717: Two Hot Streaks

We changed out the entire server that the site runs on, and no one seems to have noticed. So that's a good thing. There's only one post update error that I had to fix, which is pretty good. Yeah, that's, what was it? I don't remember anymore. It was like, oh, some page. It was like the store page.

The store page has like embedded PHP, and it was like a thing that could potentially send output after the HTTP headers had already been sent. And, you know, PHP sends it warning to the log that says, " Hey, you tried to do a thing, but I already sent the HTTP headers." That was literally the only thing I had to fix. And that was basically content-based, like because of the data in the store page that triggered that error.

So that's pretty good. And that's, I don't blame myself too much for that because at the time that I did the PHP update branch, I didn't have that store content, page content, so I couldn't have caught that beforehand. Yeah, I think it went very well. I was, I mean, it was heavily AI assisted, I will say, which is amazing for like server transitions and stuff.

Like, figure out everything on this server and make sure it also works on this server. And set it all up and help me transition and everything. Oh man, what a, what a help that was. It did miss a lot of stuff, I have to say, though. That's why I was telling you, and this was a miscommunication with us. I was trying to tell you, hey, before you switch it over, set up the new server, but don't change the DNS yet and let me hop on to the new server and fix all the stuff.

And you didn't do that. You switched over the DNS and said, "Now it's switched over, check it out." And it was missing a lot of stuff, but it wasn't that bad. Yeah, I don't work well with others. That's basically what that is. Just to get a little bit extra drama of me to jump in. Oh, this is already the live server. I guess I'm fixing it live .

It was mostly just like, like periodic jobs that I needed to set up and directory permissions that I needed to change and stuff like, yeah, but it was fine. It wasn't things that were causing the functional site to not work in the moment, but it was, you know, if I hadn't fixed it by that evening, there would have been problems.

Marco, Marco, a few days ago, we'll do it live. Basically. I'm just assuming you didn't understand what I was saying when I said, "Let me look at it before you switch over." I probably just forgot. That's the... I had the momentum going. Again, arguing for giving us access to your Unforgetful.

Yeah, well, that's... yeah. Did you say that publicly? I think that might have only been privately. No, that was like after we went offline or something, but ... Yeah, so John made a tongue-in-cheek plea for he and I to have access to Marco's Unforgetful so we could put things on his to-do list.

Which, by the way, you can do, because like Unforgetful just reads the system reminders. Well, system reminders do have like a shared list that you can set up. Don't tempt us with a good time, Marco. Yeah, but the problem with this, like we didn't discuss it seriously, but the real problem is like the potential for griefing is massive and that would just make Marco start ignoring it or delete that shared thing.

Yes. You know what I mean? Yeah. Because we would... Casey in particular would not be able to resist the temptation to put stupid stuff in there. Oh, come now. I would only do it once every two or three days. Mm-hmm. Only. All right. Let's do some follow-up and we should start with a great big thank you to anyone who purchased merch. The store is closed.

So far, I think I've only seen one person say, "Oh, no. Oh, God. It's me." But I did see one of them. And so you, the listener, did a pretty good job this year. I'm pretty happy about that. Usually, we see at least a smattering of it. And I believe I've only seen one. Or maybe you're just learning to be quiet about it.

Maybe you've shamed them so much that they don't want to write and say anything. That's it. That's it. That's it. Yeah. So anyway, what we did complete the sale, obviously, Cotton Bureau takes it from here. And we hope that they'll get you stuff by the holidays, but no promises. But we told you we would talk more about the ATP Pride and ATP Trans Pride t-shirts and what we're going to do with that money and all that stuff.

So this is mostly John's bag. So take it away, John. Tell us, how did this turn out? So people really responded to those two shirts. I think if you combine them, I think they are our single best-selling product ever. People just showed up for those charities. So we're very proud of our listeners.

Yes. Here are the preliminary totals we have. As we said, we're donating 100% of our profits from those two products to charity. How much will that be? Now, the disclaimer, they're preliminary because Cotton Bureau always gives some amount of time for, like, payments to settle and everything. So we don't have final, final numbers, but it'll be within the ballpark, within like 10, 15, 20, 100 bucks of what we say now, depending on how various failed payments get sorted out.

And the second thing is, of course, we have to wait for Cotton Bureau to send us the money before we can send it to the charity. So we're not making the donation today, but I wanted to tell people immediately, as soon as the sale ended, what the totals are. And we did, collectively as a community, we did amazing. So the total is just over $10,000.

Holy shit. Seriously? $10,071 in change, but that's going to, that's going to fluctuate as it, but it's going to be around $10,000. So that's around $5,000 for each charity. But wait, there's more because our listeners are incredibly amazing.

Mark Zarin, a long time ATP listener, was inspired by our Pride and Trans Pride products, and has pledged to match our donation to Trans Lifeline and Transgender Law Center. So it's not $10,000 in change, it's $20,000 in change. And he will make his donation when we make ours, when I give him the final numbers. That is flabbergasting and amazing.

That's awesome. That is incredible. Wow. That is incredible. Apologies, I've been a little sweary so far, even though I'm in a very good mood, maybe that's why. That is being incredible. And incredible, incredible thanks to anyone who bought one of these shirts, to anyone who wears one of these shirts, but most especially to Mark.

Mark, please feel free to reach out directly to me. I will send you a pile of stickers if you're interested. This is incredible, incredible work by everyone associated, but most especially Mark. Thank you to every single one of you. I'm genuinely, really and truly, extremely proud of all of us. I really, truly mean that.

Yeah. And also, we heard from lots of other listeners who donated directly to charities, either the ones we picked or other ones. They didn't want to buy a shirt because it was too expensive and shipping and tariffs like we understand, but they still donated directly to charities. It wasn't just Mark. Lots of people. And they didn't just give small amounts. Some people gave a small amount, but some people gave literally hundreds of dollars to these charities.

So it's been an amazing response. Yeah. Our listeners are the best. Like I know there's other podcasts out there. We love them, etc. Our listeners are the best. And I think, you know, you all out there, you prove it. You have a number of occasions per year where you prove it.

And this is definitely one of those occasions. Yep. I'm incredibly, incredibly thankful. I could spend the entire next two hours just saying that over and over again. I will try to shut up about it, but truly from all of us, this, I'm so incredibly proud. I mean, I don't want to get into politics, but like the last Tuesday was an incredibly strong and good day for the three of us here on the show.

I think it was probably a strong day for anyone who would buy one of these shirts. I am, I'm riding a real happy wave right now. And this, this feeling of what, what is it called? Hope, I think is what it's called, is one I'm not familiar with. And I'm just, I'm so incredibly thankful and proud of genuinely everyone involved. So thank you so much.

I also wanted to take note. We got some feedback from a listener of the show, Zinnia Bynum, who, as it turns out, is the major gifts manager for Transgender Law Center, which we had no idea was a listener. And that's freaking awesome. Zinnia writes as a member and longtime listener of ATP. I was delighted to hear that you all will be supporting Trans gender Law Center with this year's merch sale. I wanted to send you a bit more information about TLC. Here's a link to our one pager, which I'll put in the show notes, which goes over our work in a bit more depth. Since you are all dads,

I'd love to shout out our recent campaign that shows how dads can have a huge impact on their community. Also in the show notes, we have a toolkit for how to start these conversations in your neighborhood and how to support trans people in your life. We also have a short advertisement about this campaign and radio attached and video form, which will be in the show notes, if you'd want to share it out. Please feel free to mention that the email address, giving at transgenderlaw center.org. I will repeat that, giving at transgenderlaw

center.org. You can use that for anyone who wants to reach out with questions or if they'd like to know about more ways to give. Yeah, this is amazing that they were already an ATP listener, so that's great to hear. Yep. Really, really good stuff. So thank you. Thank you. Thank you. Thank you. I'm so proud of us for what we all did in September. I'm so proud of us for what we did in October slash November. I think we're going to stop asking you for donations for a little bit, but nevertheless.

This wasn't like strategically planned or anything. It just so happens to come after we did, you know, September childhood cancer awareness month, but it just, you know, that makes it all the more amazing that people reacted so strongly to this and didn't go, oh, I already gave to charity with the cancer thing. Do I have to do it again? No , the response was amazing.

Yep. So thank you. Thank you. Thank you. Thank you. All right. Let's talk some nerdy stuff. Let's talk MacBook Pro with ProLED XDR display. HTX Studio, which is becoming our favorite YouTube channel, has a video about the MacBook Pro and I believe the iPad Mini was stuck on the end of that video.

But interestingly, at three minutes and 17 seconds, again, link in the show notes, they note that the front facing camera has software stabilization to combat screen wobble caused by touching the screen. And they do a demonstration of that. It looks pretty good. That was in that was in like the bullet points that we read on like the release day.

I'm like, what does that mean? This is what it means. I forget what the wording was. It was something about advanced camera stabilization, blah, blah, blah. This is what it is. It's when you poke the screen, we do on the fly software stabilization to try to reduce the wobble. Obviously, if you poke it hard enough, the screen is still going to wobble.

But that's that's pretty neat. Additionally, there is the illuminated logo on the back of the display. And we're going to talk about that on and off for a while now. But anyways, there is a knowledge base document, which we will link in the show notes.

And it reads to make the Apple logo dark in all lighting conditions. You can turn off its illumination using terminal. And then you have a defaults right that Apple provides. We'll put that in the show notes or if not the defaults right. We'll at least put the link in the show notes. This change affects all macOS user accounts, including when starting up for macOS recovery.

When migrating to a new Mac, the setting transfers to the new Mac. To turn logo illumination on again, you do a defaults delete. Reinstalling macOS also turns the illumination back on. So when I see a tech note like this, this is officially from Apple. We're going to get to some third party stuff and people are experimenting. And last episode, we were like, oh, people are exploring.

And they think, you know, I think it was Guy Rambeau. We're like, they found the private framework and blah, blah , blah. Apple put out a tech note basically at the same time they released these with a pseudo defaults right command. And it's like, Apple, if you find yourself doing that, just put a damn GUI in settings for it.

Like how long before this GUI appears? It seems so clear that they like maybe they ran out of time , couldn't fit the GUI in the whatever 27.2 or release. But they're like, we just want people to know. Because it makes sense. Like if you're like in a studio type of environment or like you're shooting on a movie set or something, sometimes you actually would do want to turn that logo off.

so light isn't shining in a particular direction. So surely there's a way to do it. And there is. But Apple's like, yeah, just defaults right until we can get to the GUI for this. There is an app called Backlow, B-A-C-K-G-L-O-W, that lets you customize the glowing Apple logo.

Joe Rosignal at Mac Rumors writes, a new Backlow app created by designer Juan Buiz lets you customize the illuminated Apple logo on the lid of the new MacBook Pro models. The free version of the app allows you to manually adjust the brightness level of the Apple logo independent of the MacBook Pro's display brightness level. You can also turn it off entirely with a single click rather than having to use the terminal.

There you go, John. Another free option is a Breathe mode that makes the logo's light gently pulse, similar to the sleep indicator on older MacBooks. Rest in peace. God, I love that so much. A one-time payment gives you two additional modes. Music mode that makes the logo's light move to the beat of the music you play on your Mac, while agent's mode makes the light pulse while Claude Code is working and blink when your attention is needed.

It's clever. Speaking of lights on the outside of MacBooks, like the breathing one, that used to be the light on the power adapter, which, you know, I think it's, I suppose they could still do it today because MagSafe has that. And there are some USB-C connectors that have little screens on them and light up. But anyway, that made sense with the breathing because it was the power supplying thing.

Maybe the logo doesn't make that much sense as breathing. But anyway, the other lights that used to be visible on the outside of Mac laptops, where they had like a series of lights for the battery to show you what your charge remaining was on the battery in units of like quarters or whatever, it was like three or four little tiny lights, just like on your little DeWalt battery.

If all four lights are lit up, your battery is full. If three of the four lights are lit up, it's 75% full or whatever. It's very crude, but it was a way for you to check your battery charge level without opening your laptop. Uh, and they were so small. They were tiny pinpricks and they wouldn't be on all the time.

You would press a button or something near them and they light up. And I've just been waiting for someone to discover those in the past. And I go, you know, we could add those pretty cheaply. They don't take up a lot of room. We could do some clever thing where the button is like one of those invisible capacitive buttons. It doesn't look like a button or something.

And I do think that does serve some function. If you're just peeking at your laptop and your bag and those lights were along the edge of it, just to sort of touch a thing and see three out of the four lights light up and know whether you need it to charge or not. I don't think it's essential, but it is in that same family of cute, interesting features they added back when they thought putting stuff on the outside of computers was a good idea.

I would argue the battery ones, even though personally I would love to see them again, I would argue that's a lot less of a big deal now that these things have effectively infinite battery life. But you're not wrong. But when you're traveling, though, like you are going to run out of battery eventually if you're working it hard.

And like laptops in particular, like why wouldn't you just open it up and check the batteries? It's in your bag. Like it's closed up. Like you're peeking. It's the edge of it just poking out of your backpack. I still think it has utility. No, I agree with you. I definitely agree with you. I just think it's less critical than it used to be. Nevertheless, Austin Evans in his video review says that the newly divided air slots on the sides of the MacBook Pros are less sharp than they used to be.

And he was very excited about that. We'll put a link in the show notes. Lots of reviewers have said that. It's not just him. So they look less sharp in the pictures and everybody says they run their fingers along them. They are less sharp, which I think is great because I said last time when I revisited the old models and when I was looking at the Duo in the Apple store, I was struck again by how sharp those slots are on the old models.

Yeah. In fact, I actually went to an Apple store yesterday and handled these for the first time. And that was one of the first things I did. Of course, I pick it up. I twirl it around, see how light it is. And immediately, the next thing I do is run my finger along that bottom edge. And it was actually useful because, of course, I've been using a 14-inch for years now.

I know pretty well how they felt. But they're still selling those models. And so immediately next to the new one I was playing with was one of the old models. So I could compare them side by side. And yeah, the difference is night and day in terms of the sharpness of that slot. You can, if you really dig your finger into it, you can still feel a little bit of a sharp edge.

But it's nothing like before. Austin also noted, regarding the fans and screen cooling, he also says that he was told that the exhaust no longer blows up. Now it blows down and out. And as someone who runs his computer clamshell most of the time, this sounds pretty good.

Yeah, Max Tech had some, in their review, took out their little thermal camera to show the front. Like the laptops are open in front of you. So you can see the keyboard and the screen. They showed the old model and the new model. And on the old model, you can see two big, bright, hot stre aks coming out and going up the screen.

You know, they're just above the keyboard where that crack is between the keyboard and the screen. And you can visualize where the fans are inside there. And they're blowing hot air up and out across the screen. So you get two hot streaks going up the screen. And then you look at the new model and you see the same heat under the keyboard where in the SoC and everything is.

But you do not see hot streaks going up the screen. Instead, the screen looks much, much cooler. Yep. So then Geeker One used smoke to show the airflow change. This is very slick. I didn't see this when I was studying. Yeah, this was awesome. I'm going to have to go back and watch this.

But we'll put a link in the show notes to that video as well. Yeah. So it's funny. Like, so first of all, Geeker One was another, I don't know if it's a Chinese YouTube channel, but it's a non-English speaking YouTube channel that's dubbed into English for us. And the side views of these computers, it looks like the computers are farting.

Let's be honest. It looks like they're shooting out. All right. So how did they do this? They went to Taco Bell tonight. Yeah. As they noted in the video, these were like not ones that they had purchased, but they were the review units. So it's like, well, we can't really crack them open because they're not ours. Apple kind of frowns on you taking apart their computers.

They loaned you. But apparently they felt fine about letting it inhale smoke through its input ports and shoot the smoke out of it. And so you compare it. They were comparing the 2025 model to the OLED one. And the old model has fart gas coming out the back, you know, basically shooting out, shooting directly out the back, like along the table, you know, away from you.

And also pretty much an equal amount of smoke going straight up the screen. Right. Because, you know, some of it goes out the back, but some of it goes up the screen. And the new one has barely any going in front of the screen and a much, much bigger plume shooting out the back.

That's what Austin Evans was getting at is that the new ones fart out the back way more than the old ones did. And additionally, Geeker Juan noted that in the old model, if you open up the screen really, really wide, first of all , the fart gas starts going down at an angle, which is funny .

But also it blocks a lot of the back port, the one that shoots out the back of the laptop, because the screen is blocking it. As you move the hinge down, the screen blocks more and more of that port. And even more hot air goes shooting out across the surface of the screen. With the new ones, the hole in the back is bigger.

So even when you open the lid to its maximum openness, still there's enough room for all the fart gas to come out the back down at an angle. And there's still barely any going up the screen. So that is the cooling thing. It is sending less hot air across the surface of the screen , because before you never had occasion to touch the screen because it wasn't a touch screen.

So however hot the screen got, you didn't notice that your fingers were never going to touch it. But now they are. So they're trying to focus all the heat out the back instead of going up the screen. There's also a chart that they provided about power use. Can you tell me about this? Yeah.

So that's the other thing people were doing is like temperature testing. So with all this, you may be wondering, oh, is that comprom ising cooling or anything? Apparently, the new models, the new OLED models, have better performance when the lids are closed in the default fan mode.

So if you close the lid on both of them and you run them, the thinner, lighter one has better performance. And they measure this by saying basically how much power does it consume? Because that's that's the proxy for like, is it thermal thr ottling? Like if it's not throttling, it should be using the maximum amount of power.

You just measure how much power it's drawing when they run whatever benchmark or whatever. So the chart is hard to understand because some of the labels are not in English, so I can't read them. But according to the narration, which is in English, the top two lines are for the lid open and the top two bottom lines are for the lid closed.

And in both lid open and lid closed mode, the OLED model is drawing more power, which means it is thermal throttling less than the old models, which is pretty amazing. Again, it's lighter. It's thinner. It has fewer places to eject hot air, but still it is thermal throttling less than the old one.

Now, I will point out that they did when they tested it, they forced the fans to max RPM. And we'll get to fans in a second. When they forced both fans to max RPM, they both perform identically. So apparently both have cooling systems that are capable of cooling it to the point where they either don't thermal throttle or thermal throttle equally.

But no one runs their fans at max speed all the time. Like that you use the software that the computer comes with to throttle those fans. And maybe the throttling curve on the new one is better than they were on the old one. But the bottom line is if you just use it out of the box the way it comes and allow it to control the fan speed, you will thermal throttle less with the new ones, which is amazing.

Because remember, that's the same exact SOC in there in a thinner, lighter laptop. Yeah, it is bananas. Apparently, there's some bad news about fan speeds, though. No, you think this is bad news, but this is the thing. So they measured it. They were comparing, again, the 14-inch old to the 14-inch new OLED one.

And the 14-inch OLED new one has faster, higher fan speeds. And actually, they're different max fan speeds. So the old ones were like 7,800 RPM for both fans. The new one is 8,400 RPM for the left fan and 9,000 RPM for the right fan. You're like, wow, that must be terrible. They must be louder.

No, they're higher RPM and also they're quieter. 50 decibels in the old model and 48 decibels to the new one , which is amazing. Like, I don't even know how they did this. That's why this channel was like, we can't wait to crack these open and see, did they get a better model of fan? Do they have different fan blade designs?

But the new fans can go faster. And at the faster speed, they have lower noise level. They did a bunch of experiments. They're like, let's just make them make the same noise level and test the cooling. Or let's make them run the same RPM and then test the temperature. They really did a good job with these laptops.

They basically just took the old ones and bettered them in every possible way. This is such good news so far. Again, we had a lot of damage in this podcast about the previous screw-ups where it just seemed like every time Apple touched certain things about the Mac, we were like, oh, no, don't make it worse.

Because they kept making it worse and had a history of making it worse and telling us they were making it better. But it was actually worse or the same. We're staying broken. Anyway, not that we're upset. But this is such an amazing progression. Because in many ways, this is actually really boring.

And that's what we want. In many ways, it's like, yeah, it's like the new one, but better. That's it. In many of these ways are pretty small. Many of these ways are pretty big. That is exactly what we want from this product line. And it's such a weird thing. Like the fact that they decided or whatever.

Like the fact that we don't have M6 Pro and M6 Max, right? It gives us this weird opportunity. Because when they do change the SoCs, it's like, well, how much of this is attributable to a better design? Or how much is it attributable because a new SoC on two nan ometer? But they kept the SoC the same, which, again, is not good.

That's not a good thing. But it gives us this opportunity to say, have they improved fundamentally the design of this laptop? Or are we just seeing the benefits being reaped by a new So C? And now we have apples to apples to parents. They have the exact same chip in there. And it's clear that even with the exact same chip, they made it faster, lighter, quieter, better in every possible way.

Oh, I can only imagine what this will be like when it has an M7 Pro or Max or whatever at two nanometer in it. So, John, I know we're kind of in the middle of follow-up, but I can't help but ask. I know you're not a laptop person. And I know we all concluded, two-thirds of us concluded at least, that now is not the time to buy one of these.

But if you were to buy a new laptop today, would you spend the $400 and get the OLED one or would you stick with the previous one? Why am I buying a laptop? I don't know. Just go with it. What is my hypothetical? I don't know because you're… Am I forced to use it instead of my desktop or am I looking for a portable thing or what am I doing here?

Pick your poison. Whatever hypothetical will actually get you to answer this question. I mean, when I do muse about potentially getting a laptop, like because right now when I go on my Long Island vac ations, I obviously use a laptop to do all my photo processing and stuff. And I basically use one of the kids, quote unquote, kids laptops.

They're not theirs. We bought them. But anyway, like they're like the laptops we bought them for college, like the M4 MacBook Air and M2 MacBook Air. And I think, well, we'll probably give those to the kids to be their laptops when they finally, you know, leave the nest someday if that ever happens. So I might need to buy myself a laptop.

And when I think about that, I'm like, oh, an M6 MacBook Air. Like that's what I'm thinking because for the purposes I have, it's like I want a small light laptop that I use one week a year and there's no way I'd want to pay it and spend eight grand on something like that. You know what I mean? And hey, if I hold out long enough and or if they don't leave the house long enough, maybe I'll get like an OLED MacBook Air in 2032 or something, right?

So that's what I'm currently thinking. I wouldn't spend the eight grand on these, even though they are amazing and accomplish great things. You know I wouldn't spend it. The screen corners. Can't do it. Can't do it. Won't do it. And that's both the OLED and non-OLED, right?

No, just the OLED, I think. No, I think you're right. I think it is just the OLED, actually. Just wait until all their desktop monitors have rounded corners and you're going to really hear me go ballistic. But anyway, I'm hoping they will snap out of that one before it comes to that. How do you snap to grid when it's a curved corner?

You don't. You just make the UI squirm out of the way. It's great. Make one corner of the window and have a different radius. Apologies for the digression. I couldn't help myself. Coming back to the ProLED, Max Tech talks about, well, where did all the weight go and noted that the aluminum corners are carved out a bit.

And there's, I think, some other things that he had noted, if I recall correctly, when I was watching this earlier. He had some guesses. He was like, oh, it looks like the batteries are physically smaller. But when I was watching the video, I'm like, this is great, Max Tech, but we're just waiting for the iFixit video because they're going to put things on scales. And they did. Yep. So where did the weight go?

Indeed. So that brings us to the iFixit video. And again, all these will be linked to the show notes. The 14-inch ProLED compared to the M6 MacBook Pro, the battery is 2 grams lighter. The trackpad is 8 grams lighter, but it's also not as tall.

The chassis is, excuse me, 92 grams lighter. The bottom cover is 44 grams lighter. The lid is 101 grams lighter. So the total is 203 grams or 0.55 pounds. There you go. That's where the weight went. That is all of it.

Most of it went, shockingly, they saved the most weight in the lid, like the screen, the backlight. The lid is physically thinner. And second place is the chassis, that part that had the corners cut out of it or whatever. But yeah, a little bit here and there adds up to a half a pound in the 14-inch.

Yeah, and even like the, you know, iFixit broke it down further, like the bottom cover, not just like the whole frame and chassis of the bottom assembly, but the actual bottom cover, like the part that if you unscrew the screw is that part that comes off. That was like half the savings of the lower chassis savings right there.

So like... No, that was in addition to the chassis savings. That's why I have it listed. Chassis 92, bottom cover 44. Yeah, it's, I mean, this is an incredible savings due to a lot of, you know, different manufacturing and, you know, carving out of the aluminum. And I'm looking forward to hearing more about this because, like, some people had speculated that maybe they were using the same process they were using for the MacBook Neo.

Because remember when the Neo launched, they said that they , basically they said that they found like a cheaper way to make the aluminum. But the Neo is not light. The Neo weighs as much as the MacBook Air. So I don't think that's what it is. Like... The bottom cover might also be stamped.

But like, when you have something that is flat, like the bottom cover that just unscrew the screws and it comes off, whether that's stamped or machined almost doesn't matter. It's like if I have a flat sheet of aluminum, if I machine it flat or if it's stamped flat, it's just flat. Like, the difference in the final product is very minimal.

Whereas if you have something that is a tub or a frame or something, that makes a big difference whether you're stamping that or machining it in terms of the strength. And how precise you can be with removing material. I'm looking at like the corners and stuff. It reminds me of the things car manufacturers do to like, especially with like advanced computer design.

That's why they have all the suspension arms that look like they're like bones from an animal. Because they just put into a computer, they say, here are the stresses. Here's the durability we need from it. Here's where it's connected. Here's all the stuff that's going to connect to it. Here's how much the car weighs. You know, they put it into a machine. They say, we just want it to meet these specs.

And it just comes up with the least amount of metal you can use to meet those specs. And it turns out, perhaps not surprisingly to biologists, that it ends up looking like a bone. Because our bones are, you know, evolved over millions and millions of years. And the animals with slightly less efficient distribution of calcium in their skeletons had a slightly lower chance of reproducing.

And so we get things that look like bones. And so that's why you get cars look like bones. And the inside of this chassis doesn't look like bones yet. But some of the carved out portions are like, I can't imagine a person choosing to do that. And how do they know that that's a safe thing to do? I don't know if they're using the same design techniques that they would use for car suspension parts.

But that's the direction this kind of stuff goes. If you really want to save weight, rather than having a human guess about where you can remove stuff or just do some stress diagrams and like, let's carve some stuff out here. With computer-controlled milling machines, you can come up with almost anything you want. And then it will mill it for you.

And they're traveling that path with these laptops, it seems. This reminds me, I think we had this in the show notes, but never quite got to it. And I think it even fell out of overtime. But Ford, when they've been designing their new pickup truck to replace the Lightning, there was an article on The Verge or something about how they were obsessive about efficiency.

And how, you know, giving the car like a half mile more range. Well, if you do this 500 times, that's 250 miles more range . Like, obviously, I'm being hyperbolic, but you know what I 'm saying. And I feel like this MacBook Pro is a similar thing.

Like, where can we get a few grams here, a few grams there, a few grams here? And suddenly, that adds up to 203 grams, which is, as we said before, over half a pound. Yeah, it's like Gordon Murray Automotive. Gordon Murray famously obsessed with weight in every single screw and every single piece and everything in his car. And it's the same strategy.

It's like, you don't care about this one screw now, but there's 300 of these in the car. And if you just repeat that over and over again, it adds up eventually, which is why his cars are very lightweight. You got it. All right, continuing on. The hinges, iFixit found.

It's hard to paint a word picture about this. But if you imagine the hinge, there's like, there's a lateral bar that has what appears to be like rings on it. John, I think, in the show notes theorized that they're spring windings, which makes sense to me. But on the MacBook Pro without ProLED display, there's something like six or eight of these.

And on the MacBook Pro with the ProLED XDR display, there's two more of those rings. So the hinges are very, very similar looking. But there's two more of these rings, which I presume has to do with it being a little more sturdy. Yeah, there was some speculation like, oh, the hinge and the new one feel so good.

They must have totally redesigned it. It looks like it is exactly identical with two more rings. And the two more rings translate into it being a little bit stiffer, which makes sense. Like they didn't totally redesign this hinge, but it seems like it's just the old hinge with more stiffness. And, you know, that's good.

You want it to have a little bit more to defend against it. But it's not like they went back to the drawing board and came up with an old new hinge, it seems. Honestly, like when I, knowing about this, you know, the redesign hinge, when I was in the Apple store, put them side by side. I poked them both because I'm like, it's not my screen.

I don't care if I touch it. So, you know, I poked them both and I, you know, moved the lid back and forth a few times. And, you know, I could not tell a difference between the two hinges, to be honest. If you poke either one of them, it does like jiggle a little bit. Like if you think the hinge is going to be super rigid and you can like touch it and have it not wiggle.

Nope, it's not. But neither was the old one. And to me, they felt exactly the same, to be honest. All right. Then coming back to iFixit, they also took apart, of course they did, the logo light. And what they found was there's four LEDs that are all along one edge that shine into a diffuser.

And that's how you get the glowing logo. It's edge lit. Yeah. This kind of puts a damper on the hope that you'll be able to do RGB stripes or anything with the LEDs because it's just not the way it's set up. It's just edge lit into a diffuser with a little reflector behind it. I don't think we mentioned this last time, but the light up Apple logo, when it's not lit up, it looks like the past mirrored ones.

Because it's basically like a one-way mirror or a two-way mirror. Isn't it fun how one-way and two-way mirrors, both of those phrases in English describe the exact same thing, which is a mirror. If you look at one side of it, you see your reflection. And if you look through the other side, you can see through it. Like you see in the TV shows when they're interrogating somebody and people are like, yeah, yeah, yeah. That thing.

Anyway, that's what the Apple logo is. So when the light is off, it actually looks very similar to the old non-light-up ones. Not exactly the same because the old non-light-up ones were like shiny metal and this is a piece of mirrored glass. We are sponsored this episode by Base44 because every builder needs a base.

Now, when your idea of your product is actually working and people start showing up, when all that stuff happens, you start feeling traction and everything. The system behind it has to become real too. And usually that means stitching together a pile of separate tools. Base44 takes a different approach.

It's an AI platform where you describe the software you need in your own words and get a working build, whether that's a customer portal, a dashboard, or the system your whole business runs on. But this is not a mock-up or a prototyping tool. You're actually building the real thing with Base44.

The actual product, the website, the app, the internal system, the AI agent, and you can keep shaping it as the work changes. And all this is done by describing it in your own words. You lay out the pages, the data, who uses it, what happens next, and what the tool should remember. And Base44 turns that into software you can run live.

The hard parts are already handled for you. You don't have to think about databases, logins, backends, or hosting. All of that's built in. So you can focus on how the thing should work. And this is all not a stack of separate tools trying to integrate with each other. This is one build. So your product isn't spread across a whole bunch of different tools.

You build them all together in one place. That is Base44. And you can be ready when people show up. It helps you turn interest, demand, or a clear workflow into software that supports the next step. And it's made for how you work. So it's made especially useful when you need something very specific to the way that you serve your customers or what you sell or what you do.

You stay in control. You use it. You change it. You add what's missing. And you keep building. So go to Base44.com slash ATP for 30% off the Builder plan for your first year. That's Base44.com slash ATP for 30% off your first year.

Thanks to Base44 for sponsoring our show. All right. Then we've been talking on and off about these Chinese YouTubers over the last few weeks because we're now getting exposed to them. And Elijah writes, like HTX Studio, Tim Pan also did a Tim Cook collaboration.

We'll put a link in the show notes to it. Elijah continues, also regarding MediaStorm's size, they also run a clothing brand, Storm Crew. My roommate from China has one of their shirts. MediaStorm is the HTX people. Is that right? I forget who's who now. No, I think that's Tim Pan.

Okay. My bad. Adiko writes, Tim Pan is more like a Linus, Linus, Linus. Where did that come from? A Linus Media Group operation from China. They have courses, massive e-commerce, clothing, and started as a TV production company. HTX is a much smaller operation, similar in size to MKBHD.

I don't know if that's really true. Does MKBHD have a couple of people? I think they're bigger than MKBHD, but I think the idea is that Tim Pan is like bigger. Does Linus Media Group that much bigger than? Anyway, this is what they said. So I'm just reading some feedback. Watching their behind-the-scenes videos, they just seem much more focused on having big production quality, different style.

But Geeker Wan, as we mentioned earlier, is also a good Chinese YouTube channel. Then Wairon writes, HTX Studio just posted a Q&A video that answers a lot of your questions about what HTX Studio is and where it came from. And I did watch this, and it was very good. Yeah. Like, it's so clear. Like, the reason we're seeing these now is because they're, first of all, they're posting their things to YouTube before they were doing whatever the Chinese YouTube is, like Billy Billy or whatever that's called.

So we wouldn't see them. And second, they're dubbing them in English. Like, they're going after an English-speaking audience by taking their existing content and either, like in the case of HTX, the guy speaks it in English because he can speak English. And then, like the other ones, like the Geeker Wan video, they just dub it in English to get an American audience.

So HTX Studio, I guess this is the question everybody had when they saw the videos. Like, we all had, like, who is this person? And where did these videos come? Because they just, for us, they came out of nowhere because we weren't watching them on the Chinese YouTube or whatever. And that's what this video is about.

Who are we? How do we get where we are? And it is a very MKBHD kind of story. In fact, he said he was a kid growing up watching MKBHD videos. And he wanted to be a YouTuber as well. And he mentioned a few more of his influences. And yeah, he was just a kid who wanted to make videos and started to be successful.

And it snowballed. And yeah, watch the video. It's great. And it's funny. All right, then with regard to Marco and I's best friend, the Apple business specialist, Rich Richman writes, the business specialists at Apple are fairly useless around phone launches. They can try their best to advocate for business customers, but there's not much they can do.

In fact, at the store I used to work at, management would take phones by the hundreds to sell to resellers that could have gone to business customers, but they'd count them as quote unquote business sales. Source, I was one for five years. I'm writing a book about my time at Apple Retail. Read it next year. There you go.

An ex-Apple business specialist writing in to say, it's a farce that he was part of for five years. Then we, I'm not sure why, John, you wanted to include these in the show notes, but I'm going to read them and you're going to talk about them. We had some suggestions for John Ternus, the CEO. Mattias writes, if I were advising John Ternus, here's a software bet that I think would put Apple in the lead.

Patch continuously without restarts. AI is about to surface a steady stream of severe bugs, many exploited before they're announced. Fixes should install automatically as often as hourly without interrupting the user and usually without a reboot. macOS already has the bones for this. Is it XNU? Do you know?

What's the verbalization? I've only ever read it. All right. XNU sits on mock, which was designed to run OS services as replaceable user space processes and driver kit and FS kit already moved drivers and file systems out of the kernel. Finish that move and let components update independently.

And most fixes become a background restart of one process. Only kernel fixes would need reboot and those would be rare . So we think. Whoever gets there first owns the security story for the next decade. Yeah. So the reason, before you get to the next one, the reason these are in here is this follow up on us talking about Ternus as CEO and people were all ready to say, here's what I think about Ternus as CEO.

And this one was a particular suggestion I thought was worth talking about because we are currently in a time when security bugs are being found and people are scrambling to patch them and we're scrambling to apply patches. We didn't talk about this story, but most recently with related to macOS, one of the SSL certificate, whatever authority companies had hilariously insecure private keys on some of their SSL certificates that became exposed.

The private key and one of them was for the number four. Yeah, this was amazing. That was the private key. They, they, uh, had to pick a 256 bit private key and thought that meant like a number, a number between zero and 255. Yeah.

And so they picked like four as like, the other one was 16. Yeah. Like this is some bad secure. Anyway, that was discovered. And so there's this mad dash to be like, Hey, if your system trusts those keys, which it used to, there's a part of like the root authorities for, I don't know the terms of this, but like these are, these are certificate authorities that every device trusted because they were part of like the trusted set of root authorities.

But obviously once it's discovered that the key is four, that's trivially exploitable. Anybody can imitate that blah, blah, blah. So there's this mad dash to run around and say, make sure none of your devices trust those keys anymore. Uh, and that has been difficult for Apple.

They patch pretty quickly. Like the, the system, like certificate authority stuff, you know, so like Safari and whatever would work with it, but there's other sources of that information for command line tools. Like there's a, there's a file and Etsy something, whatever . Like there's in fact, multiple sources on a default Mac install.

One that like the Python tools trust and one that like command line tools like curl trust. And Apple didn't patch those. And not only did they not patch those, I mean, I, I patched them manually. First I did it manually. And then people were releasing scripts and GUI tools to patch it or whatever. But not only did Apple not patch it, but if you patched it manually, like I did, and then you updated it to 27.2, they undid your manual patch.

So their relays process is so far behind being caught up on security flaws that they're undoing fixes that people apply. That is too slow in the current age of fixes, especially since this is not like, Oh, we have to figure out how to fix it.

And we have to come up with a, no, this is very straightforward. You just literally delete the, the, uh, the compromise certificate authorities from a file that say, don't trust these anymore. You just delete lines of text that fixes immediate, obvious , straightforward. Everybody knows what it is.

And they couldn't get it through their release system to get that out to people in a time and fashion. And they couldn't even get it out to the point where like, you know, Oh, we're about to release 27.2. Let's at least make sure we don't undo their changes. No, they couldn't do that either. So they're not quick on their feet.

And two things. So, so this particular feedback of like, Oh, they should be able to like do it without rebooting. And, you know, use these technical, uh, you know, features of the OS to, uh, apply patches faster without disrupting the user. The first thing is, as is clear from the, uh, certificate exploit thing, the, the first hurdle you have to overcome is bureaucratic and not technological.

There's nothing technologically stopping them from patching the .pem file in some Etsy directory. And they couldn't do that. That's not a tech problem. That's not a microkernel problem. You don't have to reboot for that. That is a bureaucratic problem. And if you can't solve the bureaucratic problem, all the technical stuff doesn't matter at all.

As for the technical stuff, uh, the mock microkernel is the underlying thing in all of Apple's OS's, but, uh, it's the XNU or XANU or whatever. That is the name of the whole kernel. And then as mock plus the entirety of BSD welded to it in the same memory process.

And if there's a bug in any of that BSD kernel code and everything, and I think all the, like the BSD, you know, there's, it's a huge amount of code . It's not just mock. If there's a bug anywhere in there, it's a reboot. Uh, now there are ways to patch even this type of system without rebooting, but historically Apple has not.

Just not taken on that task. There are a lot of things that you could look at the way Mac OS is designed and say, you should be able to patch that without rebooting. Right. And the reason Apple doesn't take it on is they're like, well, maybe, but what about all the apps that are currently running?

What about all the frameworks? Does everything in the entire stack handle this service disappearing and coming back in a graceful way? It's not too hard to make the services, especially things that they've moved into user case, like a driver kit and FS kit. It's not too hard to make them be able to, you know, reboot and replace themselves on the fly.

But then you have to validate, does every other part of the system that is above this handle that and not just flip out and be like, oh my God, I don't know what's happening. The thing I was talking to is gone and that was a new thing . And I don't know. And that's even before you get to like third party apps and stuff that are built on top of it.

So it is a much, much bigger problem than you think it is, which is why changes like this happen very, very slowly. You really have to get everybody gradually used to it. And they are traveling in that direction. And I bet they're going to get closer and closer to that goal, but the bureaucratic problems have to be fixed first because those impact are impacting us now, even for the easy fixes.

And those other changes are going to take a long time because you have to do them slowly and carefully. And it involves everybody in the entire ecosystem becoming acclimated to changes to things they thought would be constant and unchanging. And that's going to take a really long time. The other suggestion we got was from Kari Helgeson.

Kari writes, it occurred to me, listening to you guys talk about John Ternus' wish to squeeze more out of developers, that one potential interpretation is not to squeeze more from developers already paying, but from developers who get away without paying anything like Meta and Netflix. That primarily offer free apps through the App Store.

The new platform fees that are charged in the European Union are an example of how Apple could change requirements for larger developers that currently bypass paying Apple. That is an optimistic scenario. We may get to more hopeful and optimistic listeners who wrote in to us later in the show, depending on what we cover.

But trying to get more money from Netflix and Meta is probably really hard. And every time Apple has pushed in that direction, Netflix and Meta have said, yeah, no, we are backing out even farther from your platform. We will take the hit.

We will just not do it. And we will just continue to offer our free apps. Apple can do whatever they want, but they have to be careful because those companies have some leverage over Apple as well. Because if you buy an Apple device and you can't watch Netflix on it and can't use Facebook, that currently hurts Apple a lot.

We are sponsored this episode by Uncommon Goods. All right. Did anybody realize just recently this week that the holidays are like right around the corner? I actually have a reminder set for every early November to remind me of this because it always sneaks up on me. And I always start stressing out. Oh, geez, I have to shop for everybody.

What do I get? What do I get? Uncommon Goods makes holiday shopping genuinely fun with thousands of one of a kind gifts for everyone on your list all in one place. So if you have, you know, the friend who has everything or the in-law who's impossible to shop for, no matter what it is, you will find something that makes them say, wow, where 'd you find this?

Now, if you wait until December, they'll be fighting everyone else for the same gift. So get ahead of it. Uncommon Goods helps you do this. Uncommon Goods looks for products that are high quality, unique, and often handmade or made in the U.S. Many of them are crafted by independent artists and small businesses.

So it really helps out, first of all, artists and small businesses. And also, it makes your gifts feel meaningful and truly one of a kind. So when you're all out of ideas, you know, for whatever people need to get for gifts, Uncommon Goods is there with fresh gift ideas you probably haven't seen before. And they cover all sorts of interests.

Moms, dads, kids, teens, book lovers, history buffs, die hard football fans, foodies, mixologists, whatever your people are, whatever they like, you will find thousands of new gift ideas that you won't find anywhere else at Uncommon Goods. And with every purchase you make there, they give back $1 to a nonprofit partner of your choice.

They donate more than $3.1 million to date. So don't wait. Cross those names off your list now before the rush. To get 15% off your next gift, go to uncommongoods.com slash ATP. That's uncommongoods.com slash ATP for 15% off.

Uncommon Goods. We're all out of the ordinary. Thanks to Uncommon Goods for sponsoring our show. All right, let's do some topics. And let's start with Siri AI. And there was a news item in the Apple developer site about Siri AI.

And Apple writes, starting on November 3rd with the 27.3 betas, a new developer preview enables any developer to integrate their app with Siri AI. Today's preview allows any app to bring its content and actions directly into Siri AI by using App Intents. So people can just ask for help in their own words.

And Siri AI carries out any custom action you've built from checking a delivery to booking a reservation to placing an order. For complex requests, we're also introducing skills to let you teach Siri AI how to combine several actions into a single experience. And Siri AI can now surface any entity in the Spotlight Sem antic Index, not just the ones that fit an existing schema.

So I've been very busy with like life stuff. Everything's good. I've been busy with life stuff. And so I haven't had a chance to dig into this really at all. But I don't think I understand at least the technical side of this. Now, maybe we're not going to get into that this episode.

But I am deeply unclear what I am supposed to do with Call Sheet in order to support this and how it works and so on. So, John, if we don't want to talk about that today, that's totally fine. But tell me, what do we think about this? I can tell you the high level overview of it.

Up to this point, for your app to integrate with Siri AI or Siri at all, it had to basically conform to a certain preset list of types of actions. And it would be like notes, ride sharing, reminders, you know, it's like stuff like that.

And so that way, if your app was somewhere close to that, you could say, all right, my app, like, you know, for Unfor gettable, my app complies with the, you know, the create task API for Siri. And I implement a few endpoints and it says, okay, this is how you create a task.

This is how you update a task. This is how you delete a task. And Siri kind of presupposes or it predefines like, okay, well, what a task means is that it has to have a title, a description, optional URL, optional image, like that kind of stuff. And it presets all of these things in like a big definition .

And your app can say, all right, I comply with that or I don't. And the only way to integrate with Siri was to comply with one of those preset roles that they had defined. What this means is that they are basically allowing an app to just basically describe itself and in a way that the AI will read and will try to understand.

So your app can expose whatever kind of API it wants with all sorts of different capabilities. And it just describes to the system what these capabilities are. And then the on-device and probably cloud LLM will try to interface with that and try to figure out, okay, so the user just asked, like, look up a movie in CallSheet.

CallSheet has an API that says, I can list movies for users and get information about them and their cast and things like that. And so CallSheet describes to the system in like English, you know, what it is, like, you know, what it offers. And then the system says, hey, this person just asked for this request.

It sounds like something CallSheet can answer and that maybe they want CallSheet to answer it. Let me try to figure out how to get what they said to match this API over here that CallSheet says it advertises. So this is the promise of Siri AI. This is the whole thing. And Apple Intelligence, like, this is part of what was advertised, you know, whatever it was, two years ago and, you know, didn't ship until now.

Like, this is a core part of that. It's a big deal. I, um, based on how Siri AI is going so far, I'm a little hesitant to get excited about it, to be honest. Also, this is like the fourth time they've tried this.

And so, again, I'm a little hesitant to get excited about it, but they do occasionally make progress in this area. So maybe this is the time that this will actually work. But until they really show us that, I'm not going to believe it. So do you plan on doing any of this for any of your apps?

I think it, I mean, Unforgetful, I don't really need to. Unforgetful already complies closely enough to the reminder role that just continuing to comply with that is probably the best move there. Sure. Unforgetful, I think, overcast complies with the, like, the play media role and has for many, many years.

And again, I, and I think overcast basically, like, that's basically all it really needs. Like, I can expose a little more functionality that's kind of overcast specific. But I don't think, I don't think I'm heading for, like, major changes there. It's going to be more, like, minor additions.

And Fineliner is basically a notes app, but not quite. Like, and I already expose the notes API to Fineliner, at least for reading, for spotlight indexing. And so that you can ask Siri, in theory, hey, what did I say?

You know, what did I write in my grocery note in Fineliner? And it should be able to tell you, although, honestly, I've never gotten it to work. But, so, who knows? I have some investigating to do there, I'm sure. So, already, my apps are kind of close enough to the existing stock endpoints that already were there.

So, I don't foresee major changes for my apps. But, again, minor changes, probably. Like, you know, adding little things to the API to Siri that, like, maybe Siri's template didn't have that my app has. So, things like that. I got you. The interesting part of this is the guts of it and the implementation of people who haven't been messing with, like, I mean, coding agents is what I've been messing with.

But I'm sure it's true of other type of things. And you've used these things called skills. Like, that's a term of art. You know, there's plug-ins, skills, whatever, blah, blah, blah. But skills has become a term in the industry for things that you can install that will teach your agent how to do something that it didn't previously know how to do.

And it sounds really cool and sophisticated. But you know what skills are? Text files. With English in them. Because they're large language models. And if you open up these text files, they have English that you can read. For example, we said this on a past episode.

Like, the Xcode 27 or whatever version they shipped this in . Came with a bunch of skills. To. About, like, Apple computery platform stuff. And you can export them. You can do XC run something something export. And it will spat out a folder with a bunch of text files in it.

And you can read them. And they'll be, like, modernizing your tests. And it will be a page and a half of concise English saying, here's what you do to modernize your tests. If you're using XC tests, you know, Swift testing is the new thing. Here's how Swift testing works. Here's best practices. Here are things to avoid. There's a SwiftUI one.

Modern SwiftUI. If you're using old SwiftUI, you might want to think about this, that, and the other thing. Here are the, like, it's like a miniature, very concise, like, lesson, you know, that you can imagine being a series of slides or a transcript of a WWDC session. Explaining in English concepts.

These are built into Xcode. So, when you use Xcode's coding agents, it will read these text files into the giant context window along with all the other things that it puts in. And that advice will influence what it does. You can find third-party skills from a bunch of people. Like, Paul Hudson has a bunch of them. He's hacking with Swift.

He's got a Git repo. He's made a bunch of skills. Go look at them. They're just text files in his repo that have two pages of text from Paul about things to avoid and things to do with Swift, with SwiftUI, with Swift testing, with whatever, whatever. That's what a skill is.

And so, this, not only are they breaking you out of, oh, you have to say that you're a notes app. You have to say that you're this app. And then you define your intents or whatever. This is saying, define whatever the hell app intents you want. And then provide some text files that describes, hopefully in a concise way, because you don't want to blow your context window.

And there's actually fairly strict limits on a lot of coding agent harnesses or whatever. Like, don't allow the skills to be too long, you know, because otherwise you'll just fill the context window with your skill that's never being used. Anyway. And you can enable and disable skills, which is a good idea to narrow things down. But just describe how your app works.

So, for example, you'll say something like, you know, pretend that you had a reminders app, but it's not like, for example, you say, my app organizes things into buckets. And the buckets are named after the seasons of the year, winter, spring, summer, and fall. So, if the user asks to do something, first find out what the current season is from the calendar.

And then based on what the current season is, look in that season bucket and whatever thing they said, look in there first. But if you don't find it there, look in one of the other season buckets. You can just write that. And then your hope is, fingers crossed, like with everything, your hope is. Then when they say, you know, hey, my cool reminders app, add this to my whatever list.

It will look at those. It will have that instruction in the context window and say , oh, I'm not just going to look for that whatever thing. I have to first find out what the season is with a calendar API that I hope I know how to do. And then I know which bucket to look into. Like you're describing the user model, the information structure of your app. Instead of just saying, I'm a reminders app and I have reminders.

Good luck. Here's a bunch of APIs. You tell it like how to do this stuff. You know, these are the steps you should take. And it will look at your app intents and look at that skill and be like, oh, I'll call this one to do a list. Like you can see it if you're ever using like the Xcode MCP and you see what it's doing. It's like I'm listing all your workspaces.

I'm listing all your open tabs. I'm seeing what the current schema is. And like all that stuff is informed by a skill that is telling it when, you know, when the user asks for whatever, here's how my app works. And that is incredibly freeing. And it is also just throwing a dart at a dartboard and being like, God, I hope I'm what I'm writing here has some beneficial effect to Marco's point.

Like, yeah, you can write whatever you want in that text file. Is that going to result in a user saying a thing and getting useful response? Or is it just more more words thrown into the context window and you're just hoping something nice will happen? And on that front, based on my use of skills for a bunch of stuff, like you'll see skills for everything, by the way.

Like, for example, when working on the payment back end of my hopefully soon to be released app, there are Stripe skills for the Stripe payment processor. You can, the Stripe itself distributes skills that says, hey, load these into your coding agent if you're trying to debug some, some Stripe problem.

And it will tell it, here's how, here's, here's some common pitfalls for using Stripe. Here's the thing to look for. If you're checking a configuration to make sure it's right, look at this, this and that, right? Incredibly useful just to read as a human, not forget the coding agent. But also you're hoping the coding agent will get something from that and make itself more powerful.

Now, Siri AI is slow, even on big, powerful computers. And it's frustrating and it is not as good as the top tier, you know, chat GPT or Claude, current and former sponsor. It's just not, you know, and, you know, it's free or free- ish and it comes with your Mac.

So it does have advantages, but it's sometimes frustrating ly slow. But the thing that makes me hopeful about, yeah. The thing that makes me hopeful about incorporation of skills and opening up third-party apps to Siri AI is that I 'm much more willing to tolerate slowness if it's doing something sophisticated and impressive.

I would much rather be able to ramble off a big, long thing and have it read in all the skills for these three apps that I'm using and actually figure out how to use those specific apps the way the authors of those apps intend them to be used so that it knows to check the current season.

And then check the season bucket and then call the API that searches by task name instead of just doing a task name search across the entire thing and picking the wrong one because it doesn't know how my app works. I will wait for that. I won't wait for, you know, remind me in 10 minutes to change the laundry.

That should be fast. Okay. But the more sophisticated things get, the more I'm willing to tolerate waiting. So here's hoping that this change finally gives Siri AI something that plays to its strengths, something that is acceptable when it's slow.

Because the things, the simple things they can do now are not acceptable when they're this slow. A more complicated thing that takes even longer, I would find more acceptable because in the end, you just want it to do the thing. And once you can trust it to do this because the apps that you're talking to asking it to use or whatever have skills that teach it how to do it, then that makes you more confident.

It doesn't help the set a reminder for 10 minutes case because Apple needs to just fix that. Oh my God. It's so bad. Like I can't even set timers anymore. Like timers. Set a timer for 10 minutes. Even that, it'll just do. You see the spinner for a long time, don't you? It just spins for.

And the other day, I just couldn't set one. It just, it just didn't. It never worked. I mean, maybe if I would have sat there for another 20 minutes, but it was a 10 minute timer. Like it just doesn't work. I'm, I am so anti Siri AI right now because it has broken so many things for me that used to work a lot better than this.

That's why a lot of people are downgrading to classic Siri. I forget how you do it, but there's like ways you can say, I don't want to use the Siri. I want to use the old one. My stance on that is I had enough of the old one, man. Like I know it had strengths and weaknesses and I know it was faster to set timers, but like there was no hope of the old one ever doing anything sophisticated.

And for some people, it's worth it to them. They're like, I've just wanted to do the simple thing fast. Like I get that. But I am, I am, you know, I'm, I'm going forward. Like I just, I've had enough time with the old Siri. It's so frustrating. I want the possibility of it doing something smart. And I'm currently willing to pay the price of having to wait 10, 15 seconds to set a timer.

Because in the end, for me, it has still eventually set it. It's just embarrassingly slow. Apple needs to fix this. Let me ask something. Is there anything that Siri AI does reliably for you? I mean, it does. When I do ask it to set timers and reminders, it, other than it coming back to me and asking me to clarify pronunciation when I'm in a noisy environment, which I find frustrating.

But other than that, it has never failed in the way you described. Like it always does eventually set the timer. It does eventually set the reminder. When I clarify the pronunciation, it does accept my clarification. Although, again, the, the alternatives it gives me is like, which one of those do you think is plausible?

One of them has the name of my child in it. And the other one is just gobbledygook. So anyway. The amount of words in your response makes my point. When I asked a very simple question and you just had to give like three paragraphs of qualifications to your answer . Well, I, you know, it's like it can do more things that the old one couldn't do.

That's not what I asked. But it still does the things that I want it to do eventually. So I'm, I'm willing to tolerate the, I'm willing to tolerate the performance decrease for the potential of it doing things that are more sophisticated. I see. I don't want to harp on this, but just so far, Siri, Siri AI for me is the same playbook, a huge amount of promise, some impressive things it can do, but so much unreliability and slow performance that I just never want to do it.

I never want to ask because it lets me down so often. I mean, I'm still, I'm still asking it to do the same things that I ask it to tell me when to change laundry. I ask it to remind me to do a thing when I, when it occurs to me, when I'm on a walk, I, I'm using it the same amount I always was. It's just that every request now takes longer.

Great. What, what an upgrade. Yeah. But all I'm saying is it's not about, to my point is it's not, it's not actually deterring me. It is annoying, but it is not actually deterring me. It deters a lot of people though, again, which is why they downgrade to classic Siri, which again, totally makes sense . But for me, I'm sticking it up.

So Marco had made mention an hour ago about all the different domains and things. And there's a post that we're going to link to where, uh, swiftjective C.com. They talk about, uh, Apple's finally freeing us from domains. We'll link that in the show notes. Oh, I'm sorry.

This is Jordan Morgan. Uh, before 27.3, unless your app fit into a blessed domain, there weren't many opportunities that we could take advantage of in our apps. Why? It's simple. Apps do not fit into pretty little buckets. Uh, now Apple is finally freeing us from domains. You can supply your arbitrary app intent definitions to the system and even supply skill files and markdown, like John was saying, to help Siri know how and when to use them.

All right. Uh, there was a link that I didn't mess with until earlier today and was both very cool and mildly alarming, but mostly cool. If you ask me, uh, Dario Pasquini and Mike Michelle Baisley . I hope I got that somewhere near right.

Uh, at eighth rank, right? What if we told you that every Mac ships a pair of demons that whose only purpose in life is to build a graph about you and the people around you, who you are, whom you know, how those people are related to you and to each other, how you communicate with them, where you go, with whom, what do you do and when this graph is built on device by cross-referencing pretty much everything on the machine, context, calendar, photos, messages, mail, location, screen time, and a long tail of behavioral signals.

And more importantly, it's held in a queryable store that L LMs retrieve from allegedly, or that LLMs allegedly retrieve from, I should say. Would you believe us? Well, that's what's happening. And interestingly enough, there's a little, there is a little documentation of how this whole mechanism works on your Mac. Well, TLDR, they built or they vibe coded a front end for this that you can run on your machine and it only runs, you know, privately and, and you can try it, which I did.

And let me tell you, there was an astonishing amount of information on there, including like who my spouse is, who my parents are, who my children are and stuff like that. Now, I don't think I'm particularly worked up about this. Like this is how this is supposed to work. I'm a little worked up that it's just sitting there and easy to read as long as you have full disk access.

Well, easy-ish to read. But by and large, I am not currently worked up about this. Ask me again at the end of the segment. Maybe I will be. But I mean, this, this is, how do you, how do I expect Siri to be helpful if it's not going to try to figure all this out for itself? Yeah, this, this article went around and it's phrased a little bit provocatively in the beginning.

Like, what if I told you that your computer is spying on you? And for people who aren't, you know, ATP listeners and keeping up with every nuance of the industry, it is probably is surprising for them to know, what do you mean? My computer is compiling a database of information about me . I didn't ask it to do that.

That seems invasive or whatever. But if you have been keeping up with Apple's announcements, you know, especially the 2004 24 one that never came to pass, you know, that Apple has talked about this again and again by various names, a semantic index, blah, blah, blah. Like the things they're demonstrating, you know, when his mom's flight arriving, how can it know that if it doesn't know who your mom is?

Like, like it's not, I'm not saying that people are wrong to feel that this is invasive, but it is also wrong to say that Apple is trying to do this sneakily behind your back. Like they're literally advertising that this is what their thing's going to do. In fact, they're advertising things that they hope it will do that it can't currently do, but they're trying.

And the, this is basically like, how is that implemented under the covers? Well, of course it's implemented, you know, in a way where they build, you know, whatever the semantic index or whatever. I think semantic index means something slightly different than this. But anyway, in the end, they have to build a data store that knows this because they're not going to derive it in real, even though it seems like they are, they're not going to derive it in real time when you ask your question.

They're not like, Oh, I better start searching their contacts to find out who their mom is. No, it's going to do that in the background ahead of time and build it in a concise way so that when you ask a question, it can in theory respond without taking a half an hour to dig all this information out of your messages history, which that's the laboriously download from iCloud or whatever.

Um, but it makes for a good sensational story of like, I cannot believe that Apple is building up this dossier on me. Uh, and to your point, Casey, if it is unencrypted at rest on the disc and exploitable and grabbable, that is a way for people to steal information about you.

But on, you know, on the flip side, if they have that kind of access to your computer, they can just take your actual context database and your actual email. Like they don't need the index. They can just get the raw data. But either way, like I can understand why people are upset about it. But when I read this, I'm like, Apple, whatever you're doing, do it harder because it's still not good enough.

Like, but that's me. Like I want to be able to ask it when his mom's flight are riding and have some faith that it's going to get the correct answer in a reasonable amount of time. And I currently don't have that faith, but maybe that's not a thing people want their computer to be able to do. Like, I feel like this is kind of a thing where it's like, it's all just like a threat and evil and everything until it passes some threshold of reliability and usability.

Then everybody's like, oh, well, that's too usable now. Now I accept like the, uh, that it's building this index of my relationships because I didn't want it to build a relationship when this was not benefiting me in any way. I didn't ask for this. I don't want it. Turn it off. Get rid of it. But if it ever crosses that threshold where you're like, oh , actually I did want to know when mom's flight arrived and it told me and it was right.

And now I trust it. Then suddenly you may be willing to accept the cost of having this big index being built. And of course the complaint is I was never asked about this . This is not a thing that I did with my consent or whatever. And I get that perspective. But again, it's not like Apple has secretly been doing this and not telling you they've been advertising.

They were advertising it before they were even able to do it. Famously, they had a TV ad, but a thing that they could not do, but they were telling you we want to be able to do this and we're trying. We just suck at it. So anyway, it's cool. There's like, you know, things to visualize this database and take a look at it and stuff.

But most people I know who have tried it have said, yeah, it figured out who my wife and kids are and it knows where I live. And it's like, you know, anything thing anybody can get by looking at your contacts, calendar, photos, messages, mail. Like, yeah, you can get that information by looking at that stuff. By the way, Google knows all this about YouTube as they read all your Gmail.

Right. Yeah. Yeah. Yeah. I mean, it is worth if you're interested anyway, it is worth playing with this app that they built and it's open source. Again, vibe coded, but open source. It was it. I don't know if I'd go so far as to say fascinating, but it was definitely interesting.

But yeah, it's worth checking out. Yeah. And the other part of this, of course, is like who has access to the information and to what purpose? Apple's whole pitch is this information is private to you. We can't see it. We don't store it. You don't send it to us when you send it to our servers.

They're private cloud compute. Those servers can't store anything about it. We never see it. We can't peek into the servers and watch the data you're sending. We can't do any of that. Like this is your data. Yes, it is on your Mac and it's a potentially a security problem if someone breaks into your Mac and steals it and so on and so forth. But Apple does not want this information and carefully constructs its system so that it can't even get it.

Even if someone asks for it, they can't get at it. Google, on the other hand, has had complete access to the plain text of your email and Gmail for the history of Gmail. And Google does use that information for its business, for advertising, targeting and figuring out stuff like that or whatever.

So, again, I'm not saying this should make you feel better about Apple. Like if you don't like what they're doing, you don't like what they're doing. It is what it is. But if you look across the industry, it is, I feel like, scarier to give for companies who want access to this information.

Facebook wants to know this about you. Facebook wants this information because they use it for targeted advertising. They want this information. They don't want anyone else to have it. They want to sell ads against it because they would say, hey, we know all the people who are currently in the market for a new lawnmower. We're not going to tell you and we're not going to sell you that information, but we'll sell you an ad that you can send to people who are currently in the market for a lawnmower who live in this state, who are this old, who are blah, blah, blah, blah, blah.

That's their business. So they need that information and Google, same thing, search results and stuff like that, that they need that information to power their business. Currently, for the most part, Apple doesn't want and doesn 't need this information. Famously, Apple absorbs tons of information about what you 're doing on the app store because they sell search ads there.

But that information is like, what apps did you look at? What, you know, what things did you click on? What apps did you buy? So stuff you do in the app store, it's kind of like being in Google land where Apple is absorbing all that information. But stuff like your contact information and where you live and who you're related to and who your kids are.

Currently, none of Apple's business models require that information as input. And so when Apple says that we keep this private and we can 't see it if we wanted to, I believe them because it is not useful for them and in fact is a liability to them. This is as compared to the app store activity, which is useful to them and they sell ads against.

So that's the way I measure things. Like, is there a reason the company would want this information? And that influences whether I believe that they really are never going to touch it. We are sponsored this week by Quince. So the seasons are changing.

In fact, we're about to get a cold front here in Virginia. And now is the perfect time to take stock of your closet and decide, maybe I could use some new pieces for this or for that or whatever the case may be. But it's more than just your closet, right? Maybe you want to redecorate the house a little bit. Maybe you need some different sheets or different pillows or whatever the case may be.

Well, let me tell you, Quince is your kind of one-stop shop for basically all of those things. What I really love about Quince is obviously they have a ton of men's clothing, but they also have clothing for women, clothing for kids, and housewares and travel stuff. There really is an astonishing amount of stuff on Quince.

So if you've heard this ad before, I would say give them another look because you'd be surprised what's there. They're always adding new stuff. What Quince does is they focus on high-quality staples made with premium materials, like 100% Mongolian cashmere, 14-karat gold, and Italian leather. These are timeless and versatile pieces that feel special to unwrap and become everyday favorites long after the holidays.

And what's great about Quince is that everything is priced 50% to 80% less than similar brands. They work directly with ethical factories and cut out the middle people so you can give luxury-quality pieces or purchase for yourself luxury-quality pieces without the luxury price tag. They were kind enough to give us a little money to spend, and we bought, among many other things, for Michaela, an organic cotton long-sleeve fit-and-flare pocket dress.

This is the perfect fall colors. It's like a coppery color. The official name is Caramel Confetti Floral, and it has pockets, which is great. And so she loves this thing, and it is perfect for this time of year. And as I'm recording right now, it's under $25, which ain't bad.

So, hey, if you want to give the gift of quality this holiday season, download the Quince app for app-exclusive offers or go to quince.com, that's Q-U-I-N-C-E.com. You can get free shipping on your order and 365-day returns . Quince is now available in Canada and the U.K., too.

And when Quince asks where you heard about them, you should let them know. It was from the Accidental Tech Podcast. It's the best way to support the show. And speaking of supporting the show, thank you to Quince for supporting ATP. All right, let's do this thing that we almost forgot about called SKTP, and we're going to start.

Don't worry. Upgrade will periodically remind us. Yeah, right. We're going to start with Dennis Berger, who writes, I'm genuinely curious why you guys use dedicated Apple TV devices instead of the built-in Apple TV app on your smart TVs. Is there a noticeable difference in picture quality, or do you use your Apple TV for things other than watching Apple TV Plus content?

For me, I'm trying to be not-so-sweary. I have a LG C9, which I love as a television, but all of the built-in stuff is trash. It's slow. It's not very user-friendly. It's garbage. Or at least it was five, six, seven years ago, whenever it was I bought this TV.

And maybe it's gotten better since then, but thanks, but no thanks. I do a lot with my Apple TV, not just watch Apple TV Plus content. Plex is on there, although I'm mostly not using that for myself. I have several different jellyfin front ends that are on there, which is what I am using for myself.

I forgot to put the show notes. We're going to have to talk about it at some point. The death day for my cable card has come and passed, and so my new alternative for television is on there, but before that, channels was on there. Like, the Apple TV was literally everything that I watch on TV.

It's not just Apple TV Plus. And let me tell you, the Apple TV, while hilariously expensive and kind of ridiculously overpriced, if you ask me, is far and away, not even in the same universe, better than everything else I've personally tried. And I've tried a handful of these things.

Not as much as Snell has, but I've tried a handful of these things, and it is way better than anything else. I cannot fathom hating myself enough to use the stuff built into my TV. Thank you, but no thank you. Yeah, I think it's less about picture quality for the most part, and much more about the overall user experience.

Now, tvOS is fairly half-butted by Apple, but even when Apple is only using half their butt, it is still so far, so many miles ahead of what everyone else produces in this area.

It is really quite staggering when you are accustomed to Apple's, you know, Apple TV experience, and you go try to use smart TV software from anybody else. Now, if that's the only world that you know, it's similar to, like, if you only ever use Windows PCs or Android phones.

Like, if it's all you know, you might think it's fine. And for many people who know more, they might think it's fine. But when you are accustomed to Apple's way of doing things and Apple's standards for doing things, everything else is full of ridiculous paper cuts.

It's also just incredibly user hostile. Like, we, somebody in my household recently connected our LG TV to the Wi-Fi. This did not happen on my watch.

This would never have happened on my watch. As a result of this, the LG TV is now stopping, is now showing in the middle of us watching TV, fully bright white dialogue boxes randomly advertising upgrades to us.

Like, I'm so unbelievably annoyed by this. And just, yeah, this is, this is the standards of what the software does. Apple, they used to promote baseball games to me. They don't do that anymore. But, like, Apple does not put a bright, fully, you know, full white overlay over your dark TV show in the middle of you watching it, telling you, hey, you should really upgrade your software for the Apple TV right now.

It's like, no, they don't do that. They just quietly take care of things in the background. It's just one of many examples. And again, tvOS is not perfect. And it's, it's really not a high effort platform for Apple. Um, and the new Apple TV 4K is horrendously priced at $300.

Uh, so there's a lot of reasons. Well, there's one big reason not to get the Apple TV, which is the price. But if what you want is a nice experience, then it is unpar alleled. It is the only option in the game.

Nothing else in this market provides a nicer experience or even a close to as nice experience as the Apple TV. Yeah, as Casey, uh, touched on something that I think is, uh, a key to, uh, the general advice to, if you, if you can stomach the price to get an Apple TV box for your television, people don't replace their TVs that often.

They don't need to, uh, you can buy a TV and use it for many, many years. Casey's C9, I believe that nine is for 2019. Yep. I believe that's right. That was the year numbering model scheme that they have. Um, yeah. And as your TV ages, the tiny little chip and Ram and whatever, and storage in there that is running the smart TV OS does not get better with time.

It just becomes more and more outdated, more and more wimpy compared to what modern technology can provide. Um, Apple TV boxes for all their expense, the day they're released are tremendously more powerful than the chips that are inside even modern like TVs that are contemporary with them.

And as times go on, the Apple TV doesn't get updated that often, but it maintains that lead. I'm not sure there has ever been an Apple TV that was sold long enough to see the chips in televisions be more perform ant than it is. There's, there's plenty of fast chips in, in, in telev isions, but they are running like the image processing for the TV, right?

They're not running the smart TV apps. Those always get scant resources compared to what an Apple TV can do. So Apple TV is massively overpowered for what it's asked to do. So the performance is usually pretty good. Um, one of the things I will say about built-in smart TV apps is there have been times in the past and probably somewhere, somewhere today where a built-in app will have better picture quality for certain sources of content than the Apple TV app.

Is simply because the built-in app is made in cooperation with the people who make the TV and has access to TV controls or is willing to use TV controls internals of the TV that Apple TV can't. Because Apple TV is a generic device you would connect it to any TV, whereas the built-in apps can really get, they, they know what TV they're inside.

So it's at various points, there have been built-in apps that are like, if you want to see this with the best picture quality, um, with the highest bit rate, use the built-in whatever app, because it is better than the other one. But that is a rare exception. Almost all the time, the truth is that the built-in apps are slower and also uglier and have worse picture quality and fewer features than the Apple TV one.

Which again, makes sense considering the smart TV apps are running on $1.50's worth of the hardware that's inside there. Whereas for $300, you're getting at least $100 worth of hardware, probably. All right, maybe $50 worth of hardware. But still, it's an order of magnitude more hardware being thrown at the problem.

Now, I have huge numbers of complaints about Apple TV. It is the worst except for all the other ones. So that is the frustration with Apple TV. Like again, I said when, you know, pausing during the previously on, it didn't allow me to unpause. Plus, that is an unacceptable level of crappy functionality .

But its competitor products, even if they don't have that reliability problem, have worse things. So their business models are based on throwing advertisements in your face and crap like that. Or they have slow hardware or some combination. Whereas thus far, Apple TV's business model is mostly focused on getting you to subscribe to Apple TV Plus and getting you to buy the $300 box.

And there is not as much, you know, hurting you towards viewing ads or whatever the other cheaper things do. Including, as Marco pointed out, the built-in apps on your TV. And this is before we even get into LG apps spying on you and Samsung TVs listening in and all that other stuff.

So, you know, Apple, again, Apple TV has a better privacy story there because they, you know, they say they don't do that. And we believe them because no part of their business model , you know, would be benefit from that. Whereas people selling low margin televisions, some companies like Vizio or whatever, if they're still in business, I forget.

But a lot of the lower cost television products, they sell their TVs at a loss and make all of their money by spying on you. So it's not a question of like, oh, is this beneficial to their business model? It's their entire business model and, you know, and showing you ads and stuff like that. So, yeah, there's a reason everybody uses Apple TV boxes, even though we complain about them incessantly because everything else is sadly worse.

It doesn't mean Apple TV is good, but it just means everything else is so much worse. I could not agree more. All right. Richard writes, and I actually dug up this email because I replied to it when it was sent in the 5th of August. So if you're curious, do we get to ask ATP stuff in a timely fashion?

No. That's why we respond by email, by the way, because we know like, oh, this might be on Ask ATP, but the poor person needs to hear now because they have a problem now. Which, to be clear, we are not offering tech support. Don't hear what John did not say. We are not a tech support solution. But for what it's worth, we do try to answer if we have something worth saying.

But anyways, Richard wrote on the 5th of August, Casey often espouses the benefits of using his Synology while away from home via Tailscale. While I have port forwarding configured on my home router in order to remotely access my Synology DS423+, I don't have a Tailscale VPN set up for authentication and access, nor do I have HTTPS set up for accessing my various Docker images.

What does your Synology VPN topology look like? How did you set it up for authentication? And are there any memorable challenges you encountered along the way? All right. So let me just start by saying Tailscale is a former sponsor. I think they are a future sponsor as well. I freaking love Tailscale.

And if this sounds like an ad read, truly from the bottom of my heart, I'm not trying to make it that way. It's just I really, truly like this software that much. So if you want to skip this chapter, I totally get it. But I think I have some interesting things to say. Now, as a very brief overview, Tailscale uses VPN technology.

It uses WireGuard under the hood, but it's so much more than that. And what Tailscale is, is if you think about a typical VPN, it's generally speaking hub and spoke. So you are dialing in, so to speak, to a server. And then that server will get you onto a network where you can ping away at other servers, right?

So you're dialing into like your employer. Your employer has one or more servers in a particular location. And once you are dialed into that server, I shouldn't say dial in, but you know what I mean. Once you're connected to that server, then you can connect to other things behind your corporate firewall. So Tailscale, on the surface, uses the same technology, but it's a mesh network.

So the idea with Tailscale is as many of your devices as you can will get onto your quote-unquote tail net, which is to say it will connect to Tailscale. But you're not really, when I say connect to Tailscale, you 're really just advertising, hey, I'm here if you need me.

And all of your other devices say, oh, I see where, let's say, the Synology is. And what Tailscale lets you do is it's a mesh network. So any of your devices, if they need to access your Syn ology, they directly access your Synology. You're not going to some Tailscale server and then bouncing back onto the Synology.

You're going directly from your node, from, say, your Mac, to your Synology. Now, there's asterisks and double daggers and daggers and double asterisks and so on, but I'm trying to keep this less than an hour. What this means, though, is that effectively, it's as though any device on your Tailnet is on your local network.

So all of your Tailscale devices get 100.xyz IP addresses. And as long as you're connected to Tailnet, you're, and I'm making this up, you're 100.1.2.3. That will always be your Synology. Always.

I don't care if you're in your house. If you're at Marco's house, don't be creepy. If you're overseas, wherever you are, 100.1.2.3 is always your Synology, no matter what. And your Synology can be on a different continent, different network. It does not matter.

And it's that that makes Tailscale incredible. Tailscale also has really trick ways to, like, punch holes into NAT situations. If you don't even know what that is, it's probably for the best. But basically, if your other device is behind a firewall, Tailscale doesn't always, isn't always able to poke through and get a direct connection between your two devices.

But it's, like, damn near always. And if it can't do it directly, Tailscale, here's one of those asterisks, Tailscale actually does run some servers where it will be an intermediary. So you connect to it, and your other device connects to it, and it just is a pass-through if it can't make a direct connection.

This happens very rarely. Everything that does get transmitted is fully encrypted, so they have no clue what you're sending. But it is an option that they fall back upon. So with all that said, what do I use it for? Well, basically, all my devices are on my Tailnet, so I can always connect to anything.

It doesn't matter. My instance at Linode is on my Tailnet. That's in the D.C. or – no, it's Atlanta, maybe. I forget where geographically it is. But it's on my Tailnet. So it's effectively on my local network. Not literally, but effectively. I have a droplet at DigitalOcean for the kids' school.

That's on my Tailnet. It's effectively local. I have a Synology at my parents' house. That's on my Tailnet. It's effectively local. And you can see where this is going. My KVMs that Marco and I have talked about are on my Tail net. My NUC, my NUC that runs all my Docker containers is on my Tailnet.

And so on and so on. So what does Tailscale provide more than just connecting your devices? Well, one of the really great things it does is if you're a Docker sicko like I am, and you have all of these different containers, well, typically you would access them by like 192.168.1.7 with a port number on the end.

And that's not SSL secure. It's HTTP, not HTTPS. Well, you can use one of the features of Tailscale called Tailscale Services, and it will proxy – well, it's complicated, but the short short of it is it'll proxy those, and it'll terminate an SSL endpoint for you.

So you're going to, instead of 192.168.1.2 colon 12345, you're going to, I don't know, like jelly.sumname.ts.net, and that could be your Jellyfin server.

And that's HTTPS jelly.sumname.ts.net. And now you have SSL termination for all of your different Docker containers and stuff like that. So it's things like that that make it really cool. They also provide an OIDC provider, which sounds like a corporate thing, but with a lot of these projects that I'm running on my Docker server, like Home Assistant, like Paperless NGX, and among others,

you can actually configure them to use OIDC as their authentication setup. And then what happens is when you log into Paper, is what I use, paper.sumname.ts.net. That's my Paperless NGX, which is like a PDF repository. If you log into Paper, or if you want to log into your Paperless NGX instance, you go to that URL, which is SSL terminated.

You don't have to fiddle with port numbers or anything like that. And you hit log in with Tailscale, and then you're logged in. It's awesome. The big thing, though, and what I said to Richard when I initially replied to his email, and I said, this is verbatim, I mean this with respect. If you're opening ports, you've already failed.

That's kind of the number one benefit of Tailscale, is you don't have to open ports. So, when I back up, when my Synology backs itself up to my parents, or the Synology that's mine, but at my parents, it's doing that via Tailscale. They don't have any ports open on their router.

My Synology, actually neither of my Synologies, are accessible via the internet, but they are accessible via Tailscale. And so, you never have to worry about port forwarding. Because why? Because, effectively speaking, all of your devices are in your local network.

I know that's not literally true, but effectively true. So, that's what makes it powerful. It's the SSL termination. It's not dealing with port forwarding. It's that they're always available, as long as your two devices are connected to the Tailnet. It's really, really incredible. And again, now I'm definitely sounding like an ad read, and I am sorry for that.

But it is some of my favorite software I've ever used in my entire 30-whatever years of using computers. If you have multiple computers, or multiple devices, I should say, that need to talk to each other, it is truly amazing. Now, I would like to hear John's input if you have any, but Marco, you have a lot more experience than John with Tail scale.

Is there anything you'd like to add? No, you basically covered it. This is exactly the kind of thing that I would hear about for years, never really try, and just think, I probably don 't need that. Correct.

And then once I tried, I was like, oh, this is useful. Yes, this was exactly my experience, too. Exactly my experience. And this is why, again, I think what you said a minute ago, if you're opening ports, you're doing it wrong. If you're having to deal with dynamic DNS, you're doing it wrong.

Oh, well put. These are problems that we had to deal with in the past that we no longer have to deal with. Even simple things like if you want remote access to, say, a Mac mini that you're running somewhere or 47 more of them , it's much more secure and much easier for you to have a tail net going than for you to give them all public IPs or route them over public IPs somehow.

Because then when something comes around like that remote desktop exploit, anything on a public IP or on a public port might be exploitable. Whereas if it's all within your tail net, you have a lot more protection. So there's all sorts of reasons to do it. But mostly it just makes things work a lot better with a lot less fuss from you.

Yeah, and actually one other thing I'd like to add, and I think you might have said a second ago, helping out parents or whatever. Or maybe I just made this up. But anyways, my parents, they're getting older. And while both of them are fairly technically savvy, they still have questions that still have problems.

And I have this other Synology at their house. And Tailscale has this feature called subnet routing. And what that basically means is you can use a Tailscale node to be kind of the hub and spoke or the hub in the hub and spoke model of a traditional VPN.

And what that means is I have my Synology at my parents' house allowing me to access any of the devices on their network. And they also have a separate but related feature called something like IPv4 over v6 or something like that, where basically, and it's really complicated, and I'll put a link to where I wrote this up in the show notes.

But basically, if I use a particular host name, that Tail scale will then translate internally to mean, oh, this is actually this funky IPv6 address. Oh, actually, this funky IPv6 address means he would like to get to such and such IPv4 address, but on the other network.

So, again, I would have to explain for like another 20 minutes to get it verbally. You'll just have to read this post that I'll put in the show notes. But what this means is I can log into my parents' my dad's Mac, his MacBook Pro, as long as I'll tell him before I do it, of course. But I can say, hey, I'm going to log into your computer.

And I can screen share into his computer from my house. And his computer has never heard of Tailscale before. It's really incredible, incredible stuff. There's so many different things you can do with it. It really, I cannot agree with Marco more. I'd heard about Tailscale for years.

All my nerd friends were like, oh, you got to try, got to try, got to try. Yeah, yeah, yeah. What does it really do for me? I don't need any of this. And then I finally saw the matrix, and now I cannot live without it. John, any thoughts? Are you good? So on the topic on Richard's question, like, I know Syn ology has a bunch of VPN stuff that ships with their software or whatever.

Is your advice basically don't use any of the Synology built-in stuff and just do everything with Tailscale? I mean, that is what I do. I'm not trying to, like, throw shade at Synology stuff. And you could make an argument that maybe you'd want to stick with that for whatever reason. But for me, Tailscale is universal.

And I bought into a lot of Synology's ecosystem. But I don't want to have to use Synology's ecosystem for, like, VPN-related purposes. And again, there's more to my network and more to my devices than just my Synology. All right, Chris Kerner writes, I'm going to sound incredibly idealistic here.

But with the combination of insane price increases on RAM and other hardware combined with the efficiency and bug squishery of agentic coding, will we see a golden age of software over the next few years? Listening to you all talk about how these new tools have allowed you to be more productive has me thinking, will we see fewer bugs, more features, better efficiency?

From indie apps like y'all's, to video games, to websites, to open source projects, to big software and OS packages. Could you imagine? Am I too hopeful? I don't know if you're too hopeful, but I don't know if I share the same amount of hope that you have. I don't know, man. I am really enjoying my job right now.

And I am also very worried that my job will effectively disappear any second now. So for the hope on efficiency, this has come up a few times when we've been talking about this, like increased prices in RAM and stuff like that. Like maybe it will make people make more efficient software .

I'll address the AI and bug things separately from the RAM. But like the RAM thing is like it's a hardware situation. RAM is more expensive now. There's less of it in devices. We have to presume, like we don't know, but like we have to presume that if the RAMageddon had not occurred, devices sold today, some of them might have more RAM.

Like that there was a planned RAM increase that if the roadmap had been followed for these products from whatever manufacturer, they might have had more RAM in it. But because RAM is so expensive, they're like, oh, we can't . We can't do that RAM increase. We got a whole line. Maybe even iPhones have been subject to that. And then obviously for people buying their own computers, they're buying them with less RAM because it costs so much money.

I just bought a Mac with less RAM than I would have wanted because it costs so much money. So that RAM situation is real. And the question is, how does the world react to that? And the hopeful Chris in his hopeful way saying, maybe this will motivate programmers to use RAM more efficiently.

And I think this is how this is actually going to play out. There is some motivation to do more with less RAM when RAM is expensive. But the way that is manifesting, the way I've seen it manif esting, which makes sense to me, is there is a lot of research, development, and effort on the part of the biggest new RAM consumers to find a way to do what they want to do with less RAM.

And those RAM consumers, LLMs. There are so many different technologies surrounding LLMs that the purpose of them is we want to do what we want to do, but we want to do it with less RAM, whether it's the sp arse models that don't load all of the weights and everything or trying to find a way to make it.

You see it in all the LLM vendors, trying to take a model and like, here's a new version of our model. What is the selling point? It costs less. It uses fewer resources. It requires less RAM, but still is as good or better than the one we came before. That is where I'm seeing a lot of the motivation because those are the people, they're the cause of Ramageddon, right?

They're sucking up all the RAM, but also they're incredibly highly motivated to figure out how to do what they do more cheaply because that increases their profit margins and it makes their product better. And like they, you know, they want more margins. They want to lower prices for consumers so they get more business because it's a competitive business.

Come to our thing, we'll charge you less money and our things are better. And so they are highly motivated to do that. Apple is highly motivated to figure out how do we run models that are worth a damn on phones that we can't increase the RAM on. And so, so much of the effort that has gone into Apple intelligence has been about how the heck do we get whatever models that we just bought from Google or whatever models we have, how do we get them on the phone?

How do we get them to actually run locally and do something useful? And that has been a huge effort within Apple. What it hasn't motivated is let's make pages on the Mac use less RAM. No, that's not happening, right? Let's make the finder use less RAM. Let's, let's do a RAM audit of the entire system.

Probably some of that is going on, but you don't get a lot of bang for your buck for that because all the action is happening in the, you know, the big, let's, let's look at the big RAM consumers. The big RAM consumers are, oh, you're suddenly running, you know, umpteen billion per annum or LLM on your phone.

And it's using like a huge percentage of your RAM. Forget about how much RAM notes is taking unless it is massively increased. Seriously, we're going to get the most bang for the buck for figuring out how to make our models use less RAM, but still be efficient or whatever. So you'd like to think that like pressure in one area will cause this like, uh, reevaluation of the whole thing.

We've, we've been wasting RAM all this time. We should be more efficient everywhere. Like the idea of let's lighten all the screws on our, you know, Gordon Murray automotive vehicle. I'm just going to make all the screws lighter everywhere and we'll be such a big savings. No, what they do is say, look, the engine block is made of iron and we're going to make it out of aluminum.

And that's where we get the most bang for our buck and forget about the screws. So that's, that's my take on the RAM thing. I think there is motivation to do it, but it is all happening in the parts that you don't, not that you don't care about, but like in, in the LLM's parts and not in the whole rest of the operating system.

Even though in theory finding, you know, replacing all the screws in every single app on our entire Mac with titanium screws would give you a aggregate overall savings. It's much better to, instead of having a thousand people do that, put a hundred people on figuring out how to make the models take up less RAM.

And as for the coding and bug squishing and stuff like that , all the studies I've seen about this is that a lot more code is being generated and a lot more bug reports are being generated. There is still some debate about how, you know, what is the overall productivity benefit?

Because you have to filter through the bug reports. Like I saw a recent sort of summary of like, look, coding agents are finding a huge number of bugs. Some of them are serious. We've seen the ones that are serious, right? But there's just so much volume that the percentage is still low. So the volume of bug reports we have to deal with and sort of validate is huge.

And because the volume has gone up so much, it's also finding more serious ones than, you know, ever before. So is this increasing productivity or is it weighing us down with having to deal with all of the false ones and figure that out? And the same thing with code. You can generate a lot of code faster.

You can make individual developers much more productive. But I think it is still an open question about the long- term viability of continuing in this direction. Do you end up with a code base that's 20 times the size that no human can understand anymore? Is that fine if the models can understand it?

Or do we reach a breaking point where now the models can't understand it either and the humans haven't understand it in 10 years and now you're kind of screwed? We don't know how this is going to end up because the models aren't static. Like things aren't staying the same. Every day they are changing and, in theory, improving. So we don't know where this is going to end. But I don't predict a beautiful golden age where everything uses less RAM and all the bugs are erased and everybody is more productive and wonderful.

I don't think that's where we're heading. We may get there eventually in 100 years, but it's going to be rough between now and then. And it has to be. Like I compared it to early on when we were talking about this. I compared it to the Industrial Revolution. The Industrial Revolution was really bad for a long time, especially in the beginning.

And we're glad we have it because we like indoor plumbing. Like we ended up in some good spots, although we also cooked the entire planet. So that's bad. I enjoy the optimism. People think this is going to produce a golden age, but you can't really hold that vision in your head unless you're willing to not look at any of the downsides, short term and long term.

And there will be short term and long term downsides. There will be upsides, too. But it's too optimistic to think we're going to get all the good things and none of the bad things. We're going to get it all. I think going back to the actual question of like, you know , can AI produce, you know, better software or will this result in better software that's more efficient, that's less wasteful of resources and everything.

And I think there's a couple angles to this that I also want to add. One is that, you know, if you look at what actually uses lots of RAM on your computer, a huge amount of most people 's actual RAM usage is the web browser. And that's mostly just related to basically active tab management.

How many tabs do you keep open? How many tabs does the browser kind of keep in memory versus some kind of serialized format or some kind of like, you know, sleeping format? Yeah, John. Don't blame me. My browser. I can test the fact that browser tabs are the major user of memory. You know what the top user memory of most people's Macs is right now as you speak?

Chrome. File system caches. Well. Because unused RAM is wasted RAM. Any RAM you are not using, macOS will use as a file system cache. You don't think about it too much, but like you will always , ideally, macOS will use 100% of your RAM at all times.

But a lot of that RAM is going to be file system caches so that when you try to do some file IO, it doesn't even go to the file system. It's like, I've already got that in RAM. And then it just gives it to you. Because that's a good operating system. It's like, don't waste it. Don't let it. Obviously, if you need to allocate memory, it will pull it from that pool or whatever.

But like, yeah, file system caches. But I get what Marco is saying in terms of like, forget about the operating system infrastructure. What apps that most people are using and using RAM? Browsers do use a lot. Yeah. And of course, you know, if you're a developer, Apple does not care at all about the resource usage on your computer of Xcode and its tooling.

Apple will use all of the memory because it certainly seems like their own engineers use very high-end computers and they don't care about memory or disk space at all and will waste all of it. And honestly, most of us would trade that if it would make the compiles faster. Yeah.

But for the most part, I think going back to the question, it is possible to make software use less RAM. But you have to care about that and seek that out. Now, what the LLMs make easier and better is whatever you want to make happen, they can help make that happen more quickly and many cases with fewer bugs than you trying to do it yourself.

But definitely more quickly. And what that means is if you want to tell it, hey, examine this algorithm and see if there's optimizations you can make to make it faster. It will do that. But if you didn't ask it to, it might not. And I have found this with my expansion of AI coding use myself.

When you just ask it to make a feature or to build some kind of functionality, the first attempt it does, it does not pay that much attention to performance because maybe it didn't need to. Maybe it didn't think it needed to. Maybe it didn't have enough data or maybe it just tried something random until it worked.

But then when I ask it, try to find a way to make that faster. It immediately finds a bunch of ways to make it faster. This has been true of software efficiency forever.

If you cared about making your software efficient and you were able to and had the initiative to take time to try to speed it up, profile things, see what actually is taking up a lot of time or a lot of memory, and then try to figure out, is there any low-hanging fruit here?

There almost always is. There almost always is. But you have to have that as a value and a task list item, and most people and most companies don't. Most people in most situations either never think about performance at all, or their jobs require them to keep doing stuff at a certain pace or to prioritize different things so they don't really have the resources they need to care about performance.

And that's a shame. But LLMs make it easier and faster to improve performance and to reduce memory usage and to be more efficient. But you have to ask them. And not everybody will. So I think we are in a golden age of software development right now with AI development.

But I don't think this is an inevitable side effect of it. What we have is more software coming out faster. And if people ask for it, it can be better software. But that's a huge if. We will at least, we are guaranteed to have more software faster.

That it's very good at. But the quality part still depends on the creator of that software valuing that and asking for it. It's like the bureaucratic problem with security updates. There is the problem of how can we update macOS without rebooting. That's a technical problem.

But the bureaucratic problem is how can the organizational structure within Apple get security updates out the door in a timely manner. And that is not a technological problem. That's exactly the same problem with like, oh, make this use less RAM or spend your time hunting bugs or whatever. That's a bureaucratic problem. That is a problem of like organizationally, where do we ask people to spend their time?

Forget about the tools. Forget about anything. This is the same problem it's always been. It's just that now we have these powerful tools. But to your point, Margo, the powerful tools are worth nothing if what you're asking them to do is bad. If you're never asking them to find bugs. That's why, you know, the whole release of like, you know, a Snow Leopard type release or like the 27, we spent a lot of time fixing bugs.

The first part is you have to say you can spend your time looking for bugs. And then we have these amazing tools that will help you find the bugs. But if we don't tell you, you can spend that time and say instead you have to spend your time implementing features X , Y, and Z, the bugs aren't going to magically fix themselves. Somebody has to spend the time to fix them.

And, you know, so that's that's the other, you know, human nature being what it is. It's very difficult to, you know, have people say, oh, we 're in this crisis where we don't have as much RAM. Everyone will look inward and find their inner optimizer and spend their time optimizing instead of spending their time adding features that will produce more sales in their apps.

Nope, that's not going to happen. Like it doesn't change people's hearts, right? It doesn't change the hearts of the managers at Apple. It doesn't change the profit incentives that that dictate what software is built and how it's built. If anything, the ease of making new software is like, you know, I can make a buck faster by slapping together some crap and throwing it out there because that is the optimal strategy for making the maximum money.

You're like, but wait, that's not good software design. And it's like you have misread my motivations. I'm here to throw out some crappy apps that are vibe coded and make a quick buck off of them. And yeah, that's I mean, that's that's true of all these things. Like, and you know, there's there's an upside to it.

Like, for example, YouTube democratized and made it easier to make video. And there is way more garbage video than has ever been before. But there's also way more good video. And YouTube, part of the magic of that is that, you know, in its way tries to make it so that you are not burdened by the crap, even though there is, you know, I forget what the ratio of like.

How many hours of video are uploaded to YouTube every like hour, every hour of clock time, how many hours of video uploaded? It's astronomical. But as a user of YouTube, you don't have to see all that. We'll see how that shakes out in software. We all know way more software is being created than has ever been created before, because now non developers can create it and developers can create software faster.

And it's overwhelming if you're in this industry. The number of like, you know, unsolicited emails we get about, hey, I have this new app that's coming out is 10x what it used to be for sensible reasons. The app store is seeing that burden, too. I'm not sure what the equivalent of the YouTube algorithm is to try to surface the ones that are good.

But that is the challenge we're faced with. And this is now it manifests as a burden on all of us of like, I can't keep track of this. Every single person is creating apps now. And it's amazing and beautiful. And it's great for those people. But like it does make the market for software, perhaps at least temporarily, extremely overwhelming and hard to tell good from bad.

All right. Thank you to our sponsors this episode. Uncommon Goods, Base44 and Quintosh. And thanks to our members who support us directly. You can join us at ATP.fm slash join. One of the many perks of membership is ATP Overtime, our weekly bonus topic.

This week on Overtime, we're going to be talking about future phone form factors. This is an interesting time for this. So there it is in Overtime. ATP.fm slash join to listen. Thank you so much for listening, everybody. And we'll talk to you next week.

And you can find the show notes at ATP.fm.

Because it was accidental. Accidental. It was accidental. Accidental. And you can find the show notes at ATP.fm. And if you're into Mastodon, you can follow them at...

C-A-S-E-Y-L-I-S-S. So that's Casey Liss. M-A-R-C-O-A-R-M-N-G. Marco Arman. S-I-R-A-C-U-S-A-C-R-A-C-U-S-A.

It's accidental. Accidental. They didn't mean to. Accidental. Accidental. Marco, where we left our hero, I had begged of you, begged of you, if you're going to try the iPhone Duo, give it two weeks if you can and a week if you can't.

I received a, or we received a message in Slack like 12 hours after that conversation saying, 18 hours with the Duo as my iPhone and I'm out. You complete freaking wuss.

What is your deal? Well, I reconsidered my out. Oh, I take it all back. See, I should have started with honey instead of vinegar. I'm sorry. I take it all back. I take it all back. All right. Did you last at least a week? I lasted exactly a week.

Okay, but then I have no poo-poo to give you. I did this ridiculous thing for you. All right. Well, then I take all my snark. I take it all back. You just didn't want to get yelled at on the show. That's right. I think that's mission accomplished in my book. Okay, so now I don't have to be quite so cranky about it.

So what happened? Okay, so I used it as my phone for one week. I had a lot of things to do this week. It had a lot of varied use in a lot of different places and cases. Let me tell you what I liked first.

When I wanted to browse the web or look at a photo largely, opening up to the inner screen and having all that screen real estate was very nice, actually. I really did enjoy and appreciate being able to view more of a web page, being able to view a big wide photo.

At some point, a couple times I watched video by kind of propping it up on something because you can't prop it up on itself in big mode. But okay. I watched video that way a couple times. Those were all very nice. I also, I guess I can say one more nice thing about it.

I was like, were we at the end of the good things with the one item? Wow. Well, I mean, it's a big item. I will say my concern about Touch ID was mostly mitigated. Touch ID is not a problem if and only if you wear an Apple Watch full time and you enable the watch unlock.

If you do that, it is fine. It is not perfect. Neither is Face ID. But it is about as convenient as Face ID if you wear an Apple Watch. Okay. There's a lot about this that didn't work for me. Now, a lot of people are really excited about the Duo and it's working great for them.

Let me tell you some small issues and some big issues. First of all, it is a very difficult phone to hold comfortably because it's a very difficult phone to have good cases for. So if what you want is a little bit more grip or a little bit of protection, what you're going to add is a huge amount of bulk.

And that bulk largely either won't cover the hinge, which is the main area that you'd like touch and it needs a lot of protection. Or it has to add even more bulk and have some kind of folio thing like Apple's folio case. The case situation is not good.

I decided instead of a case, a few days into it, I added grip tape to it, which I can actually strongly recommend. So there's, it's just, there's literally, it's like, you know, rubberized kind of grippy, you know, sticky decals.

There's a bunch of them all over Amazon. I added a big square to the back and I added a nice thin skinny one to the edge on the right side when it's closed that does not have the buttons. So like the top edge and having grip tape on it.

And unfortunately, where I most wanted it is the hinge, but you can't do that. Obviously you can't cover that. You can't, you can't put tape on the hinge and we'll close or it won't open rather. So the grip tape actually made it a lot easier for me to handle. Pull it in and out of my pocket with the, you know, the big square one on the back of it.

That really helps like help get my grip on it. The side one on the right side kind of makes me feel a little more secure holding it. And it actually stayed the whole time. So I'll put a link to the one in the show notes that I used . I think it's called like dragon tape or something. Dragon grips.

That's it. Is it going to come off clean? Is it going to leave sticky junk? It will probably leave sticky junk, but I can just, you know, I can wipe it off. It's, I mean, it's on a polished steel surface on a piece of glass. So it's not one of those like magic gecko feet, like non sticky stuff removing stuff.

No. Oh no. Like the micro suction. No, it's not doing that. It's, it's real sticky stuff. And I even did the thing where it says like, use a hairdry er or a heat gun to make it stick better. So I did. I went into the bathroom, got Tiff's hairdryer and I blew them on there. I tried this at home.

Your poor phone. It was fine. It's supposed to be connected to a cooler, not a blowing a hairdryer at it. I mean, for a couple minutes, it was fine. Anyway, so it stayed like the edges didn't peel up or anything. So that's actually been really nice. And I have now, I'll spoiler, I've switched back to a different phone and I've actually put a little square of grip tape on the back of it because it's actually, I like, I like that so much during, from my week experiment here.

So I'm trying that without, instead of a case on the new phone also. Anyway, fundamentally, I think they got the aspect ratio wrong. And a lot of things fall out of that. What most of the Android world has settled on is basically a regular tall rectangle phone that unfolds into something close to a square.

And we've heard over the last few months as Apple was getting ready to launch this thing that Apple had a better solution. This is much better. And I think a lot of people might still agree with that. But I think after using it, I don't agree with that. When it's closed, it's a terrible iPhone. The number one problem is text input with the keyboard.

Because when the keyboard is up on a closed screen, you have room for very little in terms of messages or notes or whatever you want, whatever you're writing it into. And messaging is one of the most common uses for a phone. It might even be the most common application people use on their phones.

Some form of messaging, whether it's WhatsApp or iMessage or whatever. And messaging on this is terrible. Because when the keyboard's up, you can only see a few lines. Now, if you open the phone up, well, if you're opening it up in landscape orientation, it's still just as bad.

Because the keyboard in landscape is still half the screen. And in the left column, you get like the sidebar. And so then on the right, you have the same crappy little like, you know, two and a half, three lines of text that you had when it was closed. You really only get more useful real estate with text type applications if you open it up and also rotate it to portrait.

And then you get a little more space. But even then, it's not, you're not seeing that much more text, even in the giant portrait mode with the keyboard up as you would see on a regular shaped iPhone. You do see more, but it's not as much more as you would think.

Because again, the keyboard is massive. I just think they got the aspect ratio wrong. What I wanted the Duo to be was when folded closer to a regular iPhone shape and then opening it would simply be additive.

But it's not that. It really is a pretty bad iPhone when it's closed. And then it becomes a pretty nice, you know, large screen when it's open and in portrait. I would say open in landscape, I didn't find a lot of use for it besides like watching videos.

I think that was amazing. But open portrait, I did like that for browsing things, you know, reading things. However, the software is not there yet, even for that. Because here's the problem. If you, again, I think one of the, one of my favorite things to do on it was open it up and like read, you know, a webpage, a document, you know, something in an RSS reader or whatever.

But if you open it in portrait, the margins are all wrong on almost every iPhone app. Because these are not iPad apps that are running. They're iPhone apps with different bounds. And when you lay out an iPhone app, you lay it out with pretty thin side margins.

So what you have when you open this up is a very wide column of text with very small margins in almost every app. Now, over time, I think the apps will probably start designing their way out of this. But what you have today is your fingers are too close to the text.

And the columns are actually too wide to comfortably read them. My least favorite app to use on the Duo is Mail. Mail sucks on the Duo. Because you have, okay, when you're composing, you have the keyboard problem.

Where it covers up half the screen in almost every orientation. And it's terrible. When you're reading Mail on the Duo, if you're on the outer screen, you have, you know, no space. If you're on the inner screen, both orientations suck. Because if you are in landscape, again, it's kind of the short squat landscape.

You don't have a lot of reading space. But at least you have the two columns. If you're in portrait, there's no way to pin the column on the left so it's always visible. So you're constantly opening up the sidebar to deal with all the messages and everything. Also, though, all the swipe gestures on the message don't work very easily. Because one of the swipes is close the sidebar.

So I found one of the most obnoxious things to do on this phone was anything related to mail. There is no good layout for mail on this phone. And that, unfortunately, probably will apply to a lot of different types of apps. Like using the AI apps that are basically, you know, a side bar list of your AI chats.

And, you know, the stuff on the right. You have a similar problem there. Not quite as bad as mail. But really, mail and messaging. I just couldn't stand on this phone. And the problem is, I do a lot of those things. That is actually most of what I do on a phone, as I have learned.

Also, when you are in portrait mode, reading something long , which is, again, this is all these tradeoffs you're making to get this phone. You're doing it to get the inside screen. So using the inside screen should be amazing in as many cases as possible. Well, when you are reading things on the inside screen, you know what's really not amazing?

Is that under screen camera. When you're holding the phone in portrait, it is like right where text will most likely be that you are reading. Because, again, nothing's split out with very wide margins. So almost everything is going to put text right under that in portrait orientation.

And that was very distracting for me as I was trying to read things on the Duo. I was hoping the Apple Pencil use would be really useful to me. But because the Duo just rocks like crazy on any flat surface, it's not.

Unless it is handheld in very, very specific ways. I found one use for the Pencil that I actually succeeded in , which was basically I was doing one of the furniture layout apps. So I had scanned a room, except, wait a minute, you can't scan a room for a furniture layout app with the Duo because there's no LiDAR sensor on the camera.

Yeah, yeah, yeah. So I had to take out my 17 Pro, scan the room, transfer everything over. It was such a farce. But then when I got everything over to the phone, like, I was out, like, getting coffee. And I just had the phone and my bag with me.

And I had the Apple Pencil in the bag. And I was trying to, like, move this little tiny nightstand around this app with my finger. And I couldn't quite do it. I'm like, oh, let me get the pencil out. So I got the pencil out. And I was able to, like, precisely move it with the pencil. That was nice. But that was, you know, handheld. Taking notes on it, and unfortunately, as somebody who makes an app to do this, I just don't think it's very useful because it just does not sit flat on it.

surface whatsoever. Now, I actually custom designed and 3D printed a little desk, basically like a wedge that it can sit on. So it sits flat on a desk. And maybe I'll release the model. I don't know. It's not really done yet. But it's done enough to work.

That made it a little bit better. But, like, I'm not going to carry this piece of plastic everywhere I go just for the chance that I might want the Apple Pencil on it. Also, the USB-C Apple Pencil sucks and constantly dischar ges itself. It needs to be recharged. I take it out of my bag and I try it.

It doesn't work. Like, ugh. Anyway. So not the pencil aspect. The pencil use cases have not really come through for me. Video playback. I mentioned it's good if you are standing it up in portrait mode with something else. But if you want to, like, watch a video on it when you don 't have anything to stand it up on and kind of use it as its own stand, that seems like a foldable win.

Except the way that you would do that where you kind of fold it like a laptop, well, then the outer screen is face down on a surface. Which, if you care about screen scratches, you would never want to do that. Like, what are you putting it down on?

Is there a speck of grit on that surface? Are you putting it down on anything rough? Like, forget it. That's why Apple sells that folio case, which both protects the screen and screen down, but also, I believe, has a kick stand that allows you to prop it up in landscape open mode. Yes. But that case is very nice.

I felt when in an Apple store for a few minutes, like, I think the folio case makes a decent case for itself, but it adds a lot of bulk. And then you also then, it's closing over the front screen all the time. So you have to, like, flip it open to see your front screen .

Already, this phone is kind of clumsy to, like, you know, get it in and out of a pocket and start using because it's so bulky and heavy and touch ID. So, like, to add more friction to that process, I don't think is amazing. I also did have the problem of frequent pocket lint accum ulating on the inside screen.

Like, when I would open it up, like, the inside screen crease area would have a little bit of pocket lint or, like , one hair or something. I have to brush it away. A lot of the use cases for this, I think it's optimized for someone. And that someone is definitely not me.

And you have a, it's kind of, this is not a great example. Kind of like owning a boat. Gracious. To make owning a boat worthwhile, there's a ton of hassles and downsides and costs.

But if you really love boating, it can be worth it to you. But if you just want something to get you from place A to place B, buying your own boat is probably not the best way to do that. I thought the boat analogy you were going to use is the two happiest days of a boat owner is the day you buy your boat and the day you sell it.

Yeah, I heard that too. Two happiest days of Marco's iPhone due ownership is the day he finally got his order in for one and the day he stopped using it. That's more cruel but not entirely wrong. You have a lot of tradeoffs and downsides to get that inner screen.

And so it has to be really worth it to you. And you have to use your phone for apps and priorities that are pretty different from mine for that to be worth it to you. So it's not that it's for nobody or that it's a flop or anything.

It's just a, it's a very extreme set of tradeoffs. And for me personally, it is not what I want at all. It might be for you. And I think if you don't do a lot of text entry, especially if you don't use Apple Mail, if you don't use, if you don't do a lot of messaging, if you don't do a lot of text entry, if you do a lot of like, you know, media consumption type

stuff, reading, maybe like artistic things. It might be better for you. Maybe if you want to like view large documents, like if you 're working on spreadsheets and stuff, maybe that might be better. But it is a pretty rough iPhone for the things I actually do on an iPhone.

Yeah. When I would talk about the aspect ratio, that's why when we were talking about like the cleverness of the software, I've always been framing it as like, given the form factor, the chosen aspect ratio. If you're a software designer, like, you know, make the best of the situation. But it's always with the idea that, you know, you're not determining the form factor and the form factor determination is made, you know, like that the software falls out of that.

But like that form factor, it seems like it was made to emphasize what Apple thinks are the most important use cases. And they're not the ones that you just mentioned. Right. So the obvious downside of taking like an iPhone 17 type thing and just making it fold in half is that when you open it, it's squarish.

And then worst case scenario, I don't know if the measurements on this, but if you're if you're tall phone, regular phone is more than twice as tall as it is wide. When you open it and you watch a, you know, 16 by 9, like full screen video on it, doesn't matter what orientation you turn it in.

That video may actually be physically larger if you hadn't unfolded it. That is the downside of a square screen. It's like I've got this big thing. And when I unfolded, my TV show gets smaller. Yeah, because now you're taking, you know, you're you're taking what was a vertical, you know, tall thing and turning it sideways and sticking it on a thing that is not as wide as your phone is if you had just turned it sideways .

And that makes me think that Apple has really prioritized 16 by 9 ish media consumption as a use case. And again, before this phone came out, we talked about MKB HD video where he was just harping on this because he'd been testing folding phones for years and years.

And like his complaint, his specific complaint was, look at all these phones. I open them up. I try to watch a TV show or movie and it's smaller than if I just kept it closed and rotated it. Or if I use a regular phone that didn't fold and just rotated it. And that seems absurd. And it does seem absurd. But Marco, what you point out is in exchange for not making that absurd, there are other sacrifices.

The thing is shorter. Like the keyboard thing really hits home with me because I frequently find myself adding things to our Google Docs show notes document from my phone. I have an 18 Pro and before that a 16 Pro. It's a tall phone.

And every time I go to that Google Doc and have to pull up that keyboard, like I cannot see anything anymore. And this is not the iPhone Duo. This is a very tall phone. It's not a Max, but it's a very tall phone. And I'm like, the second that keyboard comes up, I think I 'm looking through a tiny like slit in a tank or something.

Like I can't see. How can I edit the document? I can't see anything. This keyboard is taking up the entire screen. So I cannot imagine doing edits like that on the Duo with its much less tall screen. But yeah, like the prioritization, like Apple picked this form factor seemingly to solve that specific problem of video.

And maybe what they know and we don't is that most people, the thing they want a bigger screen for is to watch TV shows and movies. And like there's something to be said for that because I do see a lot of people watching video on their phone in various scenarios. But on the flip side of that, I also see a lot of people watching vertical video.

And vertical video is not great on the closed Duo either. No, it's really not. Right. Now on an open Duo, vertical video is also good. Again, that is the advantage of having it open to be rectangular and not square because it gives you a long edge . So vertical video, horizontal video, there is always a long edge.

You just rotate the thing and it will work. Right. But it just seems like this use case is like that they really prioritize that. And maybe, you know, to your point, maybe that is the correct prioritization for most people. Maybe most people aren't doing email on their phone all day . And maybe for messaging purposes, people don't care that they can only see one line.

Like, I don't know. You know, we'll find out from the sales figures. But that is the like, why is the phone look like the way it does? So when it opens, it's not squarish. It is instead more rectangularish. And Apple, Apple's not the first one to do this. Their copycats beat them to it or whatever. But the market will decide, is this the form factor we want this to be?

There's tons of phones that are regular, like iPhone 18s that open and close. We have the flip phones that bend in half vertically. People have tried lots of different form factors. Apple is essentially popularizing this one, which was not popular before they did it. And we'll see what happens. Like, the good thing about this folding phone tech and everything is like, changing the aspect ratio is not fundamentally rethinking it.

You still have to have a hinge. You still have to have the screen. Like, if they decide, actually, we have some different ideas about aspect ratios, you can make that change without saying, we've got to throw everything out and start all over. Because a folding phone is a folding phone is a folding phone. It's just a question of the sizes of the rectangles and how you apportion the space.

But yeah, I do wonder if they've made the correct trade-off . Obviously, not the correct trade-off for Marco, but I do wonder everyone else who has this phone, how they come down. Hi. For whatever it's worth, by the way, though, on that front. So I also paid attention this week. I spent a good amount of time in Manhattan this week, and I saw a number of people with duos.

A few, like, passers-by. I was getting my eyes checked, and, you know, there was somebody, like, in the waiting area who was using a duo the whole time. And then I was at a coffee shop, and I saw somebody using a duo for, like, a half hour.

None of them ever opened it. All of them were using it in closed mode the whole time. Wasn't sponge-worthy. And I noticed, too, like, in my own usage of it, a lot of times I had to remember, oh, yeah, I could open it right now.

But I would almost feel like, ah, I can't be bothered. Because it is kind of, like, a thing. Like, it did draw a little bit of attention sometimes. Like, some guy on Subway platform asked if he could hold it . I said no. Like, I did get a lot of, you know, a lot of questions.

Oh, is that worth it? I'm waiting for mine. How do you like it? And, you know, I was honest. I'm like, eh, it's not really for me, you know. But in reality, most of what most people do seems to be typing. Like, the one in the coffee shop, the guy was literally typing like mad the whole time I was there.

It was like a half hour. And he never opened it. And at first I was like, why isn't he opening it? But then I thought, well, if I'm, like, chatting with somebody on messaging for a while, I don't open it either. Because the open version of messaging isn't better. Because the keyboard's the wrong size. And so, like, most of what I was doing with it was closed.

And I think, again, I think most people, most of the time, they're going to be using it closed. Because that's what most people do on their phones. I think it would be in Apple's best interest to be willing to reevaluate the aspect ratio decision. Because that really, like, everything else falls out of that.

And, you know, the design of the software, having the stuff on the right sidebar. And, like, they did, basically they optimized the phone for making the best, in their mind, the best open experience for media.

And the elegance of having the inside screen just be two of the outside screens side by side. And having, like, you know, split view and everything. And, okay, those are, you know, moderate use cases for sure . But is it better than messaging?

Should that take higher priority for someone's phone than messaging? I'm going to say no. So much of the phone would be better if the outer screen was just taller. And you could make it also narrower.

Make it, like, if the outer screen was basically a regular iPhone Pro size. Or maybe even a little bit smaller, but still generally iPhone Pro shaped. It would solve so many of these downsides.

And the only resulting downside from that change would be the inner screen would be closer to a square. And it would be a little less elegant in certain ways. But, honestly, I think that's a worthwhile tradeoff for a phone. Because it's not a TV.

It's a phone. And a lot of that's, like, so many of those limitations and downsides and problems would be alleviated by just changing the aspect ratio and making it a little taller and skinnier . Are you ready for me to tell you how wrong you are? Yes.

Please. Let it rip. All right. So here's the thing. All snark aside. I actually don't particularly disagree with anything you've said. Like, I disagree with you, but you're not wrong, maybe is a better way of putting it. That I don't know that I would make the tradeoff of a taller outer screen and a, you know, more squarish inner screen.

But nothing you said about that was empirically incorrect, right? So as you were talking, I was like, all right, I do think Marco has a point. I'm thinking to myself that there is less visible on the duo, on the internal screen, on the innie, than you would have expected, right?

But am I right about that? Is he right about that? So I brought up, and I'm not going to put these pictures in the show notes because it's kind of private stuff. Like, I'm going to share it with Marco and John, but I'm not going to put it in the show notes. And I apologize for that. But I just, I didn't know we were going to go this direction, and so I didn't prep anything. But what I did was, we have a shared note, the three of us.

So that's ATP merch promo codes, where Cotton Bureau will occasionally send us promo codes, and occasionally we'll, you know, tweet them out or whatever. But we have a shared note where it's a preamble at the top, and then just a long list of numbers from Cotton Bureau. And I looked at it on my 17 Pro, and there's a preamble at the top.

And with the keyboard up, with the keyboard there, I can see two promo codes on my iPhone 17 Pro. If I'm on the inner screen in portrait, I can see one, two, three, four, five, six, seven, eight promo codes with the keyboard up.

So this is apples to apples. The keyboard is up. Now, obviously, I'm on the inner screen, not the outer screen. But the keyboard up on the inner screen, I can see twice as many of these promo codes, which is my closest I can get to John's point about, oh, my God, I can't see anything on Google Docs.

Now, Google Docs, of course, is not updated for the Duo. Probably won't be for 10 freaking years. And so I can't do anything about Google Docs specifically. But hopefully you see what I'm driving at, is that there's actually a fair bit more in notes on the Duo when unfolded, when in portrait.

So, I mean, I'm not trying to take away everything you're saying, Marco. I did not look at it on the outer display, partially because that ruins my whole point, but also because I just didn't have the chance. On this specific topic, though, again, we're not going to share the note here, but, like, the preamble is a paragraph of prose followed by a bulleted list that is short.

And for the purpose of messaging, if people are mostly sending short one-line messages, the message bubbles and the margins and everything are going to be pretty harsh. Like, the benefit you're getting of this being open in portrait is that the paragraph has longer lines, so it wraps less.

And that's what's pushing all those things down. Now, physically speaking, you are right that there is physically more pixels and more data in the portrait mode open Duo. But, like, that is a massive, massive open screen compared to a closed, tall, skinny phone.

So, I think, like, the space efficiency of your entirely open best-case scenario keyboard up Duo depends a lot on how long the messages are that you are sending back and forth. If it's single-line short messages, it really eats into the advantage.

Where if it's paragraphs of text, then the open portrait keyboard up Duo wins. But I feel like this is kind of not a fair fight. It's your point. Like, oh, well, this one does unfold. But, like, you have given the best possible scenario for the Duo. No, I certainly have.

But I knew you were going to say this because I know you, John, and you're a pain in my ass. And so, the next thing I did was, and again, we will not share these, but I looked at a text thread between myself and a couple of friends. And on the iPhone 17 Pro, I can see the last three messages in their entirety.

And then the fourth message, so the fourth oldest message, you can read it, but it's peer with the name of the group chat, if that makes sense. So, it's, like, barely, it's barely within the acceptable range of visibility. That's four messages I can see with the keyboard up.

With the inner screen, I can see one, two, three, four, five, six messages. Now, is that night and day different? Of course not. But it is more. But again, you have a message that wraps three lines, and then it wraps the five lines on the narrow phone. If they were all one-line messages, it would be eaten into.

So, you do need the wrapping advantage because you're taking advantage of the width of the width of the unfolded thing. Like, if that middle message wasn't there, it would be closer. I think you're probably right. But, I mean, I don't know. I'm looking at, like, my conversation with Aaron. You know what that is? It's a bunch of links that we sent back and forth.

So, it's a similar thing. I mean, I get your point. And here again, you're not wrong. I'm not trying to say that Marco's wrong. I'm not trying to say that you are wrong. But I think you're both overblowing. I'm not overblowing. I don't have this phone. It's not my opinion at all. I was just explaining why the Duo was shaped the way it was and how it maybe means that Apple is really concentrating on the video consumption thing.

I will also point out that the message, the second from bottom message, also wraps to two liners in the narrow phone. So, really, the wrapping is doing it here. Yeah, but, I mean, that doesn't negate my point, though, John, that it's wrapping or not. It's still more stuff on the Duo. Right, right. What I'm saying is if they're all single-line messages, the advantage is very, very small.

If they're multi-line messages, the advantage grows. And specifically for messaging, like a better case would be like reading an article because that's not going to be single-line paragraphs. You know what I mean? You'll definitely get more use there. But then if you're reading an article, you wouldn't have the keyboard up, you know? So, anyway. Right.

But, I mean, how many – I surely do not send paragraphs every time I send a text message. But I also wouldn't say that my text messages are one or two words each. Like, again, you're not wrong. But I feel like I talk in such a way that I have messages that wrap, you know, on the thin phones and on the wide phones.

So, I don't think this is as unreasonable a demonstration as you seem to be implying. But, nevertheless, again, I don't think Marco is wrong. But I do disagree. I was doing the dishes a little while ago. And I have been re-watching – well, I watched most of And or, and it did not click for me.

This was months ago. Didn't really get it. Just hold on. Just hold on. It didn't really get it. This does not shock me, but go on. No, hold on. Hold on. I didn't get it. It didn't really click for me. Like, it was fine. And remember that, like, I enjoy Star Wars. I'm not a Star Wars super nerd like everyone else I know. But I really thought, you know what?

I don't think I gave this a fair shake. And over the last week or so, every great once in a while, I'll sit down and I will properly watch the show. And I think that's where I went wrong before is I was doing the standard like, oh, this is on and I'm also futzing around with my phone. And so I think that's my fault.

I did it wrong because I wasn't properly focused. And now that I'm properly focused, it turns out it's actually not bad. And so I was watching the show and then I realized I needed to do just a small amount of dishes. And so I thought, okay, well, I would like to continue to watch the show.

I knew the general, the beats of the story that were going to happen. So I didn't need to super duper focus. But I'd like to be able to watch the show as I'm, you know, doing dishes. And so what I did was I put the thing in tent mode, sat it on the windowsill that's behind our kitchen sink, and I watched the dishes while I was watching the show.

Can I pause here for a second? And point out that you, quote unquote, properly watching the show is you watching it on your phone while you do dishes. No, no, no. Okay. Yes, I know. Yes, yes, yes. I know. But this was the particular scene that I was watching. I knew exactly what was going to happen. I didn't. Again, you're not wrong, but focus.

So the point is I could do something with this phone that I couldn't do before. Now, to be fair, I typically was a PopSocket person in the past. I probably could have propped it up on the PopSocket, and it arguably would have been a bigger screen. But the fact that I could do this at all, I find to be really freaking cool.

And I was at lunch today with a couple of friends. Well, it was Aaron and me and a couple of friends. And I wanted to show them a short video of something. So what did I do? I opened up the screen, and now poof, I basically have an iPad mini with me. Again, Marco isn't wrong, but I don't find the trade-offs, with the exception of the camera, both internal and

external, I don't find the trade-offs to be too burdensome. I echo what Marco says about Touch ID being mostly a nothing burger. It is unquestionably not as good as Face ID, but mostly, as long as you're wearing an Apple Watch, mostly a nothing burger. But I feel like it is incredibly great and fun to be using NetNewsWire with it open in landscape.

And I do this with my email. I happen to use the Fastmail app, but your mileage may vary . But on the left hand, like third, is your either list of articles or list of emails. And you can just go bloop, bloop, bloop, bloop, bloop, blo op, and triage everything super quickly.

And then if I want to properly dig in and read an email or read an article, you know what I do? I flip to portrait and now I have this giant screen where I can read that email or read that article. And when I'm done with it, I flip back to landscape, bloop, bloop, bloop, bloop, bloop, bloop, and I triage everything.

Now, you, anyone, Marco, John, whomever might find that to be burdensome. Look at this fricking acrobatics I need to do in order to use this device. I don't find it burdensome at all. I think it's kind of fun and kind of neat. And I don't find the external, the Audi screen to be incredibly burdensome.

Unquestionably, it would be a better phone. It would be a better phone if the outer screen were taller, but then it would kind of ruin the inner screen as we just talked about. So the outer screen is maybe a little too squat, but I've gotten used to it and I mostly don't mind it.

Does it make Marco wrong about what he was saying about messaging and so on? No. But do I mind it? No, not really. And so I can't say that I'm right and Marco's wrong. And in fact, again, I've said it several times because I really want it to stick.

Like, I don't think you're wrong, Marco. It's just the things that you find burdensome, I either don't find as burdensome or I actually think they're kind of an advantage. And especially being able to flip that screen open whenever I want. And I have no shame about doing this publicly.

None at all. Like, I do not care. I'm not trying to get people to talk to me about it, but I do not care. And I will open it publicly and I will do what I need to do and I will close it and put it away and that's that. Getting it in and out of the pocket, not a problem for me. The lint on the inside, occasionally, but you just, and then it's gone.

Like, it's not been a big deal. I had somebody write me on Mastodon, like, is yours a fingerprint magnet on the inside? No. Like, there's a little bit of fingerprints, but no. No, I didn't have that problem either. And I actually think the matte screen looks great. I concur with everyone and I know John is probably seething.

It is not perfect color reproduction. I totally get that. But I use my phone outdoors from time to time. And I think it looks great. And I actually kind of like the way that it's a little more muted. Maybe I'm just convincing myself to like this phone and maybe deep down I don't.

I just don't realize it. But I actually think it looks really great. I also agree, though, Marco, that the internal camera is way more noticeable than anyone told us it would be before these things got released. And we said it to each other when we were in New York and I still think that.

And coincidentally, when I'm holding my phone in landscape and I'm doing my triage dance in Fastmail, the freaking delete icon is like right on top of the camera. Yeah. And so it looks fuzzy every time I tap it. And that's annoying. But all in all, I actually really like this phone.

I don't know if next year I will get an iPhone Duo 2, if I will go back to a slab phone. I genuinely, truly, hand on heart, do not know. But sitting here today in basically mid-November, three weeks in or whatever it's been, I'm still really liking this phone.

And I would really encourage you, the listener, to ask yourself, like, A, do you really want to drop $2,000 plus on this freaking phone? But B, do you see yourself doing the sorts of things that this phone is great at, like triaging and reading and watching movies or videos or what have you? And if you do, you should really give it a shot and check it out.

And I would also say that I am glad you lasted a week. I really, truly am. And if you don't, if you're not even getting a hint of joy out of the phone after a week, it's probably not for you. I think two weeks would be better, but I'm not complaining about one week. If at one week you're really not getting any joy out of it, it's probably not for you.

But let me tell you, the first time that I was able to, that I realized I could do the open it up and landscape and bloop, bloop, bloop, bloop, bloop, bloop, bloop, and triage everything, rather than having the, what I never thought before was burdensome, but now I find burdensome. Like, okay, tap on an article, animate in.

Oh, I don't really care. Animate back. Tap on the next article, animate in. Oh, I don't care. Animate back. And so on. Like, not having to do that is really freeing. And I'm making a mountain out of nothing, and I get that. But it's really freeing. Additionally, going too up on the inner screen when it's in landscape, when you're trying to, like, copy something down from a text message into your calendar, or, I don't know, get two Safari windows side by side for whatever reason, that's really freaking convenient.

Really convenient. Do I remember that I can do that? Almost never. But when I do, it's really freaking convenient. And so, again, you're not wrong, and obviously, it's your opinion regardless, so I can't even pretend to say you're wrong. But I find a lot of the things that you find burdensome to be actually kind of fun or useful, and you, the listener, may fall anywhere on the spectrum between me and Marco.

And to reiterate what I said when I was playing with it in the Apple Store, again, I just have to speculate because I didn't live with this phone, but my initial impression of it was that I found the outer screen less objectionable than I thought. I said, I can imagine using this weird short squat thing. Like, and that surprised me. Like, not that it was ideal, but that it wasn't as bad as I thought it was going to be.

And I immediately saw tons of places that I would use the inner screens at larger screen space, mostly for content consumption, whether it's reading a big complicated article with lots of, like, code samples and diagrams in portrait mode. Oh, yeah. Or watching video, obviously, or stuff like that.

And I confess that I did not think a lot about how it would perform when doing things that involve typing. And I already said how annoyed I get about Google Docs on my tall phone, so maybe that would bother me. But my first impression, never having lived with the phone, was the outer screen is not as bad as I thought. And I could, all the cases that you're talking about, Casey, about the inner screen, I also envisioned myself using those and enjoying those.

Again, not having actually done it, I can't tell you what the reality would be for me. But I think I am kind of in between the two of you. Like, that I, that aspirationally, I feel like I would enjoy it as much as Casey. I don't know if I would end up where Marco is. Beep, beep, beep.