
00:49:15
◼►
And then I can watch things that are in full quality.
00:49:19
◼►
It's not like sitting there buffering.
00:49:21
◼►
You forget that you have it on, which is fantastic.
00:49:23
◼►
So go protect your online activity today by going to expressvpn.com slash upgrade.
00:49:29
◼►
That's E-X-P-R-E-S-S-V-P-N dot com slash upgrade, and you can get an extra four months for free.
00:49:36
◼►
Expressvpn.com slash upgrade to learn more.
00:49:39
◼►
Our thanks to ExpressVPN for their support of this show and all of Relay.
00:49:45
◼►
So you mentioned a minute ago about spending Apple's money.
00:58:43
◼►
If you're talking about an interesting startup that's doing things with AI, Software Applications Incorporated,
00:58:49
◼►
the company making Sky, who was previously shortcuts, like workflow and shortcuts.
00:58:53
◼►
Like it really feels like that product and like that team, like they're making a thing and it's interesting and it's doing some cool stuff.
00:59:03
◼►
But would also now, especially now we've seen a new Spotlight, would fit really well into that team now.
00:59:11
◼►
Like, you know, Spotlight is doing a lot of interesting stuff and then Sky is doing more of it.
00:59:19
◼►
So software applications, these are, this is the old shortcuts team.
00:59:22
◼►
Who got acquired by Apple and then they left after their deal was up and they have started this new thing that is basically just a, you know, AI control Mac utility kind of thing.
00:59:34
◼►
So my question about that is, it was a very impressive demo.
00:59:36
◼►
It's impressive that they got it out before WWDC, even though it's not actually out yet.
00:59:40
◼►
They showed it to Federico because it allowed them to make an impression before stuff that might steal some of the spotlight from them.
00:59:52
◼►
I, I, I am not opposed to spending money to, to reacquire those people, that team again.
01:00:01
◼►
Um, I, I, I don't know what the dynamics are.
01:00:06
◼►
Did they, are the people who, do they like people at Apple and hate other people at Apple and who would they be working for and all of that?
01:00:13
◼►
But Apple has also had a little bit of an attitude adjustment in the last year.
01:00:27
◼►
This is, this is what I was thinking about earlier when I was like, if you cared about this and thought you had something interesting, maybe nobody else wanted to do it.
01:01:09
◼►
They've always been like, oh no, no, we only hire eight players.
01:01:11
◼►
It takes forever to hire people at Apple.
01:01:13
◼►
A lot of Apple is understaffed for what their ambition is.
01:01:17
◼►
So I look at something like Sky and software applications and I say, you know, oh, it's those people we lost.
01:01:26
◼►
We bring them back and now we've got a little group that is doing this thing that we've struggled to do because even though our people could do it too, they're doing this other thing that also is important to us.
01:01:36
◼►
Like, that's one of the reasons you do that.
01:01:38
◼►
And you pay them, you pay them a premium to, because they're literally building.
01:01:43
◼►
I mean, I think the Sky demo is really interesting.
01:02:00
◼►
But, like, whatever they're worth, and I know they have an open AI investment and all that, it's like, just, I don't know.
01:02:07
◼►
I mean, this goes back to some other things, which is, why were they allowed to leave?
01:02:11
◼►
Probably they left because they got paid and then Apple wasn't letting them do the stuff that they thought would be an interesting path forward for them.
01:02:21
◼►
So do you say we've changed and we're going to buy you and we're going to put you in charge of this because this time we're going to listen to you?
01:02:29
◼►
I hope so, because what they're doing is really interesting and it is the direction that Apple should go in.
01:02:34
◼►
And I'm not convinced that they've got, again, that they've got the people who are on this.
01:02:39
◼►
Now, it's also possible that that Sky thing, internally at Apple, they're like, yeah, we're already building that.
01:04:03
◼►
But if you view Apple not as a computer company, which they're not, but as a global technological and manufacturing brand, and I still think we're in an era where there is room for companies with that kind of skill set and that kind of brand to make money in the auto industry.
01:04:27
◼►
Because the auto industry is being really shaken up and it's going to be different when all is said and done.
01:04:32
◼►
And it's, you know, 10 years ago would have been better.
01:04:38
◼►
My alternative suggestion, by the way, is, and some people are going to really love this and some people are really going to hate it, and it may not be who you think, is wait for the inevitable complete fall and crack up of Tesla's stock because they have been completely mismanaged for the last five plus years.
01:05:13
◼►
They're just overvalued because people believe that Elon Musk is going to create lots of value with his AI and his robots.
01:05:19
◼►
As opposed to the truth, which is they've got some really interesting electric cars that have been mismanaged for five years because the guy who runs the company is distracted by other stuff.
01:05:28
◼►
No one would be happier than Mark Gurman if Apple wore Tesla because then they would like the robots, you know?
01:05:49
◼►
I'm going to pick one that is purely self-serving, Whoop, because if they bought Whoop, it would mean that they wanted to build non-watch health products, which I still desperately want Apple to make.
01:06:01
◼►
So I can get my Apple health data, my rings, and wear a regular watch.
01:06:08
◼►
I want that more than anything, and they won't do it, but I wish they would.
01:06:14
◼►
Just give me a bracelet or a pendant or whatever.
01:07:44
◼►
And the HomePod, the whole product line is a failure.
01:07:47
◼►
The products are fine, but, like, if you think about our expectations when Apple started doing the HomePods, like, first off, the Siri part is a failure because Siri is a failure.
01:08:15
◼►
This is one of those examples where, like, the best time to buy Sonos was probably 10 years ago, but the next best time is today because they're better at it than Apple.
01:08:24
◼►
And whoever is doing this stuff at Apple, and, again, I'll say, maybe it's that they're understaffed.
01:08:29
◼►
Maybe it's that they're not a priority.
01:09:06
◼►
So, by Sonos, they're in trouble, but their base is pretty good, and they're very Apple in the sense that they sell expensive products, probably at really good margins.
01:09:18
◼►
But they've built a whole product set with a soundbar that Apple doesn't make, and portable speakers, and, like, I don't know.
01:09:29
◼►
It used to be a cliche to say Apple should buy Sonos, but I feel like it's come all the way back around where Apple should just buy Sonos.
01:09:38
◼►
Especially, they're worth, like, a billion dollars now or something, Sonos.
01:09:41
◼►
Like, it feels like something they should consider, I think.
01:10:38
◼►
And they're doing more stuff with, you know, AI and voice, which means that they're really, and they've got a, like, a new home product that we talked about.
01:13:02
◼►
You would need a division inside your company that you purchased that was completely focused on this and that would completely replace the existing culture of your approach to games.
01:13:33
◼►
In that case, you could argue that then they just need to set up a game studio somewhere else and hire a couple of brilliant executives and have them go to town.
01:13:43
◼►
I think the argument there is that it's just too close to Apple's core business and that there are going to be issues there.
01:13:49
◼►
But if you make a big purchase and you empower a lot of people who are senior at your purchase company to define the culture, you could maybe make a difference.
01:13:59
◼►
I don't know if it makes sense or not, but like that's the idea here really ultimately is what are the parts of Apple?
01:14:06
◼►
If Apple wants to buy something, what are the parts of Apple that for all their talk just can't do the job?
01:14:12
◼►
And could you buy something and bring it in and not?
01:14:16
◼►
And I would say ideally not crush it, right?
01:14:19
◼►
Ideally, it would become an engine inside your company for change, not something that gets crushed.
01:14:26
◼►
Like PA Semi, I keep coming back to it, but like that was an engine to change the culture of Apple in terms of making chips.
01:14:48
◼►
Yeah, like realistically, no, they shouldn't buy Valve, but they could maybe buy like a really good game publisher that could help them bring in new game, you know, like someone who's like really respected and could help them kind of like bring new stuff on board.
01:15:04
◼►
Because like realistically, if they bought Valve, they would crush it.
01:15:06
◼►
They just would because it's everything's made for a PC.
01:15:39
◼►
I want to just talk to you real quick about the Apple intelligence transcription stuff that I've been seeing a bunch of people post about, including you.
01:15:46
◼►
So OTJ, John Voorhees wrote an article at Mac Stories that shows off what Apple's new speech transcription APIs can do.
01:15:53
◼►
And he had his son put together like a command line tool to use.
01:17:50
◼►
Well, I mean, that comes with being fast, right?
01:17:52
◼►
I think that it's just all about trade-offs.
01:17:54
◼►
It would be nice in the long run if Apple let developers and shortcuts users, yes, make some choices about, you know, a little slower, more accurate, or a little faster, less accurate.
01:18:16
◼►
And that's probably why Apple is so much faster in some ways.
01:18:22
◼►
But it is, yeah, it's mostly like capitalization and punctuation, but also like phrases.
01:18:29
◼►
Like Whisper, it looks to me like Whisper's context window is a little bit wider.
01:18:35
◼►
And the point there is that if you do a transcription based on a word at a time, you have to just use the sound of that word to guess what that word is.
01:18:47
◼►
Your accuracy will be poor because you don't know any of the context.
01:18:52
◼►
If you have a large context window and you know the whole paragraph and you know what that sound is, not only do you know that sound, but you can infer from the context of the entire paragraph of what that word probably is.
01:19:07
◼►
And good transcription engines do that.
01:19:30
◼►
So it looks to me like Apple's model because it's fast probably is, has a little narrower context window.
01:19:36
◼►
And so it missed some turns of phrase and trans and did a more literal transcription of them, but still pretty good and super fast.
01:19:45
◼►
And, and keep in mind on device, which is very good.
01:19:49
◼►
And probably the reason that model is a little light, more lightweight is that they want it to be work well on an iPhone too, and not just a Mac at decent speed.
01:20:00
◼►
You use this for things like, I mean, I use this to generate podcast transcripts or subtitles for YouTube podcast transcripts that are searchable or.
01:20:10
◼►
Why do you want to provide subtitles for YouTube?
01:21:25
◼►
Now the, and I filed the feedback about this, but the problem is Apple will let a public podcast say, I, I would rather you use my transcript than your transcript.
01:21:54
◼►
They do some processing of the transcript and magic stuff, but like, it's very frustrating as a person who's got some members only podcasts that I can hand them a transcript and they don't want it.
01:22:28
◼►
I find multiple times in the last few months, like I've been looking for things that I said in podcasts and I find them, um, by going through the transcripts of my shows in Apple podcast.
01:22:58
◼►
And the reason for that is dictation engines tend to have a, a narrower context window and they are trying to, uh, be as real time as possible.
01:23:07
◼►
It's just a little different and they do have a context window.
01:23:10
◼►
Cause they'll sometimes go back and fix things that they initially transcribed one way.
01:23:13
◼►
And they're like, no, no, no, it was actually this way.
01:23:15
◼►
Anyway, it's a little bit different tech, but, uh, it's pretty cool.
01:23:18
◼►
And it can be used in lots of different ways to generate subtitles and transcriptions.
01:23:22
◼►
And, uh, it's awesome that that's going to be built in.
01:23:25
◼►
Cause again, there are other tools out there.
01:23:27
◼►
Also, honestly, Mike, you know, I tell people about like how I generate transcripts and they're like, oh, how do I do that?
01:23:33
◼►
And I'm like, well, you go to the whisper.cpp GitHub page or you download Mac whisper.
01:23:53
◼►
Um, I have one bonus here bonus, which I just wanted to mention, speaking of AI stuff, that's in the new, in the 26 releases, which is I realized yesterday, I was thinking about like, I really want to figure out something useful that I could do now that in 26, uh, shortcuts has access to the on-device model, the private cloud compute model.
01:24:17
◼►
So I could like build a shortcut that's got an AI something step in it.
01:24:23
◼►
And like, it's really great in theory, but there's questions like, is there somewhere in my workflows that's practical where I could use this?
01:24:30
◼►
And I thought of it yesterday and I built it today in no time, which is I have a script that takes an image on my computer or on my iPad, uh, uh, resizes it to the right size.
01:24:46
◼►
Optimizes it, uploads it to six colors, gets back the URL that it's going to be at, and then generates H an HTML fragment of the image.
01:24:58
◼►
So I say this image, go boop, boop, it runs.
01:25:01
◼►
And what I get back is a thing I can paste into my story.
01:25:04
◼►
That is the whole HTML image reference so that the image shows up in my story.
01:25:11
◼►
As opposed to opening WordPress and uploading a file and then grabbing the URL and then going back in and then typing, you know, angle bracket image source equals paste close quote close.
01:25:38
◼►
But the net result is that now when I do this thing or I drag a file into BB editor or whatever to execute my script, it puts that HTML there with alt text describing the image.
01:25:56
◼►
Because it sends the image to the private cloud compute model because you can't do images on the local model, it turns out, for now.
01:26:03
◼►
And says, I have a little thing that's like, give me back alt text for this image.
01:26:09
◼►
You know, don't use double quotes because it'll break the alt tag.
01:26:13
◼►
And if it's a screenshot, please include the text that's in the screenshot.
01:26:17
◼►
And, you know, I uploaded that picture of the cat that I took at Apple Park.
01:26:24
◼►
And it responded and it was like, this is a tabby cat sitting on concrete with a purple flower near it.
01:26:29
◼►
And I'm like, oh, my God, that's exactly what it is.
01:26:31
◼►
And so now I literally I do this and the HTML includes as the alt text the content of the image.
01:26:36
◼►
Now, I know you could do this with other LLMs and all that, but like that it's stock Apple.
01:26:40
◼►
And it has the ability for me as a regular user to analyze an image and generate a description of it without my interaction.
01:26:49
◼►
It also means we are headed for a world where more and more of this stuff will be built into our third party apps, which is pretty great, right?
01:26:56
◼►
The idea that your third party apps will be able to just do a little bit of of AI stuff in the background to make their app features better.
01:27:22
◼►
But one reason is that it's very expensive.
01:27:23
◼►
Yeah, I don't want the cost, which I understand that they don't want the cost.
01:27:27
◼►
And I really, you know, what will really help here is if Apple's on device model supports images, which it doesn't.
01:27:33
◼►
I think right now, which is too bad and needs to do that.
01:27:36
◼►
But I also had a funny moment where I was testing this out and I uploaded my full photo to private cloud compute and it took forever to get a response.
01:27:54
◼►
So all of the delay was in having a big image.
01:27:58
◼►
It's like, I don't need it to be that big.
01:27:59
◼►
I just need you to tell me what's in it.
01:28:02
◼►
So I'm excited about that because that was just one thought I had just passing through my mind where I was like, oh, I could use this to generate alt text, probably a caption too.
01:28:10
◼►
But the fact is, most of my captions for six colors, I am making a point referring to the actual text of the article.
01:28:16
◼►
So like I'm not writing, having it write a caption for me, just the alt text.
01:28:21
◼►
And obviously I can see the alt text and I can decide.
01:28:23
◼►
But like, how great is that, that I put that cat in there?
01:28:27
◼►
Or the one where it really gets me is screenshots, where having the text of the screenshot already be in there is really good.
01:28:36
◼►
So that I don't have to say like, oh, this is a screenshot containing these words.
01:28:43
◼►
So I'm sure I will find other uses to it.
01:28:45
◼►
But that was just like a little moment of, I built a shortcut that tied into an Apple script that uses private cloud compute to generate an alt text for an image like this.
01:31:10
◼►
And like, so you can go on Apple's page and it says requires an Apple intelligence enabled iPhone nearby and Bluetooth headphones with device and Siri language set to English.
01:31:21
◼►
It's generating all of that stuff looking at your health data on your iPhone.
01:31:26
◼►
And I'm a little disappointed that there isn't like a private cloud version of this, but it would have to send all your health data, which is a lot, in order to get those insights.
01:31:56
◼►
Because I think outdoor runners that love the Apple Watch love it because they don't have to take their fun with them.
01:32:02
◼►
I would have liked to have seen them try to solve this problem.
01:32:09
◼►
But somehow, and maybe the answer is you really just need more power on the Apple Watch in order to like, and the Apple Watch doesn't have all your health data, right?
01:32:19
◼►
I think that's part of the problem is that your health data ultimately is on your iPhone.
01:32:28
◼►
And here's what really kind of bugs me about this is really what they need to do is they need to make it that if you take a cellular watch away from your iPhone and the iPhone is on the internet and your cellular watch is on the internet, they should be able to talk to each other.
01:32:42
◼►
They should be able to say, oh, and maybe the answer is that they tried that and it's just the Apple Watch maintaining the cellular connection really hurts the battery and it's really not reliable.
01:32:53
◼►
And that may be, you know, I usually am listening to podcasts when I'm out with the dog.
01:32:58
◼►
There's also an inconsistency problem, right?
01:33:00
◼►
Where it's like, what if you left your phone at home and your phone had 19% battery?
01:33:05
◼►
Well, like, then you can't do the exchange.
01:33:07
◼►
So then you just put your phone on charge every time, you know.
01:33:09
◼►
What I was going to say is that when I'm streaming, like, connected live, there are dropouts because the cellular is not.
01:33:16
◼►
Whereas when I'm just listening to a podcast, I'm not really constantly hitting the cell network.
01:33:24
◼►
So anyway, yeah, that's the short version is there's lots of reasons why.
01:33:28
◼►
But basically, yes, Workout Buddy, which I've heard from some people who say that it provides my complaints about the demo, Mike, include the fact that the demo was the worst.
01:33:39
◼►
That I've heard people say, oh, Workout Buddy actually says some interesting and insightful things about you.