PodSearch

Connected

580: Lick It Twice

 

00:00:00   From Relay, this is Connected, episode 580. Today's show is brought to you by our sponsors,

00:00:12   Ecamm and KRCS. Happy Thanksgiving, and I'm joined by Mike Early.

00:00:17   Yes.

00:00:18   Hello, Mike.

00:00:18   Happy Thanksgiving to you, too, Federico. We are recording this on Thanksgiving Day.

00:00:23   It will not be published on Thanksgiving Day because Stephen said I couldn't make our edit

00:00:29   a gym work on Thanksgiving, which he's probably right. I wasn't trying to argue the point,

00:00:34   but he did remind me, and so, yeah, we're going to be publishing it a day late. I'm thankful

00:00:39   for Jim for all the work that he does every week in putting this show together and publishing

00:00:44   and getting it all wonderful for our listener pleasure, so we should all be thankful to our

00:00:49   wonderful editor, Jim Messendorf. But we will be up a day late, and yes, because it was also

00:00:54   Thanksgiving, it's just two of us, you know, just a couple of Europeans hanging out.

00:00:58   Yeah, it's no Thanksgiving for us. Do you celebrate with a special meal or anything tonight, Mike?

00:01:06   Not tonight, but we're going to be spending Thanksgiving with the underscore family this

00:01:11   weekend.

00:01:11   Nice, nice.

00:01:12   So they're inviting us to a meal that they're having, so I'm looking forward to that a lot.

00:01:16   Nice.

00:01:17   Should we do some follow-up, Federico?

00:01:19   Follow-up, yes.

00:01:20   So I was noodling around on my iPad, as I want to do.

00:01:25   As always.

00:01:26   As I want to do. And I opened the Instagram app, which I have not been using because I didn't

00:01:31   like it, because they were loading me into this horrible Reels interface on the homepage,

00:01:37   and it's changed for me.

00:01:38   Oh.

00:01:39   Instagram, like many of these large tech companies, they love to A-B test, you know?

00:01:45   So I don't know if I'm in some kind of A-B test, but now the kind of homepage of Instagram

00:01:50   for my iPad is what I would expect it to be. It's the content from the people that I follow,

00:01:55   photos and videos.

00:01:56   Oh, okay.

00:01:56   With a distinct tab for Reels. So like, this is what I want it to be. It's actually emulating

00:02:03   the new default layout for Instagram on the phone, where Instagram on the phone now has

00:02:11   a messages tab and a Reels tab. Everything's tabbed. The swiping interface just moves left

00:02:18   and right through the tabs, which is always a real jump scare for me when I'm wanting to

00:02:23   access the messages on, say, the Cortex brand account, and then it just opens Reels. So

00:02:28   that's like a whole other thing to deal with. But yeah, Instagram on the iPad is usable for

00:02:34   me now, because it's actually works most of the way that I would expect it to work, rather

00:02:38   than whatever it was they were doing beforehand.

00:02:40   Yeah, I don't think I ever installed it on my iPad.

00:02:44   Okay.

00:02:45   Not a great iPad app, if I'm being completely honest. But I don't think we were expecting

00:02:50   much. But it is a shame after so much time for it to kind of have been a bit of a damp

00:02:55   squid. Is that the phrase?

00:02:57   Wet squid?

00:02:58   Wet squid?

00:02:59   Is that a thing people say?

00:03:01   Damp squid?

00:03:02   Is that squib?

00:03:05   Squib?

00:03:07   Damp squib, I think is what I'm looking for.

00:03:09   A damp squib?

00:03:09   Yeah. An event that is not as exciting or popular as people thought it would be.

00:03:15   A damp squib.

00:03:17   What is damp squib?

00:03:18   I don't know what a squib is.

00:03:19   Is that like, is that in the UK an expression?

00:03:22   A squib generally consists of a small tube filled of an explosive substance.

00:03:28   Like a firework?

00:03:29   With a detonator running through the length of its core, similar to the sticker dynamite.

00:03:34   Yeah. So I guess one that is damp wouldn't go off, right?

00:03:37   Because it's wet.

00:03:38   So that's probably where that comes from. Good to know.

00:03:41   Good to know.

00:03:43   Johnny Ive and Sam Altman have said that they finally have a prototype of their AI hardware,

00:03:48   which I guess is good considering the amount of money.

00:03:52   Yeah, speaking of a damp squib.

00:03:53   Wow. What an incredible unexpected transition.

00:04:00   So when these two announced that they were working together, I think we expected that they were further along than they clearly have been.

00:04:11   Because they seem to have just found this prototype.

00:04:14   Apparently it could arrive in less than two years, which is still within the originally stated time frame.

00:04:20   Like they said, they would show something in 26, I think, to go on sale in 27.

00:04:27   I'm quoting from an article by Julie Clover, Mac Rumors.

00:04:32   They say, it will be a sort of active participant that is proactive in a, quote, not annoying way.

00:04:40   Essentially, they want...

00:04:41   That doesn't mean anything.

00:04:42   No, no, no.

00:04:44   They want the device to not be trying to get your attention.

00:04:47   They talk a lot about how your iPhone feels like you're walking in Times Square in New York, and they want it to not feel like that.

00:04:53   You know, it's okay, fine.

00:04:54   Another quote.

00:04:55   I've told Altman that the design is right when you want to lick it or take a bite out of it.

00:05:01   And with the current prototype, Altman got that feeling, like you wanted to take a bite out of it.

00:05:05   I don't know why they're talking like this.

00:05:07   I got it.

00:05:09   I got it.

00:05:09   I got it.

00:05:10   I got it.

00:05:10   Plot twist.

00:05:11   It's shaped like an apple.

00:05:12   Yeah.

00:05:14   I mean, it feels to me that they're trying to emulate the Steve Jobs line, right?

00:05:22   Of like, UI looks so good you want to lick it.

00:05:26   No, I got it.

00:05:29   I think I got it.

00:05:30   I think I got it.

00:05:31   It's going to be shaped like an apple, and they're going to name it Eve.

00:05:35   Because it's also like the Eve of a new era.

00:05:38   Man, if they do that, I'll be so mad.

00:05:41   I'll be so angry if they do this.

00:05:45   I'll be like, what are we doing?

00:05:46   They speak a lot together.

00:05:48   They were at an event for, what was it?

00:05:53   The Emerson Collective, which is Lauren Powell Jobs.

00:05:56   That's her company, investment firm, that kind of thing, their demo day.

00:06:02   She's put a lot of money into, I think, both of their endeavors.

00:06:06   So it makes sense that they have an obligation, really, to come and talk.

00:06:12   And they speak a lot about wanting devices to feel playful and careful and simple and something you don't think about.

00:06:20   Like, all of this sounds great, but I have now gotten to the point where I just want them to stop talking.

00:06:26   Yeah.

00:06:27   Because they're heading towards humane territory at this point, right?

00:06:32   Like, you have to start showing me things and stop just saying nothing things.

00:06:39   Like, things that are nothing without the context of the hardware.

00:06:43   Like, I can say this.

00:06:45   Me and Federico have been working really hard on this new device, and it's absolutely going to change the way that you think about computers.

00:06:52   Like, I can say that, too.

00:06:54   I just said it.

00:06:55   But it doesn't mean anything, because I've got nothing to show.

00:06:58   It's actually true.

00:06:59   It's true.

00:06:59   It is true.

00:07:00   You want to lick it, in fact.

00:07:01   You absolutely want to lick it.

00:07:03   Twice.

00:07:03   Lick it twice?

00:07:05   Uh-huh.

00:07:06   Okay.

00:07:06   First me, then you?

00:07:07   All right.

00:07:08   Cool.

00:07:12   Yep.

00:07:13   That's how that works.

00:07:14   Same spot, even.

00:07:15   We like to take risks.

00:07:16   No, I mean, unfortunately here, the issue is that Altman is the kind of guy who likes to talk.

00:07:24   Yes.

00:07:25   Likes to say things.

00:07:27   But Johnny is historically not, right?

00:07:30   Which is interesting.

00:07:30   Like, he, I think, infamously has not spoken much.

00:07:37   He did not speak much in his career at Apple, right?

00:07:41   Like, he was never on stage.

00:07:43   He was just in pre-recorded videos.

00:07:46   He very rarely did interviews.

00:07:49   And it feels like he is, in his independent life, he's giving a lot more interviews.

00:07:55   And he is going and speaking at events and stuff.

00:08:03   Now, I feel like some of these events, like, he has lost a, he has actually, in his independent life,

00:08:09   lost some level of autonomy.

00:08:11   Because he is now kind of beholden to investors a little bit more than he was before and stuff

00:08:19   like that, right?

00:08:19   Or he is beholden to customers and has to, you know, because with his design firm.

00:08:27   And so he's, he's doing more events, which is an interesting thing.

00:08:31   Like, I like hearing him talk.

00:08:33   But I want to hear him talk about product when the product is something I can see.

00:08:39   I'll hear him talk about philosophy and his design ideas for as much as he wants.

00:08:45   Like, I like to, I just like to hear him talk.

00:08:48   But I don't want to hear him talk about, oh, we have this product.

00:08:52   It's amazing.

00:08:53   Like, I can't show it to you or tell you literally anything about it, other than the

00:08:57   fact that it's going to be super good.

00:08:59   Like, I just don't, show me it or don't show me it.

00:09:01   And until then, I don't want to hear about it.

00:09:04   I, yeah, I, we, I still feel like this is going to be some kind of, some kind of recording

00:09:09   device that also has a speaker.

00:09:11   And essentially, I think what they're trying to do is they're trying to have like a physical

00:09:16   manifestation of Chagipity that is a cute little object, object that you carry with you.

00:09:21   And it must be pocketable and it must be nice to look at.

00:09:24   I think that's where like the liquid or take a bite out of it bit comes from.

00:09:29   It's going to be some kind of soft touch material with a cute shape and, you know, I mean, I

00:09:34   don't doubt that Johnny Ive and all the people they hired from Apple, uh, they can make a nice

00:09:42   physical object.

00:09:43   I don't have any doubts about that.

00:09:45   I have many doubts about the actual performance of it all.

00:09:47   Yeah.

00:09:47   The, the utility of it all.

00:09:49   Industrial design they can, they can handle, right?

00:09:51   Like that's, yeah, for sure.

00:09:53   That's not a problem, right?

00:09:55   But it's about what does it do and how does it work?

00:10:00   And those are big unknowns, big unknowns.

00:10:04   And like, does it, does it even really make sense as a device when we have our iPhone?

00:10:12   I guess what they, if I were to imagine what they're banking on is the fact that, I mean,

00:10:18   take a look at what, what they have done just to push people to install ChatGPT Atlas.

00:10:24   So when they made a browser, you open, you go to ChatGPT and it tells you, Hey, we have

00:10:29   a browser.

00:10:30   Now imagine, imagine OpenAI really pushing an accessory.

00:10:36   And maybe if you are a ChatGPT subscriber, you get some kind of discount.

00:10:40   I was thinking about this today, actually, sign up for a year, get it for free.

00:10:43   Yeah.

00:10:44   Yeah.

00:10:44   Imagine that too.

00:10:46   I mean, ChatGPT is now, I was taking a look at some numbers yesterday.

00:10:49   They are now accelerating again.

00:10:52   They should be hitting before the end of the year, 900, 900 million subscribers, no users,

00:11:01   weekly users.

00:11:02   Like daily?

00:11:02   Daily?

00:11:03   Daily?

00:11:03   Daily or monthly active users.

00:11:05   So they're approaching a billion people using ChatGPT.

00:11:07   So Facebook size, basically.

00:11:09   Even if you sell, like, help me out with the math here.

00:11:11   If you sell even that accessory to even just 1% of those people, how much is that?

00:11:17   Yeah, but what I would want to know from those 900 million, how many people actually pay the

00:11:21   money?

00:11:21   I don't think it's that many people.

00:11:23   I was also taking a look at some projections for that yesterday.

00:11:28   I think the information probably reported that they plan to have something like 200 billion

00:11:34   or something in revenue.

00:11:36   So it's some crazy number.

00:11:37   Yeah, I don't know.

00:11:38   They can just say whatever they want.

00:11:40   Again, they can just say whatever they want, right?

00:11:43   Because that's what he's so good at.

00:11:45   Altman will just say anything.

00:11:46   He'll just say anything.

00:11:47   And he's clearly convincing.

00:11:49   Like, he's able to convince people.

00:11:51   Yeah, I don't know.

00:11:53   I just don't know what their business model is, really.

00:11:55   Like, hang on.

00:11:56   Let me rephrase that.

00:11:57   Obviously, I know what their business model is.

00:11:59   I'm not sure how successful they are achieving it.

00:12:02   But we'll see.

00:12:03   We'll see.

00:12:04   I mean, it's a very addictive product, right?

00:12:07   Like, people use it, and they want to keep using it.

00:12:10   And at some point, if you can kind of build the product right, you will get people to sign

00:12:17   up.

00:12:18   But it's about what they don't want to do is meet their access too much.

00:12:22   Yeah, it's an interesting.

00:12:24   It's very interesting.

00:12:25   Very interesting.

00:12:28   As a bit of follow-up, you mentioned this in last week's episode at the end, and you

00:12:31   did publish this afterwards.

00:12:32   I just wanted to know it here, that you have written a follow-up article to your M5 iPad

00:12:38   Pro review, as you've now got access to the on-device MLX model that you did not have, but Apple

00:12:47   was benchmarking against when the device was released, right?

00:12:50   Yeah, I was able to test the version of MLX with the M5-specific improvements.

00:12:57   And in fact, I did get the same numbers that Apple was claiming.

00:13:01   In fact, in some tests, even better numbers, like a four, four and a half X improvement.

00:13:07   And on the same day, Apple also published a blog post on their machine learning blog with

00:13:12   more benchmarks for the M5 and MLX, which is now officially in a beta release on GitHub.

00:13:19   If you're a developer and you want to integrate with that, it's in beta on GitHub.

00:13:25   And I assume, this is just my personal theory, I assume we're going to see a public release

00:13:30   in the MLX community before the end of the year.

00:13:35   So that's just great.

00:13:37   I mean, once again, local AI on an iPad is a niche within a niche.

00:13:42   But it's impressive nonetheless.

00:13:48   It's indicative of a potential thing in the future, right?

00:13:51   Like, the more powerful they can make this and the more efficient they can make this, the

00:13:57   better it will be for the point when we can have really good on-device models.

00:14:02   Like, when Apple's on-device models get super powerful, they will be able to be processed by

00:14:09   their chips at incredible speeds, which could be really interesting.

00:14:12   Yeah.

00:14:13   And you're not going to wait around too long for a response because that's really what the

00:14:17   advantage of the M5 is.

00:14:18   And you've got to imagine that.

00:14:20   So I was thinking about this today.

00:14:22   They did CPU acceleration with the M4.

00:14:25   They did GPU acceleration with the M5.

00:14:27   I kind of feel like, you know, we don't know exactly for how long these chips are in the

00:14:33   making and how quickly Apple realized, oh, LLMs are going to be a thing.

00:14:39   We've got to optimize our silicon architecture for LLMs.

00:14:42   I don't know if the M4 and the M5 are like maybe half-steps measured to optimize their chipsets

00:14:49   for local AI performance.

00:14:50   What could they do in a short period of time as opposed to a longer period of time?

00:14:54   And so I would expect even bigger improvements, you know, with the M6 or M7.

00:14:59   I mean, even if you just take a look at what AMD and NVIDIA are doing now to, I mean, they're

00:15:06   making chipsets and GPUs that are sold as AI GPUs and then people are buying them for gaming,

00:15:13   which is quite clearly that's not, I mean, gaming is now a fraction of NVIDIA's revenue at

00:15:19   this point.

00:15:20   But I wonder what can Apple do now that they had, you know, a couple of years since CHGPT

00:15:27   exploded.

00:15:27   And you've got to imagine that with future chips, they will be able to squeeze even more out of

00:15:33   them for local AI performance.

00:15:34   Yeah.

00:15:34   We'll see.

00:15:35   But yeah.

00:15:35   Yeah.

00:15:36   It's like, obviously we haven't, they haven't had enough time to create something from the

00:15:41   ground up with this in mind.

00:15:43   Would they?

00:15:44   And what would it look like?

00:15:45   Yeah.

00:15:46   That's going to be interesting to see.

00:15:47   I mean, especially it will also be interesting.

00:15:50   It'll be helpful for them.

00:15:51   Right.

00:15:51   They're going to need a chip like that in there for private cloud compute.

00:15:55   Because what we know about private cloud compute is that is an incredibly inefficient system.

00:16:00   Like, what was it?

00:16:01   They just had a bunch of Mac studios?

00:16:02   Like...

00:16:03   Yeah.

00:16:03   With the M3 Ultra.

00:16:04   I don't even think it was that.

00:16:06   I think it was like the M2 Ultra or something.

00:16:08   Yeah.

00:16:09   And then they're probably using the M3 Ultra now, but they need custom silicon if they're

00:16:15   going to want to run this themselves.

00:16:16   Especially if you want to serve the white label Gemini.

00:16:19   Yes.

00:16:20   Yeah.

00:16:21   Yeah.

00:16:22   Talking about the wonderful website, MacStories.net, you published an article with a few more, like,

00:16:30   here's a bunch more apps that have been updated for Liquid Glass.

00:16:33   Yeah.

00:16:35   And I thought it was interesting.

00:16:36   I also had one to throw in the pile, and I'll put a link to MaxWalky in it.

00:16:40   Udilse has had a nice update.

00:16:42   It's not, you know, groundbreaking, but they're doing, like, a lot of the kind of...

00:16:47   What I like most about Liquid Glass, actually, is just the kind of restructuring of the apps.

00:16:52   You know, like, it seems like the iOS 26 design aesthetic is making a lot of people, like,

00:16:59   think about how their apps are tabbed and what search is like and that kind of stuff.

00:17:04   Like, where those buttons go.

00:17:05   So that's been nice.

00:17:06   But yeah, you've posted a bunch of good examples of more apps that are, like, really kind of

00:17:12   taking this design aesthetic and doing good things with it.

00:17:17   And I think overall, like, it seems to be...

00:17:21   It seems like there were a bunch of developers that took time, but that was right for them.

00:17:25   And now we're starting to see the fruits of that labor.

00:17:28   Yeah.

00:17:29   Yeah.

00:17:29   I don't know.

00:17:32   I've been having a good time with Liquid Glass in third-party apps.

00:17:35   And there was a rush, obviously, in September.

00:17:37   And I think over the past couple of months, a lot of developers have refined as much as they

00:17:41   could.

00:17:42   And also, I think it's just us.

00:17:43   I think we've gotten used to Liquid Glass and the customization that is happening in

00:17:48   26.2.

00:17:49   I think it's also helping.

00:17:52   And obviously, that's not out to the public yet.

00:17:56   But yeah, I think, I don't know.

00:17:58   I feel like I prefer now, like, when I open a third-party app on my phone now, and it doesn't

00:18:06   have the Liquid Glass keyboard, and it hasn't been updated, I'm like...

00:18:11   I don't understand apps that don't have the keyboard, at least.

00:18:15   Like, I don't know what's going on.

00:18:16   Like, so I use Duolingo, right?

00:18:18   They still have the old keyboard.

00:18:20   So what are we doing, Duolingo?

00:18:21   Like, you update that app, like, every four days.

00:18:24   Like, could you not build for 26 at any point?

00:18:27   Or what?

00:18:28   Yeah, it's fascinating to me.

00:18:29   I don't know what's going on.

00:18:31   But yeah, I have been encouraged, I think, by the amount of apps that I wouldn't have expected

00:18:36   that have, like, adopted the new toolbars and stuff, at least.

00:18:38   So they still feel kind of at home.

00:18:41   I wanted to mention, in case you have forgotten, you should not have forgotten, we are currently

00:18:48   doing 20% off an annual plan for this show and all of our shows that have membership over

00:18:54   relay, you can go to giverelay.com to learn more, where you can also, you can sign up

00:18:59   to become a member.

00:19:00   You will get an ad-free version of this show each and every week of additional content.

00:19:05   Today, we do what we always do in Steam is Not Here, and we just talk about real gamers

00:19:09   topics for real gamers.

00:19:10   That's at Connected Pro.

00:19:12   You can just go to giverelay.com.

00:19:15   You can purchase as a gift for yourself, or you can buy it as a gift for someone else, and

00:19:19   you can do that for any of our shows.

00:19:20   Or if you want to, you can just go to getconnectedpro.co and use the code 2025Holidays at checkout.

00:19:27   Really easy to do so.

00:19:28   And you'll be supporting the show.

00:19:32   We would love it if you supported the show, in all honesty, because it means a lot to

00:19:36   us.

00:19:36   We put a lot of work into the show.

00:19:37   We love doing it.

00:19:40   Advertising is an unexpected and fickle thing.

00:19:43   Advertising is more difficult than it's ever been, which is kind of like a continued thing.

00:19:50   Like, podcast advertising has become harder every year that we've been running our business.

00:19:54   You know, it's just like it's always harder.

00:19:57   But one of the things that's really benefited us over the last five years is having membership

00:20:01   to support our shows.

00:20:03   And I'm sure, I know this is something Federico's experience with Max Lurie and MPC.

00:20:07   Like, having people support you with the content that you make and because they enjoy it, it

00:20:13   really helps.

00:20:14   It, like, takes the pressure off us.

00:20:16   It gives us a level of money that we can feel confident in because once people sign up, they

00:20:22   do not cancel, basically.

00:20:24   Like, we have very, very low churn, which means, like, people leaving.

00:20:27   Like, we bring in more people than we lose.

00:20:30   So if you are one of those people, thank you so much.

00:20:34   But please go to GiveRelay.com, learn more, sign up, you get 20% off an annual plan for

00:20:40   your first year.

00:20:40   It's just $56.

00:20:41   So it's like a dollar an episode for the next year.

00:20:44   You'll be supporting us and we'll give you a lot of bonuses in return.

00:20:50   This episode is brought to you by Ecamm.

00:20:53   If you're a Mac user who creates video or podcasts, you need Ecamm because Ecamm Live is the all-in-one

00:21:01   studio built exclusively for the Mac.

00:21:04   It looks, feels, and performs like a native Pro app should.

00:21:08   Whether you're live streaming, recording a podcast, or producing training videos, Ecamm gives you

00:21:14   broadcast-level control with drag-and-drop simplicity.

00:21:17   You can switch cameras, share your screen, cue overlays, and control audio, all without

00:21:22   ever leaving your Mac.

00:21:23   With Ecamm, you can brand your show with titles, graphics, and lower thirds.

00:21:27   You can pull in guests via interview mode or record multi-track audio for perfect post-production.

00:21:32   If you're into automation, Ecamm works beautifully with tools and apps like Stream Deck and Loopback,

00:21:38   as well as many other Mac tools you already know and love because this is a real Mac app.

00:21:43   This feels like a Mac app. It works like a Mac app. It is built by people who love the Mac and want to

00:21:49   make the best streaming software available, and it really is incredible. There was a time when trying

00:21:54   to stream from a Mac was essentially impossible, but Ecamm have unlocked that power, and all of the

00:22:00   power that's in on your Macs these days are absolutely like a Macs. They're very capable of this stuff.

00:22:05   We were just missing the right software, and Ecamm Live is absolutely the right software.

00:22:09   You can upgrade to Pro and unlock Ecamm for Zoom. This will let you feed your policy setup straight

00:22:15   into Zoom meetings or webinars. You can share Zoom comments on screen and even capture each

00:22:20   participant's audio and video separately for easy post-production work.

00:22:24   Right now, new customers can get Ecamm's best offer of the year. You can get 40% off an annual plan in their Black Friday Cyber Monday deal

00:22:34   at Ecamm.com. That is E-C-A-M-M dot com. That is 40% off annual plans between now and December 2nd at Ecamm.com.

00:22:43   Our thanks to Ecamm for their support of this show and Relay.

00:22:47   Our friend, the rumor wizard, Mark Gurman, has had a report. The report focused on a bunch of different things,

00:22:56   but the thing I wanted to talk about with you most is about iOS 27. Mark says that there will be a focus

00:23:04   in two key areas, AI features and overall operating system quality. Mark invokes Snow Leopard.

00:23:12   In fact, the image in the article is Bertrand Soleil, which is the name I've not heard in a long time,

00:23:20   standing in front of Mac OS X Snow Leopard. Now, Snow Leopard is known as the no new releases year,

00:23:28   which is the thing that is banded around a lot. Apple should have a Snow Leopard year,

00:23:33   a Snow Leopard year, which is the idea of increasing software quality and fixing bugs. That was the way

00:23:39   Apple pitched it, but actually under the hood, Snow Leopard was a massive re-architecture of a bunch

00:23:44   of different key technologies. And the introduction, I think, of Grand Central Dispatch,

00:23:49   like a bunch of like really heavy stuff under the hood. But from a user's perspective,

00:23:54   it was more about making the operating system feel better to use. And in a sense, like people forget

00:24:00   about and people don't really think about the big underpinning stuff, but they think about the user

00:24:05   facing features. Now, we'll come back to this again in a minute, but I actually think if Apple

00:24:10   does what Mark is saying here, this actually would be Snow Leopard like in bug fixes and big under the

00:24:18   hood stuff. So basically, it seems like at the moment, if you're not working on AI at Apple, you are working

00:24:27   on fixing things to make the system run smoother. Like this appears to be what Mark is pitching,

00:24:33   that like everybody is just working on cutting bloat, fixing bugs, like making the system

00:24:39   nicer to use. But also, Mark does expect, considering that we just got a full redesign,

00:24:46   that iOS 27 would feature some new styling and customization options for liquid glass.

00:24:51   So just stopping on that part, from like a bug fixes and performance improvements release,

00:24:56   what is your thought on that for the 27 calendar?

00:25:01   I feel like there's going to be a lot of liquid glass refinements. I think a lot of people,

00:25:05   I see these viral posts every once in a while showing like some glitchy liquid glass animation

00:25:13   or visual bug. I think there's going to be, I think, you know, Apple can tell a nice story

00:25:21   about like, and we looked at every nook and cranny of iOS and like this idea of, you know,

00:25:28   iOS has grown so much over the past decade and it's undergone so many changes that it's, you know,

00:25:36   every once in a while, it's got to be time for spring cleaning and, you know, optimizing everything.

00:25:41   And so I think, you know, this idea of we're going to make everybody, I think it's uniquely Apple to say

00:25:49   we are going to make everyone's phones faster and more efficient, even the older ones.

00:25:56   And Apple is already well known for being one of the few, the only company that actually cares about

00:26:04   supporting old phones with software updates. And so I think, I think it would be in Apple's best interest.

00:26:10   Also going against the idea of what do they call it? Planned obsolescence. Like you say, we're going to make

00:26:16   everybody's devices look faster, be more efficient, be more polished. That's a nice story.

00:26:22   Yeah. Cause I think maybe one of the only things that I have heard, which I feel like would make sense

00:26:28   is that iOS 26 is rough on some of the older phones that Apple supports.

00:26:35   Yeah. And I could understand that completely. Right. Like that, that feels like a thing that,

00:26:39   that totally makes sense to me and that they could maybe do some work to make that better. I mean,

00:26:46   like when you think about like iOS 26 runs on the iPhone 11. Yeah. Like remember that? That is a,

00:26:55   that is a really old phone at this point. Um, so it's impressive that they do it, but you've got to

00:27:00   assume that that is not where they are prioritizing when they're, when they're rolling out first, but they

00:27:06   could take some time making that better for later. I mean, in general, I think they've had a win of iOS 26.

00:27:15   Like I don't hear complaints. I mean, it doesn't seem like they've done the press the button.

00:27:21   It goes to everyone yet, which is they're getting pretty late in the year. Um, but I would understand why you

00:27:27   would not do that, especially because they seem to not be able to stop themselves from adding new features

00:27:32   and all the point releases right now. It's like, I wouldn't be surprised at this point if what people,

00:27:36   end up getting on their phones is 26.2 rather than 26.1. Cause I think it's gotta, it's gotta be

00:27:42   that one when they push the button to 26.2, you think? Yeah. Yeah. I think so. I think so.

00:27:48   Cause it seems like they, to me, at least it feels like if they were going to push the,

00:27:52   and when I say push the button, it's the, you know, we all, I'm expecting the vast majority of

00:27:57   you that listeners of this show go and get the, you know, they go to the system settings and ask for

00:28:04   the update. Right. Um, but there comes a point in the year when Apple will start pushing it to

00:28:11   everybody else. And so the people that have automatic updates, they're going to wake up one day and

00:28:16   their phone's going to be updated and it doesn't appear like that's happened yet. Um, and so if

00:28:22   they've not done it by now, you, I would expect they're waiting for a specific release and it might

00:28:28   be 0.2 because they, they, you know, you can tell me you, obviously you have a much better sense of

00:28:33   this stuff than me. Does it feel like they're adding more features in these point releases than

00:28:38   they would normally? Feels like, especially this early in the cycle. Yeah. Like usually the, the,

00:28:45   the, the more rich releases were like 0.3 or 0.4. Yeah. Historically speaking. Um,

00:28:53   so I don't know if, if they have, you know, accelerated or changed timelines or, you know,

00:28:59   just, they had so much that they wanted to be sure, you know, especially since the, the whole

00:29:04   Apple intelligence debacle from, from last year. Well, that's going to come at some point, right?

00:29:08   The 0.4. Like, yeah, that's going to happen at some point. And, um, but, but to go back to what

00:29:17   the sheriff is saying, I think there is an angle for Apple to essentially productize

00:29:23   performance improvements. Because the thing that I hear from people in my real life all the time is,

00:29:31   oh, should I update? Is it going to make my phone run slower? Like that is the question. And if they

00:29:37   can actually flip that and say, actually, this software update is designed to make your phone faster.

00:29:43   I think that is a way beyond emoji to convince people to upgrade. Um, and yeah, I think they

00:29:50   should do it. I think, uh, I honestly, it's been a while. I mean, I've been reviewing iOS for the past

00:29:56   11 years. Um, I can imagine that there's a lot under the hood that could use like, as like just a whole

00:30:06   bunch of engineers sitting down and be like, okay, how do we, how do we polish all of this? Like, uh,

00:30:13   there must be some old like, like, and in the process, I would imagine they're going to get rid

00:30:17   of some stuff, right? Uh, that's usually what happens. Like, I would imagine that some features

00:30:22   will be deprecated. Oh, if they're doing this, I don't know what. Um, but I would imagine that some,

00:30:28   some things will be, will be turned off or removed, or, you know, they will pick some new

00:30:33   defaults for settings or just, you know, there are so many options and so many sub settings on iOS

00:30:40   today. I would imagine that the, with this kind of focus on overall quality, you would get rid of

00:30:46   some things. Yeah. Move things around, collapse some things into each other. Like, you know,

00:30:50   there are always features and, uh, settings and all that stuff that can be tidied up,

00:30:57   right? Cause there's just stuff that's just not being used or it's not being used in the way it's

00:31:01   expected to be used. And can they kind of work some stuff, re-architect some stuff, those kinds of

00:31:07   things. But Mark Gurman obviously expects that for Apple intelligence, we should see the improved Siri

00:31:14   before iOS 27 is released, uh, but we'll continue to see more improvements there. Now, so he references

00:31:22   in iOS 27 specifically that there would be some health app features powered by Apple intelligence,

00:31:27   some form of an AI powered web search. And Gurman keeps referencing big AI upgrades, but doesn't really

00:31:36   mention what they are. Um, the, I guess the expectation is there's more to know. Um, he says that the,

00:31:44   he's not expecting them to ship a Siri kind of chat bot app, which I think is a mistake if they,

00:31:49   uh, if they don't do that in my opinion, um, because they're testing one internally, but he's not

00:31:54   expecting that they will ship that. But yeah, I mean, here's my thing. I mean, I says a few places,

00:32:00   I keep saying it. I don't think that we get the Gemini powered Siri before we see iOS 27.

00:32:09   Oh, that's my, that's my, that's my bet. This all just feels like it's coming in too hot.

00:32:14   Like to expect that in March, we're going to get this when it feels like they're still doing the deal.

00:32:21   like, uh, I, it just seems to, I mean, still within the year, right? They said it within the year.

00:32:27   Like, I think we get it in the 26 cycle, but I think it comes after WWDC.

00:32:34   Hmm. I don't know. I don't know.

00:32:39   It shouldn't. I mean, it really shouldn't. Uh, but I, I feel like I would say I could see a scenario

00:32:46   in which we do get it just in the Siri voice UI before 27, but then we get it, get it with a Siri chatbot

00:32:59   in the 27 release cycle, not necessarily in September. Uh, but I don't think there's a, I mean,

00:33:07   let's imagine this, right? Let's, let's fast forward to June, 2026. It's, it's going to be how many,

00:33:14   uh, when did ChatGPT came out? Was it late 2022 or late 2023?

00:33:18   It was late 23 because we just moved into our new house, I think.

00:33:24   Right. So it's going to be, it's going to be three, almost three, four years since ChatGPT.

00:33:31   And here we are at WWDC 2026 and Apple does not announce a Siri chatbot experience.

00:33:41   Uh, late 2022, correction, late 2022.

00:33:44   Okay. So four years.

00:33:46   Yeah.

00:33:46   I just don't find that possible at this point that they do, they, that they're going to be

00:33:55   on stage and announce a bunch of Apple intelligence features in Safari, in notes, in Apple music,

00:34:01   in health, and they're going to just show off Siri with the usual UI. You ask a question and that's it.

00:34:07   Yeah. And there's, there's like no way to access previous conversations.

00:34:11   No. Or just to type and have a conversation.

00:34:14   Especially, like, I think the thing that makes it like, it feels especially unlikely to me. Like,

00:34:17   and I understand what Mark's saying, but the fact that they thought it was important enough

00:34:21   to build an internal tool shows that they know. So they've got to get it right. I think my kind

00:34:29   of hesitance is to the overall package, right? Like we might see a better Siri before WWDC,

00:34:38   but I am not convinced we will see personal context or on-screen awareness or app intents. Like

00:34:46   I do not think that all of that ships before iOS 27. Like I think some of these features ship

00:34:52   with iOS 27. Like I just, I, I think that it's, I don't know if they're ever going to ship some

00:35:00   of that stuff.

00:35:01   I, I gotta tell you, I gotta tell you, having been really plugged into this whole ecosystem

00:35:08   for the past year.

00:35:09   Yeah.

00:35:09   And being the person that likes to use AI for automation and productivity, which is why I

00:35:18   really like Claude because of the app integrations.

00:35:21   Yeah.

00:35:21   Right. Using AI with services like Todoist, Notion, Slack, Gmail. If there's a thing that really

00:35:30   confuses and slows down LLMs is when you enable a lot of integrations. And that's partly the

00:35:39   nature of the protocol that they're using, MCP, which now all of the major companies are sort

00:35:46   of trying to fix MCP. But so far, the biggest obstacle has been as users, the fact that when

00:35:54   you enable a lot of connectors, it just confuses the model, slows down the model. Especially

00:36:02   when you have, you know, apps and services that have tools that are similar to each other.

00:36:07   Right.

00:36:08   Right.

00:36:08   It has to try and work out where to go.

00:36:10   Which one, which one, right? And Apple is saying, no, we're going to, we're going to make

00:36:14   it work with all the apps on your phone, with all the app intents.

00:36:17   Every app on your phone.

00:36:18   All of them.

00:36:21   And I mean, unless, unless they made, and look, this is totally possible. Unless they made internally

00:36:26   a major breakthrough in tool calling and how LLMs are supposed to make sense of this branching,

00:36:36   you know, mosaic of integrations. I really don't know how they're going to make it work.

00:36:44   Especially when you have even the simplest thing. If you have like millions of people

00:36:50   on your phone, you have Apple Notes and Notion. How's that going to work? Like, are you going

00:36:56   to be asked every single time? Where do you want to save a note?

00:36:59   But this goes back to the original thing. It didn't exist.

00:37:02   It didn't exist.

00:37:03   It didn't exist.

00:37:04   It didn't exist. It didn't exist. And now you have, you have all these companies like

00:37:09   the Anthropic and Google and, and, and OpenAI. They're saying, well, look, MCP was nice, but

00:37:16   we have realized that actually it's better if instead of calling these tools with natural

00:37:22   language, it's actually better if we use code. If we just call them using, uh, if the model writes

00:37:28   some code instead of talking to MCP using the verbosity of the English language. And

00:37:35   what is Apple Intelligence on device exactly going to do there? Like, I, I don't know. Honestly,

00:37:42   don't know. Uh, which is why I'm so, uh, I think they have to do it. Here's the thing. I think they

00:37:48   have to do it. First of all, I don't know if they can do it on device, which was their big

00:37:55   thing last year. Like, Oh, we're going to have a model on device. It's going to call all these

00:38:00   up intents and it's going to pull all these contacts from you. Like, okay, sure. I, I can

00:38:06   tell you that I've tested personally on my MacBook pro on my Mac studio, open source models from

00:38:12   China usually that say that they support local tool calling. And I can tell you it's a mess.

00:38:20   And those are big models. Those are like 30 billion parameter models. And they absolutely go nuts when

00:38:26   you try and run a local tool, um, that is not baked into the model. And Apple is saying, well, no,

00:38:32   actually, actually we can do it with a 3 billion model on device. And I'm like, yeah, how? Like I,

00:38:40   it's puzzling to me. Yeah. I heard this on a podcast I was listening to recently. It might

00:38:44   have been the Verge cast and they were talking about this. There's this like growing movement

00:38:49   of like LLMs, it's not intelligence and you can't get intelligence out of it. Like you can't LLMs

00:38:58   will not get us to art of actual artificial intelligence. Like it's just words like, and that's not enough.

00:39:08   Right. And like, I think this is an example of what you're talking about of like just using

00:39:13   words, like the exchange of words is not enough to understand everything you need to understand

00:39:21   quickly enough. Like, it's not like when I use words with you, you can infer all of the context

00:39:28   that's needed. Like we can have a conversation and make, and make sense of it. LLMs can't do that.

00:39:35   If you like that topic, I, it's very technical, but I will recommend a, an episode of the Dwarkesh

00:39:43   podcast. Um, uh, uh, this week came out an, an interview, an in-depth interview with Ilia

00:39:50   Suskiver, uh, uh, that one of the co-founders and a former chief scientist at open AI. And one

00:39:58   of the big takeaways from the interview was that, uh, Ilia thinks that, um, LLMs are not

00:40:03   going to get us to, to, to real, real AGI or ASI. Um, uh, we must go back to the age

00:40:12   of research, as he said. And it sort of became a meme like, Oh, we're going back to the age

00:40:17   of research, but that's essentially the takeaway. Like we need to, uh, Ilia, I mean, he's not

00:40:22   an idiot, right? It's quite a, quite a smart guy. And he's saying, yeah, I mean, we probably,

00:40:28   you know, the whole idea being like, we, we can continue scaling these LLMs and throw more

00:40:34   servers and throw more compute power at them, but we're hitting a wall and we got to go back

00:40:40   to new research approaches if we really want to, you know, take the next step. And so they really

00:40:46   jumped the gun by just rebranding this as intelligence. I think, I mean, this is, this

00:40:51   was like my thing in the beginning, right? Where it's like, this is just machine learning. Like

00:40:55   that's what this is. It is just really advanced machine learning. I think there's a difference

00:41:01   between being smart and being intelligent. Yes. And this can be, this can be really smart tools,

00:41:08   but very frequently they are not intelligent. Yes. And, and, and to go back to what we were

00:41:16   saying before, Apple went on stage and said, we're going to use an on-device model to orchestrate

00:41:25   all of these apps on your device with all of these possible commands about you. The further

00:41:32   we get from WWDC24, the more insane that feels like what they did. It was like a fever dream

00:41:41   of an announcement. Like it just, it feels like, why did you do this? The, the amount of aggravation

00:41:48   and energy they are going to have to expend as an organization to fix that mistake is unbelievable.

00:41:57   Right. I mean, I, I think it's going to look, I think it's going to happen, but it's not going

00:42:03   to be good though. No, I think realistically, because they are now doing the thing that Apple

00:42:07   never likes to do for themselves is they are forced to a timeline. Yeah. That everyone knows

00:42:13   about. Right. And the, the, the, from my perspective, the most obvious reality is that it's not going

00:42:21   to happen on device like you think it should. Yeah. I think the private cloud compute model. So Gemini

00:42:27   will make sense of your request, will get some kind of encrypted and, you know, personalized sort

00:42:34   of map of all the tools that you have. And it's going to decide which tool to use and everything is

00:42:41   going to run in private cloud compute. So Apple will sell you. There are so many reasons I don't

00:42:45   want that to happen. The local model will execute those commands, but you're gotta, you must, you

00:42:52   must be connected to the internet for the original request to be parsed and, and understood. That's

00:42:58   the thing that I strongly believe. It's just frustrating because like one of the things that I really liked

00:43:02   about what they were pitching, I was like, Oh great. If it's on device, it's like not expending

00:43:06   the energy in a data center. Right. So like I can feel like unshackled from feeling bad. I'm just

00:43:14   using my phone and it's all on here. But if we're now like, I'm sending some graph to the internet every

00:43:20   time, like, and also then it's like, we're taking a step back again because Siri got better when it,

00:43:27   or like, you know, we worked so hard to get these things to be processed on device. And now we need to

00:43:33   go off device. But yeah, I think you're right. Like I don't, I can't see how they're going to do

00:43:37   this without a data center in the middle of it, but that's so not what they set out to try to do.

00:43:43   No. For so many good reasons. I mean, obviously happy to be proven wrong, but once again, like I

00:43:51   said, like I said, 10 minutes ago, maybe they made a major breakthrough internally that we don't know

00:43:55   of. I just don't think that the likelihood of that to me feels so slim. Like no offense to them, but

00:44:02   like, I don't think breakthroughs like this are really found when you have a gun to your head.

00:44:07   Yeah. Like this isn't that this whole situation has not created the kind of good creative

00:44:16   collaborative environment that people want to work in. So yeah. Fascinating to see what 27 is going to

00:44:23   be like. And that's to kind of wrap it around. Maybe, maybe it does end up being a snow leopard year

00:44:27   where on the surface, there's a lot of focus on stability, but underneath a ton of stuff going on

00:44:34   of Apple intelligence. We'll see. Yeah. Maybe the, maybe the snow leopard approach is actually

00:44:37   for the regional enhancement of Apple intelligence. Yes. We really need to make this work. We fixed the

00:44:44   bug of not having this entire set of tools by now having them. Bug fixed. This episode is made

00:44:51   possible by KRCS, one of the oldest Apple premium resellers and your one-stop shop for all things

00:44:58   Apple in the United Kingdom. And their month of savings is well underway because right now at

00:45:03   KRCS, you can build the Mac of your dreams in their black Friday event and save 10%. Shop the new

00:45:09   MacBook pro 14 inch with M5 for just 1439. That's 1439 pounds. That is a 160 pounds saving. You can make

00:45:18   your Mac truly yours and have the specification you want or take your pick from their in-stock models

00:45:23   and save 10% until midday on the 5th of December. You can even spread the cost of affordable monthly

00:45:30   payments. With free next working day delivery on in-stock items order before 3pm, there's never

00:45:36   been a better time to get a Mac at KRCS, your local Apple experts. You can get more information at

00:45:42   krcs.co.uk slash relay or click the link in the show notes. Representative 9.9% APR of 18 and over,

00:45:50   subject to approval and T's and C's apply. Please spend responsibly. Borrowing, borrowing your means

00:45:55   could seriously affect your financial status. Ensure you can afford to make your repayments on time by

00:46:01   the due date. KRCS are a broker, not a lender. Please go and check out krcs.co.uk slash relay right now.

00:46:07   That is krcs.co.uk slash relay. These are incredible deals and to get discounts on Apple products at any time

00:46:15   of the year is quite an incredible thing. So go check out KRCS. And also, you know, I have a preference

00:46:22   for this company. They're in the UK. I love it. krcs.co.uk slash relay. Thanks to KRCS for their support of this show.

00:46:29   Considering it is Thanksgiving weekend, I thought, why not talk about some things in the world of technology

00:46:38   that we are thankful for? And we will bring back the old round robin and take this in turns.

00:46:45   Would you like to go first, Federico? Yes. Yes. Let's see. Where should I start? I am thankful for...

00:46:54   I want to start with the iPhone Air. Okay.

00:46:58   I am really loving this phone. It's made me appreciate using the iPhone in a way that wasn't

00:47:10   true for the past few years where things have gotten kind of boring. And whenever I was holding my phone

00:47:15   for too long, I would get pain in my wrists. This phone has... I use it all the time. Like it's become... I no

00:47:22   longer use my iPad to read stuff in bed. This has become the device where I watch TV shows. I read

00:47:28   books. I read articles. I watch YouTube. I... It's such a pleasure to use this device. And sure, I do miss

00:47:37   more freedom with the camera, but I'm living with it. I'm learning to live with it, you know, framing my

00:47:42   shots better, taking different pictures. And, you know, I was given the opportunity if I wanted to, to switch

00:47:50   to a 17 Pro Max, because I also, you know, I have a review unit of that. And no, I, I'm sticking with the

00:47:56   iPhone Air because I, I truly love it so much as a, it's every day when I pick it up, it feels like I'm

00:48:05   holding an object that is not supposed to exist. Uh, and maybe it won't if it's not successful, but I, I enjoy,

00:48:13   I enjoy, I enjoy, I enjoy the fact that for this moment in time, um, um, I'm using Apple's most

00:48:20   futuristic iPhone to date. It's not going to matter to you if they cancel this phone. No. You will move

00:48:26   on to the folding phone. I will move on to the folding one. And also like, uh, no future failure

00:48:33   is going to take away this beautiful moment in time that is happening now. So whatever. I'm thankful

00:48:40   for it now. I'm going to piggyback on that. Cause I also love my iPhone Air, but I, I love it for

00:48:46   some of the things that I get out of it that are specific to having a second device. So I use, um,

00:48:52   my iPhone Air as it has a few purposes for me. You know, one, I do need a device for testing and

00:48:59   screenshots and stuff like that for my work with David and Steven across forward. But also I have

00:49:05   apps and experiences on the iPhone Air that I don't put on my main phone. Like when I'm at home,

00:49:11   my iPhone Air lives in my backpack. And if I need to, if I need it for some work stuff, I'll have to grab

00:49:17   it. Now I, all of the tech space, social media in my life is on that phone. Um, the slacks that I'm in,

00:49:25   it's on that phone. So my personal phone, when I'm not at work, when I'm not at a work computer,

00:49:31   it doesn't, I don't get kind of pulled into those worlds. And, and to me, for me has been a really

00:49:38   helpful, it's been very helpful to have that separation from some parts of my work life when

00:49:45   I don't want to have that. And like to not be drawn into things that I don't want to be drawn into.

00:49:50   So to kind of have that distance a little bit more has been really nice. And it's a project that I

00:49:55   started with the pixel fold, um, which was also to understand a little bit more about kind of like

00:50:00   the usefulness and utility of folding phones in anticipation of 2026. Um, but also the idea of

00:50:07   kind of separating these devices and yeah, I found it to be, I found it to be a real nice pleasure,

00:50:12   like to have this device and having it be an iPhone has been a better experience for me.

00:50:17   Like it's easier for me to be able to get to all the things that I need to get to in the way that I

00:50:21   want to. Um, and then also it's become pretty essential for the work on like Widget Smith and

00:50:26   put on it. So yes, I love it, but I really do love having that distance and of all the phones to have

00:50:33   as a second device, the iPhone air is a real good one. Real good one. Yeah. I am thankful for, uh, and

00:50:40   this is gonna sound maybe a little counterintuitive to some folks who listen to the podcast. I am thankful

00:50:46   for AI based search tools. Okay. Like web search tools inside AI apps, uh, because they're making me a

00:50:55   more curious person. And, um, I can tell you, honestly, I've discovered more blogs and more

00:51:01   websites in the past year, since like web searching in LLMs really took off than ever before. Like, uh,

00:51:08   if it look as the owner of a website that has been scraped over and over by these LLMs, I, I,

00:51:16   I original sin. I know, I know the pain, right? I know, I know, uh, I've learned to accept the current

00:51:24   reality. And I guess what I'm trying to say is that for me, there has been a positive upside in all of

00:51:30   this, which is I've discovered a lot of websites, a lot of blogs that I've then started reading and added

00:51:37   to my RSS feeds because they were quoted in their responses by these LLMs. And in general, I think

00:51:46   I'm now using this approach of like searching Google manually, the old fashioned way, plus

00:51:52   complimenting it with, uh, usually like, uh, uh, Claude or Chagipity. Like, uh, I think, you know,

00:52:00   especially Claude doesn't really have great web search. Um, I feel like Chagipity or perplexity,

00:52:07   you know, since they scraped everything, they also have better web search results, but I think

00:52:13   it's nice to be able to find the humanity in some of these scraped responses. And for me, that has been

00:52:21   finding blogs from people, like people who publish tutorials or even finding new YouTube creators,

00:52:28   right. I was looking up some answer and, and a YouTube video is cited as a source. And I'm like,

00:52:34   Oh, that's like, that's an interesting creator. And, and I've started following a bunch of people like

00:52:38   that. So yeah, it wasn't ideal originally, but now as a user, I'm discovering these people that normally I

00:52:46   wouldn't have discovered. And I think that's a, that's a net positive for me. So I'm thankful that,

00:52:52   that at least I was able to find some positive aspect and some humanity in, in, in all the web

00:52:59   scraping. I mean, I guess that part of it sounds like a somewhat similar to when web searching began

00:53:07   in a way, like I'm sure that there was a lot of concern initially amongst some people that like,

00:53:14   is Google not just going to give the answers and then nobody bothers really. And maybe that wasn't

00:53:19   initially, but over time for sure. Right. As Google just started before even the AI overview,

00:53:24   which is absolutely the worst implementation of AI that I've come across the AI overview.

00:53:30   Like it's, it's fascinating to me. Like I think Google is like undermining their advancements in Gemini

00:53:39   by providing that tool. Cause it is so frequently wrong in a way that I'm convinced Gemini would not be

00:53:47   if I just searched on Gemini. Like I think, I think Google should get rid of the AI overview

00:53:52   overview in Google search. Like I just, I did, I don't think it is helping them as an organization on

00:53:57   a bunch of fronts. Um, but anyway, that's, that's, that's a hill that I would die on. Um, but yeah,

00:54:03   I wonder like, uh, you know, I wonder if this is an experience that most people will have. Um, I assume

00:54:10   people probably won't work like you, but I also figure that people who Google something get one search

00:54:17   result and then that's the end of it. Like I can't imagine that, you know, they're often like

00:54:23   adding these to their RSS reader or whatever anyway, but I guess people aren't getting the clicks.

00:54:28   The monetization issue is a nightmare. We're not going to solve it on this podcast. Um, I will mention

00:54:34   actually, uh, I will mention chat GPT is the thing that I am thankful for in some of the uses that I've

00:54:39   had since becoming a dad. Like there's just so much stuff to try and learn. Um, and there is a ton

00:54:46   of information out there and me getting a good place to be able to kind of like understand milestones

00:54:55   and behaviors has been very helpful. Um, and it's been very helpful to me trying to like

00:55:00   whittle something down to then be able to start doing some work on it. Right? Like this thing has

00:55:07   happened. What could this possibly be? Okay, now I can go and start doing my own research. Um, I've been

00:55:13   finding it really helpful, uh, in that. And, and also in some of some other places in my life, um, like

00:55:18   some of the things that I've been doing for cross forward, I'm like coming across terms and, and I'm on

00:55:23   meetings with people. They're saying things that I don't understand. And like being able to kind of

00:55:26   have these ongoing conversations with chat GPT that are like, okay, so what does this term mean? What

00:55:31   does that term mean? And it given me that kind of rounded knowledge has been helpful. And then similar,

00:55:36   I mentioned this before, but like one of the reasons that I was happy to start blogging again

00:55:42   is because I, there are so many tools now, including chat GPT that can help me with my grammar and

00:55:48   punctuation. Cause it's just not a thing that I feel very confident in. Um, and so it gives me the

00:55:53   confidence to feel, uh, like I can, like I can publish things on the internet and not have people,

00:56:00   uh, uh, either a being very nice and correct to me about things or be making fun of me,

00:56:05   which is, which has happened a lot in my life, uh, by people on this podcast.

00:56:11   Yeah. Not me. Yeah, sure. I'm going to follow that up with, I'm thankful for Claude, which is my LLM of

00:56:21   choice. And it's, it's allowing me and John to, you know, we're a small team. Um, and, and we do a lot

00:56:28   of things. We have a lot of responsibilities and I feel like I speak for John too, when I say that,

00:56:34   um, we like Claude, um, because of the focus on, on productivity and it's sort of allowing us to,

00:56:42   to work faster and retain our, our creativity. Like we still like to write all of our articles,

00:56:51   the old fashioned way. Um, but it's amplifying the effect of the other things that we're supposed

00:56:57   to do. The administrative stuff that, you know, I think, I, I think I'm becoming a better business

00:57:04   person because of it, like making sense of stats and, and numbers in a way stuff and like, and have

00:57:12   conversations and, and take the amount of time that you would necessarily wouldn't get from,

00:57:18   from other professionals that maybe you couldn't even afford anyway.

00:57:21   Yeah, exactly. Exactly. And so just the other day, like I, I, I, uh, exported like a full database,

00:57:30   um, uh, of all our analytics for the past year and a half. And it was like 300 megabytes of CSV files

00:57:40   and a bunch of stuff like that. And, and, uh, first of all, I used a script that John created

00:57:46   with Claude to split up the CSV into smaller files because you have a file size upload limit in Claude.

00:57:55   Um, so I split them up in a bunch of files and then I asked Claude like, okay, this is a full dump of

00:58:02   our analytics for the past 18 months. Uh, I need you to make sense of it and to visualize it in a nice

00:58:08   way. And you got to find trends, patterns, interesting things about our audience, countries, time of day,

00:58:15   day of the week, like everything. And it created this incredible dashboard for me with real data,

00:58:21   right? Uh, because everything is based on the files and it's been like an eyeopening experience,

00:58:27   not just in what AI can do, but in what AI can do when, when it works with you with real data,

00:58:32   no hallucinations and just, you know, I don't know. I keep thinking of this idea of like,

00:58:39   it's, it's amplifying the things I do. It's not replacing me, but it's me, it's giving me like

00:58:45   a collection of small superpowers throughout the week. That's how I think of it. And so you think

00:58:51   of ChatGPT for your, you know, for your family life or for parenting. And I think of Claude for

00:58:55   productivity, but we're basically saying the same thing. Uh, I wanted to mention the athletic app,

00:59:01   which I've been talking about a bunch, um, which does such a great job of pulling in so much health

00:59:09   data for me and making sense of it. And they keep adding new features that I like. Um, like they just

00:59:14   added some, there's really enhanced their step stuff. So like being able to look at the steps that

00:59:18   I'm taking, look at my average heart rate per step and all that kind of stuff. And like while I'm

00:59:23   walking, I just, if Apple don't make something kind of close to this, I don't know what they're doing

00:59:29   over there. You know, like they have all this data and all this information. Um, and then this small

00:59:34   team at athletic is just crushing it. Like, yeah, it's fantastic. Like I'd been using whoop for a long

00:59:41   time and I'm, I've canceled my whoop subscription and that's the end of that. Like this is so much

00:59:46   better and I don't have to wear a second thing. Nice. I will mention these two together, the Aeon Thor

00:59:52   and SteamOS. Okay. Um, the Aeon Thor, it's a, it's a, what if the 3DS, but launched this year and

01:00:01   running Android? Yeah. It's a, it's a, it's a handheld. It's a dual screen handheld by this company

01:00:06   called AYN. Um, we spoke about it on MPC, uh, with John and Brandon a lot. It's so far my handheld of

01:00:15   the year. And, uh, the thing is I don't necessarily use it for dual screen emulation of DS and 3DS.

01:00:21   I mostly use it to play PC games on Android, uh, through this emulation layer called Game Hub.

01:00:29   There's a lot of movement happening on that front of like emulating Windows games on Android. Uh,

01:00:35   Valve itself is going to do it, uh, next year with the Steam Frame, a VR headset. Um, and I've been

01:00:42   using the Aeon Thor as essentially a more portable, pocketable, uh, inherently more easier to carry

01:00:48   around because it's a clamshell Steam Deck. It's basically a Steam Deck emulator in such a portable

01:00:56   and cute device. And it's perfect for indie games, indie games like Ball Pit or Ballotro or, or, you know,

01:01:04   just, uh, 2D RPGs, for example, uh, stuff that is not really graphically demanding. You can just emulate

01:01:12   on Android these days with these devices and the Aeon Thor, uh, I, I, I love it because I, I, I've been

01:01:20   carrying around this device everywhere with me, like even in my jacket pocket, like you can throw it in a

01:01:25   bag, put it in a backpack. It's a clamshell, so it's protected by itself. I love it so much and it's not

01:01:31   perfect. It could be more ergonomic. Um, you know, there've been some issues with the speakers, but nothing is

01:01:37   perfect in this, in this, uh, handled scene. Especially the prices that these products sell at.

01:01:43   Like they're so cheap relatively. Yeah. Um, that they're, the, the imperfections is actually an

01:01:49   important part of the pricing process. Uh, these things aren't built with the same rigor that an

01:01:55   Android phone will be, but they're as powerful inside. Yeah. So this one is good. I'll mention a gaming

01:02:01   device, the Nintendo Switch 2. Love my Switch 2. Um, it's been a good steady stream of games for the

01:02:07   Switch 2. Uh, and the hardware is wonderful. Um, and it's just really nice to have, to be able to play

01:02:14   these games that I love from Nintendo at last gen graphics. But like, it's good though. Like I like

01:02:22   being able to play games that look good and feel good and run great. Like that is just a treat. Um,

01:02:27   and, uh, I, I do really like the hardware. Uh, I wish there were more accessories like to make it more

01:02:34   feel ergonomically better for me. I did get the, uh, D brand, not the kill switch, but just the Joy-Con

01:02:41   holders, the Joy-Con holders, the Joy-Lock or whatever it's called. And that is, that's been made my

01:02:47   experience much nicer. I recommend that to people. Um, it's possible to just buy just the, the,

01:02:55   the kind of grips of the Joy-Cons and it works great. And then you don't have to do all the

01:02:59   nonsense that they want you to do with the like external dock thing that they've got. Um, so yeah,

01:03:05   I'm a, I'm a big fan of that. I am thankful for Notion, the popular note-taking app. Um, it's, it's,

01:03:16   it's, it's not necessarily because it's integrated with AI. I can use it from Claude. I can use it from,

01:03:24   from all my other AI experiments and it also has its own AI. It's just,

01:03:29   I love the idea of modular software. I love the idea of modularity in general as a person. I mean,

01:03:38   you know me, I love my iPad because of the accessories. I love portable devices because of how

01:03:43   intrinsically modular they can be, but Notion is modular software. It's like

01:03:48   the digital equivalent of Lego. You can build so many things and so many different things. And, um,

01:03:54   this has been the year that I've gone really all in on Notion. And then we went all in on Notion as a

01:04:00   team, um, at Mac stories. And it's, again, it's a, it's amplifying the things we do. It's accelerating

01:04:07   the things we do. Yeah. I have a database with John. I have a database with my web developer and we can

01:04:12   keep track of, of each other's responsibilities, but also I have my own space. Uh, it's where I sync

01:04:18   my highlights from my read later app. It's where I write my articles and then I export my articles

01:04:23   in, in an app that I will be thankful for in a minute. Um, it's just, it's, it's so,

01:04:29   it's fun to work with it. It's, it's, it can be fun to a fault. Uh, but I've tried my best not to

01:04:35   become one of those people that just tinkers with Notion. Um, uh, you know, forever I get work done

01:04:42   in Notion and, and I have become the kind of person that religiously, um, takes daily notes and then

01:04:50   reviews his daily notes. And I've been doing that for well over six months at this point. Um, and yeah,

01:04:58   it's, uh, it's, uh, I also know a whole bunch of people that work at Notion and, and I don't know,

01:05:03   it just, they give me the sort of vibe of people that care about making great software, uh, without

01:05:10   being scammy or, or, um, you know, what is the word that I'm thinking of? Um, that's, there's another

01:05:21   adjective, adjective that, that, you know, without hyping Notion too much. They care. They just want

01:05:28   to make it great. And you know, yeah, Notion is pretty great and, and it keeps getting better.

01:05:35   And there's the old AI, Notion AI discussion, but a lot of people don't like that. I personally do,

01:05:41   but, you know, just in general as a note taking app, as a space to save documents and collaborate,

01:05:46   it's pretty great. Yeah. Uh, I want to talk about the iPhone camera system. Um, so I, I love the iPhone

01:05:53   camera on my, uh, 17 pro. I think it's really nice. Uh, I think they've done a pretty good job overall.

01:05:59   Um, and like, I, I think it's great that we live in a time where I can take such incredible photos

01:06:07   of my family, like just with the thing that I have in my pocket. Like it's, you know, like with many

01:06:12   people who've become parents, I've come to appreciate just how good that is. Uh, but also I wanted to give

01:06:17   a shout out to continuity camera. Um, I use continuity camera on my, on my Mac, like the one that I'm

01:06:23   recording on now, like whenever I do video calls or video podcasts and stuff like that. Um, and it works

01:06:28   so good. Like it's not perfect. Like sometimes I have to unplug and replug, but it's, it works

01:06:35   shockingly well at incredible quality for what is going on. Now I do use it wired like, because I want to get

01:06:42   the best I can get out of it. Um, but it's, that's, that is a fantastic, um, uh, piece of software continuity

01:06:49   camera. Uh, it works, it works really great. And I'm able to use, I have an iPhone 15 pro max that I use as my

01:06:56   webcam and it is, uh, essentially flawless. Nice. Uh, my last one is going to be drafts,

01:07:04   the also note taking app, uh, by Greg Pierce on iOS, iPadOS and macOS. Um, this has also been the year

01:07:14   that I've gone all in on using drafts, uh, and using drafts for what it's supposed to be, which is,

01:07:19   you know, what's the whole thing, like where text starts. Uh, that's how I've been using drafts,

01:07:26   um, this year. And, um, it's become the place, the, the, the app that I open every time I have an

01:07:32   idea, I open drafts and, you know, run an action that I vibe coded with Claude to save it to my daily

01:07:40   note in Notion. Uh, but it's also a destination in the sense that when I'm done writing a draft of a

01:07:45   story in Notion, then I export it to drafts using another script that I created with Claude.

01:07:53   And, and drafts is the place where I put all the finishing touches, like all the markdown editing.

01:07:58   This is exactly how I did the entire iOS 26 review this year. It was entirely written in Notion, uh,

01:08:05   and then exported at the very end as markdown in drafts, where I put all the final plain text

01:08:12   finishing touches. But yeah, the freedom, um, there's, there's a common thread, I think between

01:08:17   Claude and Notion and drafts, uh, the idea of, uh, productivity software that is highly customizable.

01:08:25   Like that I realized is like my thing. Customization has always been my thing, but I, you know, especially

01:08:32   usually for video games before. And then I realized with the iPad hardware, I really liked customization

01:08:38   too, like, you know, customizing the iPad with accessories and all those things. But this year

01:08:43   I realized how customization can be a thing in productivity software as well. Like, um, the

01:08:48   combination of Claude, Notion, drafts, shortcuts, like all these things coming together to essentially

01:08:55   helping me create my own things, my own, my own commands, my own workflows. Um, I guess it was

01:09:04   always a thing for me automation, but now that I can be helped by these AI tools and now that these web

01:09:11   services like Notion can communicate with apps on my phone, you know, like drafts, for example, or Claude

01:09:18   because of an API, it's all coming together beautifully for me. Uh, where, um, I, I don't know. I, I feel like

01:09:26   drafts is a great example of take something simple and it can stay simple if you want to, right? That's

01:09:33   how many people use drafts. It's just a, just a plain text field. But if you want to, you can make

01:09:39   it so complex and it, and it can become like this really powerful text editor that can help you deal

01:09:46   with a 30,000 word review comprised of multiple chapters and pages in drafts. And so, yeah, uh,

01:09:54   Greg Pierce, you know, also a great guy and has been working on this for what, 12, 13 years at this

01:10:03   point. Um, yes, drafts is really, really, really, really good. Uh, and I'll finish up by, uh, saying if you

01:10:12   have an iPad mini lying around, super good for a nursery device for white noise. Um, and for us,

01:10:21   at least initially in the, uh, early months of our daughter's life, when we were just sitting in the

01:10:27   nursery or she was sleeping on us, it was good for watching TV. But in general, like white noise stuff

01:10:32   is super good on iPhones. Like we try to, to find devices like physical devices to do it. And they're

01:10:39   just really unhappy with the quality and the options that were available. Um, and realistically an old

01:10:48   iOS device is the way to go. Um, and I tried a bunch of things. Honestly, the thing that I landed on was

01:10:53   just Widgetsmith. Widgetsmith has really good like background sounds in it. And, um, I like, I think

01:10:58   brown noise is kind of the best sound that to, to annoy the fewest people. Um, and it works really

01:11:05   well for us and it's obviously available on all our devices. So that's what I've been using.

01:11:09   It's got a little widget of my daughter's sleeping face and we press it and the white noise begins.

01:11:14   Uh, so, but yeah, just in general, like it's just a helpful thing to have, you know, we've got all of

01:11:19   the, uh, the baby tracking stuff on it, like in the apps that we use like Mango, which is an app that

01:11:24   I've recommended in the past. Um, using a, like that device has been, has been super helpful for us.

01:11:30   So very, very happy to, to have been able to repurpose it. I guess I got just a sub thing

01:11:36   for me, reusing old, uh, iOS devices. I've had iOS devices. This is super good. Like finding places

01:11:42   in your life where these things can be used instead of, um, like, you know, if you don't have anyone to,

01:11:48   to pass them down to, or you didn't trade them in or whatever, like these are, it's good to have

01:11:53   these, uh, devices and use them for different things. Yep. Uh, but of course, most of all, we're thankful

01:11:59   for you for listening. Um, so thank you so much for tuning in to this episode of Connected. Uh, if you'd

01:12:05   like to find us online, there's a few places that you can do that. Uh, Federico is the editor-in-chief

01:12:10   of maxstories.net, uh, and you can find him on various social medias as well. Uh, you can find me at

01:12:17   theenthusiast.net and at cortexbrown.com. When you're hearing this, um, it provided us before

01:12:22   the end of Cyber Monday, uh, we're doing a 20% off sale or Cortex brand right now. It's our first sale.

01:12:28   So I'm intrigued to see how that's going to go. Um, I would like to thank our sponsors of this episode,

01:12:33   KRCS and Ecamm. Uh, I'll speak for Steven and say that he is also thankful for you, even though he's

01:12:39   not here. Then please do not take Steven's absence as him not being thankful for you, the listener.

01:12:43   He's just taken the week off, uh, to celebrate Thanksgiving with his family. We'll be back

01:12:48   next week. Until then, say goodbye, Federico.

01:12:51   Arrivederci.