ATP

00:00:00 ◼   ► I'm kind of interested in the mini review, except I am extremely, I am extremely not interested in your mini review because I don't want to know.

00:00:07 ◼   ► Oh, my iPhone mini?

00:00:09 ◼   ► Mm hmm.

00:00:10 ◼   ► I don't want to know.

00:00:11 ◼   ► So let me tell you, I don't want to know.

00:00:14 ◼   ► I stopped the, I stopped the, the show rundown right after we got to that thing, but Marco can feel free to talk about it if he wants during the topic section, which we'll get to shortly.

00:00:27 ◼   ► Fast resolution switching on M1 Max.

00:00:30 ◼   ► Gokin Avkarogullari writes, and I believe this individual is an Apple engineer, writes, I was waiting for this.

00:00:38 ◼   ► Our display pipes are many years ahead of the industry.

00:00:41 ◼   ► And this is just one example of that.

00:00:42 ◼   ► From contrast to accessibility to color management, many teams at Apple, including my display driver team, put so much effort to it.

00:00:49 ◼   ► Glad to see it makes a difference.

00:00:50 ◼   ► Yeah, it does make a difference because I finally, well, so I finally saw a demo and somebody tweeted this.

00:00:57 ◼   ► And we'll put a link in the show notes and they have, I guess an older, an old and busted MacBook Pro on one side.

00:01:04 ◼   ► They have the current top of the line MacBook Pro, the 16 inch.

00:01:07 ◼   ► Yes.

00:01:07 ◼   ► Yep.

00:01:08 ◼   ► And then, which is obviously old and busted.

00:01:10 ◼   ► Yeah.

00:01:10 ◼   ► And then the, on the left hand side, they have a brand new MacBook Pro, which is the 13 inch.

00:01:15 ◼   ► And you know, when Marco, I think it was Marco or maybe it was John.

00:01:18 ◼   ► I think it was always John.

00:01:19 ◼   ► No, no, it's always John.

00:01:20 ◼   ► That's the rules.

00:01:21 ◼   ► Remember?

00:01:21 ◼   ► And so when John said, when, when John Marco said, when John Raffio said that the resolution switching was instant, I conceptually understood the words that were being said.

00:01:35 ◼   ► But seeing it visually is a whole nother thing.

00:01:39 ◼   ► Like this tweet, it's so ridiculous how fast this is.

00:01:44 ◼   ► I just cannot believe my eyes.

00:01:45 ◼   ► It's one of those things that like, you know, with computers, we've, we've used computers for a long time now.

00:01:51 ◼   ► And it's just one of those things that always has taken a few seconds.

00:01:54 ◼   ► Like it just, it's a thing that when you change the resolution of your screen, it goes blank for a few seconds and things kind of like pop in kind of ungracefully.

00:02:03 ◼   ► And you know, a few seconds later, everything is kind of basically where it should be.

00:02:06 ◼   ► And when you go from something that takes a few seconds and you see the transition and it's ungraceful like that to something that literally is like instant, you just like, boom, it's just like click, click, click, boom.

00:02:17 ◼   ► It's, it's shocking.

00:02:19 ◼   ► Like it, and it's, again, this is not something that they had to optimize.

00:02:22 ◼   ► This is not something that needed to be done, but it's just so damn delightful to see it.

00:02:27 ◼   ► And it's something that you don't see very often in computing where something that previously used to take a good amount of time all of a sudden takes no time.

00:02:35 ◼   ► And again, like this, most people don't change resolutions that often, although many people do plug in monitors and unplug monitors frequently.

00:02:43 ◼   ► And from what I hear, I haven't been able to test this yet, but from what I hear, that is similarly fast now.

00:02:49 ◼   ► So I am just so happy that this kind of stuff is being worked on and we can still have things in modern computing that make us go, whoa, like that really surprise us because that doesn't happen that often anymore.

00:03:03 ◼   ► Because stuff is really good pretty much everywhere.

00:03:04 ◼   ► So like, you know, to get one of those whoa moments is rare and it's just so incredibly delightful when you see one.

00:03:12 ◼   ► And this is one of those times.

00:03:13 ◼   ► I guess one of the benefits of the iOS project, you know, having started for the phone and of course they started from Mac OS 10, but they had to rip like tons of stuff out to just get it to fit into the phone and to get it to be fast enough on that very slow phone hardware.

00:03:27 ◼   ► And that sort of gave them a chance to tear everything down to the studs, so to speak, and to build back up from there without concern for legacy because they totally controlled the hardware and iOS didn't need to run anywhere except for on, you know, this phone hardware.

00:03:42 ◼   ► And so that's the perfect opportunity over the course of the next 10, 15 years to reconsider things like the display driver stack, which like Marco said, they weren't broken.

00:03:51 ◼   ► There's nothing wrong with them.

00:03:52 ◼   ► They worked fine.

00:03:53 ◼   ► Like their main job is to be reliable and, you know, do the things we want them to do.

00:03:59 ◼   ► They supported retina.

00:04:00 ◼   ► You could change resolution, you know, they, you know, it did all the things that you would expect it to do.

00:04:06 ◼   ► But if you have the opportunity to sort of rebuild that from nothing, starting from very simple display driving on the original iPhone up till today, you can end up building a display driver system, especially with the move to metal, not having to support OpenGL, not having to support other people's graphics drivers.

00:04:24 ◼   ► They control the hardware and the software stack.

00:04:26 ◼   ► It's an opportunity to do something better.

00:04:28 ◼   ► And I don't think there was huge pressure from the outside saying you really need to make resolution switching faster.

00:04:33 ◼   ► I think it is just a side effect of redoing the display driver stack to be high performance in all ways.

00:04:40 ◼   ► And one of the benefits you get basically for free for doing a good job is, oh, guess what?

00:04:44 ◼   ► Resolution change.

00:04:45 ◼   ► That's lightning fast now too.

00:04:46 ◼   ► Oh, it's extremely, extremely impressive and I'm only seeing it on video.

00:04:49 ◼   ► It's very cool.

00:04:50 ◼   ► And to your point, Marco, Sylvain Philto writes, you mentioned how fast it is to change resolution.

00:04:55 ◼   ► It is exactly as fast plugging in an external display like the ultra fine 5k.

00:04:59 ◼   ► Very cool stuff.

00:05:01 ◼   ► Tell me, Jon, about TensorFlow and forks and M1 max and so on.

00:05:06 ◼   ► So, you know, one of the things that Apple talked about in their presentation about the M1 max, they showed the system on a chip and we've talked about this in the past, how much of these system on chips is dedicated to machine learning stuff.

00:05:16 ◼   ► Obviously, you know, the benefit that Apple touted to this, oh, we'll take our cruddy 720p camera and it will make those pictures look a little bit better.

00:05:23 ◼   ► Although that may be the image signal processor and not the neural engine.

00:05:26 ◼   ► But anyway, that's splitting here.

00:05:27 ◼   ► The point is there's hardware on the M1, just like there is on all of the iPhone and iPad system on a chips that is dedicated to doing specific tasks very well.

00:05:37 ◼   ► And on the Mac, it's like, OK, that's great.

00:05:41 ◼   ► The OS uses it for certain things, but otherwise it's just sitting there not being used until someone updates their software for it.

00:05:46 ◼   ► So this is an announcement that Apple, I think it's Apple's fork of TensorFlow, which is a machine learning thing of some stripe.

00:05:53 ◼   ► The new version optimized for the M1 base max has seven times faster performance.

00:05:59 ◼   ► And how does it get seven times faster performance?

00:06:02 ◼   ► It's not in this case because, you know, the M1 is faster in CPU than than any other Mac, but it's not seven times faster.

00:06:08 ◼   ► It's getting sometimes faster because it's using the machine learning hardware that's on the chip.

00:06:12 ◼   ► The previous version didn't because there was none right on the Intel chips.

00:06:15 ◼   ► There was nothing right.

00:06:16 ◼   ► And now it's using that machine learning stuff that's on there.

00:06:19 ◼   ► Huge speed boost.

00:06:20 ◼   ► So this is the type of sort of nonlinear increases you get for actually taking advantage of the hardware that they put on the system on chips for, you know, for phones, for iPads.

00:06:30 ◼   ► And now the Mac has them, too.

00:06:32 ◼   ► And, you know, the sooner Mac developers pick up that, the better it's going to make their applications look.

00:06:37 ◼   ► I think I don't remember if what's the what's the app that has like that super resolution plug in?

00:06:42 ◼   ► Is it Pixelmator?

00:06:44 ◼   ► Yeah. Yeah. I think that one is also now taking advantage of the neural engine and stuff like it is.

00:06:50 ◼   ► Yes.

00:06:50 ◼   ► Yeah. So those are and there was someone that a benchmark of those things are showing like, you know, if you do it on the CPU, like it's still it's still faster because the CPU is faster.

00:06:59 ◼   ► But if you don't do it on the CPU, instead of doing it on dedicated hardware, it's a lot faster.

00:07:03 ◼   ► So those things are great and probably more power efficient as well.

00:07:06 ◼   ► So that's exciting and not just like, oh, we recompiled it for our Macs.

00:07:11 ◼   ► We didn't just recompile it.

00:07:12 ◼   ► We wrote to use whatever new set of APIs targets the new hardware that makes it way faster.

00:07:16 ◼   ► That is extremely impressive.

00:07:19 ◼   ► Then tell me about 8K rendering, if you please.

00:07:22 ◼   ► I think I did my own, you know, McGurk effect, Green Needle, iPhone, Fortnite, whatever thing in the last episode.

00:07:31 ◼   ► During the section where we were talking about various people's impressive benchmarks of the M1 based Macs.

00:07:37 ◼   ► One of the examples was a 8K video render test.

00:07:43 ◼   ► And the thing was that, you know, the M1 based Mac rendered it and it was competing with the Intel version.

00:07:50 ◼   ► The Intel version couldn't do it without draining its battery.

00:07:53 ◼   ► Right. And what was written in the notes and what I could swear I tried to say twice on the podcast was the percentage that the M1 based Mac depleted its battery.

00:08:04 ◼   ► And I listened back to the show and I realized twice I said what sounded like the M1 Mac used 70 percent of its battery.

00:08:11 ◼   ► And the Intel Mac drained its battery before completing the task.

00:08:14 ◼   ► But what I was trying to say was the M1 Mac depleted 17 percent of its battery.

00:08:22 ◼   ► Two more than 15.

00:08:24 ◼   ► The number after 16.

00:08:26 ◼   ► So the M1 based Mac not only did it faster, but it depleted only 17 percent of its battery, having 83 percent remaining.

00:08:35 ◼   ► And the non M1 based Mac, the Intel based Mac could not complete the task on a single charge.

00:08:41 ◼   ► So I just wanted to clarify, my mouth said what sounded like 70, but it is not.

00:08:47 ◼   ► It was 17.

00:08:48 ◼   ► That is absolutely bananas.

00:08:50 ◼   ► We're going to talk more about these new Macs and I don't want to.

00:08:55 ◼   ► I might quit the show because I like, I really like my MacBook Pro that I have from June of this year.

00:09:02 ◼   ► It doesn't feel like it's a 40 year old computer, but the way everyone is talking about this M1 Macs, I feel like this is barely better than the 8088 I had in my bed.

00:09:11 ◼   ► Room when I was like 10 years old, when it was already, you know, a way too old computer.

00:09:14 ◼   ► My word.

00:09:15 ◼   ► All right.

00:09:16 ◼   ► Tell me about running iOS apps with just the IPA files, please.

00:09:20 ◼   ► Yeah.

00:09:20 ◼   ► And we talked about this when Apple made the announcement that like, if you have an iPad or if you have an iPhone app, you have a choice of whether you want that iPhone app to be available for people to download on their Macs through the quote unquote Mac App Store.

00:09:36 ◼   ► Right.

00:09:37 ◼   ► And you can say, yes, opt in and no to opt out and every developer basically has to make this choice.

00:09:41 ◼   ► But what if you want to run an iPhone app, but the developer of that iPhone app has decided not to make it available to Mac users.

00:09:49 ◼   ► For example, Marco could have decided, oh, overcast.

00:09:51 ◼   ► I'm going to make a Mac version, so I don't want you downloading using the iOS version.

00:09:54 ◼   ► But he didn't.

00:09:55 ◼   ► So you can get overcast on your Mac.

00:09:57 ◼   ► Right.

00:09:57 ◼   ► But another developer may not make that choice.

00:10:00 ◼   ► Well, many people through experimentation have determined if you can just use I don't know what they're using iTunes or whatever, whatever way you can get to download the dot IPA file onto your Mac.

00:10:12 ◼   ► I got IPA files.

00:10:13 ◼   ► What does it stand for?

00:10:13 ◼   ► I phone package something, something.

00:10:16 ◼   ► It's a renamed zip file.

00:10:18 ◼   ► It's one of many file formats that is just renamed.

00:10:20 ◼   ► Yeah.

00:10:21 ◼   ► Anyway, it's basically how we package up iOS applications for distribution.

00:10:24 ◼   ► Probably iPhone app.

00:10:26 ◼   ► Yeah, maybe.

00:10:27 ◼   ► But anyway, if you can get one of those.

00:10:30 ◼   ► As in the legit downloaded one from like the app store, the real app store.

00:10:34 ◼   ► And you can like, you know, get it through iTunes.

00:10:36 ◼   ► But when iTunes would download stuff, you can do it in a local backup.

00:10:38 ◼   ► But I would get a legit downloaded dot IPA for your iPhone applications.

00:10:43 ◼   ► Lots of different ways to get that on your Mac and double click it on an ARM based Mac.

00:10:47 ◼   ► It'll run because it's not.

00:10:50 ◼   ► It's valid.

00:10:50 ◼   ► It's signed.

00:10:51 ◼   ► It's, you know, you have downloaded through your Apple ID.

00:10:54 ◼   ► It will do the fair play check and everything.

00:10:56 ◼   ► You know, it's not like you're hacking anything.

00:10:59 ◼   ► If you can get that IPA, which you can, you can just run it on your Mac.

00:11:03 ◼   ► Apparently now it may be broken.

00:11:04 ◼   ► There may be a reason the developer chose not to distribute it on the Mac.

00:11:07 ◼   ► And I'm not sure how sustainable this is and when Apple will close this door.

00:11:10 ◼   ► But if there's some app that you want to use that you can't right now, try it.

00:11:14 ◼   ► The person who wrote this into Mac rumors named Amy said she's used this method to

00:11:18 ◼   ► install Netflix, Hulu, YouTube, and Spotify, all of which at the time of writing

00:11:22 ◼   ► were not available on the Mac app store.

00:11:24 ◼   ► But you can get the IPAs and you can run them on your Mac.

00:11:26 ◼   ► That's very cool.

00:11:27 ◼   ► By the way, real time, real time follow up.

00:11:29 ◼   ► IOS app store package.

00:11:31 ◼   ► So I and IOS A and app store P.

00:11:33 ◼   ► I don't know if you can think of that A stand for app store.

00:11:36 ◼   ► App store is two words.

00:11:37 ◼   ► Wikipedia says it, so it must be true.

00:11:39 ◼   ► It's like the ISO.

00:11:39 ◼   ► Yeah.

00:11:40 ◼   ► What does ISO stand for?

00:11:41 ◼   ► It's something French, but it's like the international standards of organization.

00:11:45 ◼   ► It's like, I don't know.

00:11:45 ◼   ► Oh, no, no.

00:11:46 ◼   ► I thought you meant, I thought you meant the like CD-ROM archive.

00:11:49 ◼   ► Technically it's the same thing.

00:11:51 ◼   ► It's the same thing.

00:11:52 ◼   ► It's the same company.

00:11:53 ◼   ► Yeah.

00:11:53 ◼   ► They made the same stand.

00:11:54 ◼   ► Yeah.

00:11:54 ◼   ► I mean, it's a little bit of a walk to get there, but yes.

00:11:56 ◼   ► Wow.

00:11:57 ◼   ► Well, today I learned.

00:11:59 ◼   ► All right.

00:11:59 ◼   ► Tell me about crossover.

00:12:00 ◼   ► So this is something that I was not aware of, but apparently this is like a, well,

00:12:05 ◼   ► I'm going to sell it short, but kind of like a white labeled version of wine.

00:12:09 ◼   ► What is going on here?

00:12:10 ◼   ► So for those who are not familiar, wine is a thing that I used to use when I told

00:12:16 ◼   ► myself that running Linux full time was a worthwhile endeavor.

00:12:18 ◼   ► It wasn't.

00:12:19 ◼   ► We all had a Linux phase.

00:12:20 ◼   ► Yep.

00:12:21 ◼   ► Title.

00:12:22 ◼   ► So wine, last I remember paying attention to it was basically.

00:12:28 ◼   ► A shim on top of your local systems, like os APIs that acted like windows APIs.

00:12:37 ◼   ► So if you wanted to run an app that was built for windows, it would start talking

00:12:41 ◼   ► to wine, w I N E, which I think is a background name for something.

00:12:45 ◼   ► Wine is not an emulator, I believe.

00:12:47 ◼   ► Yeah, I think you're right.

00:12:48 ◼   ► Um, sorry.

00:12:49 ◼   ► So we talked to wine and wine would kind of like translate these API calls into

00:12:54 ◼   ► something that your system like Mac, Mac, iOS, or Linux or to have you would

00:12:57 ◼   ► understand.

00:12:58 ◼   ► And I guess crossover is a four pay, uh, like fancier version of wine.

00:13:04 ◼   ► That's presumably a lot less fiddly.

00:13:06 ◼   ► Is that fair to say?

00:13:07 ◼   ► I'm not sure which one of you put this in the show notes.

00:13:08 ◼   ► So I'm reaching out to the two of you.

00:13:10 ◼   ► Yeah.

00:13:10 ◼   ► Yeah.

00:13:10 ◼   ► That's basically what crossover is.

00:13:12 ◼   ► And one of the main uses of it has been to run windows based games on your Mac

00:13:16 ◼   ► without having to boot into windows.

00:13:18 ◼   ► Obviously on an Intel based Mac, you can use VMware or parallels.

00:13:21 ◼   ► To run a full windows OS.

00:13:23 ◼   ► And then within that run games, you can also reboot with bootcamp to boot your

00:13:26 ◼   ► Intel based Mac into windows and run your games there for best performance.

00:13:29 ◼   ► But crossover is like wine.

00:13:30 ◼   ► It's like, what if I don't want to run a VM?

00:13:32 ◼   ► I don't want to run a full windows operating system.

00:13:33 ◼   ► I don't want to reboot my Mac, but I do want to run a windows game.

00:13:36 ◼   ► Well, like wine, you can launch the windows game in the crossover environment

00:13:40 ◼   ► and it will basically take all your win 32 calls or whatever and say, okay, well,

00:13:43 ◼   ► uh, you know, this isn't windows, but I understand what you're trying to do

00:13:47 ◼   ► windows application.

00:13:48 ◼   ► So I will make the roughly equivalent set of Mac OS or Unix calls to do the thing

00:13:54 ◼   ► that you want and get the job done.

00:13:56 ◼   ► Um, so that's what crossover is.

00:13:57 ◼   ► We're like, well, that doesn't have any relevance on M1 based Macs cause M1

00:14:00 ◼   ► based Macs, you know, they're, they're not Intel, right?

00:14:03 ◼   ► And this is all about running windows things and windows only runs on Intel,

00:14:06 ◼   ► except for the weird arm version of the Microsoft one license to anybody yet.

00:14:09 ◼   ► But anyway, uh, what if I want to run windows games on my M1 Mac?

00:14:12 ◼   ► You might think you're just SOL because you can't run power parallels.

00:14:15 ◼   ► You can't run, you can't run windows and parallel as a VMware and you can't

00:14:18 ◼   ► reboot your arm based Mac into windows yet.

00:14:20 ◼   ► So what are you going to do?

00:14:21 ◼   ► Oh, and by the way, as we know, like big Sur doesn't,

00:14:26 ◼   ► uh, is 64 bit only.

00:14:27 ◼   ► And there has never been a 32 bit processor in a Mac in ages, right?

00:14:32 ◼   ► So how are you going to run? And most windows games are still 32 bit.

00:14:35 ◼   ► How are you going to run a 32 bit windows game? Well,

00:14:37 ◼   ► apparently through the magic of crossover,

00:14:39 ◼   ► you're able to take a 32 bit windows game written for Intel processors and run it

00:14:44 ◼   ► on a 64 bit arm, uh, Mac through the magic of translation. So here's,

00:14:49 ◼   ► here's Jeremy White from a crossover. I think he's one of his, yeah,

00:14:53 ◼   ► he works at the company that makes his code weavers. There's so much emulation

00:14:56 ◼   ► going on under the covers. Imagine a 32 bit windows,

00:14:59 ◼   ► Intel binary running on a 32 to 62 bridge and wine slash crossover on top of

00:15:03 ◼   ► Mac iOS on arm CPU that is emulating X 86 and it works. This is just so cool.

00:15:08 ◼   ► And Brendan Shanks, one of the developers says Mac OS 10.15,

00:15:13 ◼   ► remove 32 bit executable support,

00:15:15 ◼   ► but added support for 64 bit apps to create 32 bit code segments.

00:15:19 ◼   ► Crossover uses that on Intel and Rosetta emulates it,

00:15:22 ◼   ► which is how you can do it on the Apple Silicon.

00:15:24 ◼   ► So I didn't know that that you're able to like in a 64 bit app,

00:15:28 ◼   ► create a 32 bit code segment and then apparently execute that. And so that,

00:15:32 ◼   ► that ability on Intel CPU's plus Mac OS 10.15 and later is being correctly

00:15:38 ◼   ► emulated by Rosetta on arm. And you would think, okay, well this is, you know,

00:15:42 ◼   ► it's like one of those things where you run VMware inside VMware inside VMware.

00:15:45 ◼   ► It's cool and it's funny, but that's silly. It's not actually practical as it,

00:15:49 ◼   ► they show playing actual games like you get acceptable gaming performance on

00:15:53 ◼   ► certain varieties of games. It's not incredibly,

00:15:56 ◼   ► it's not as slow as you would think it would be.

00:15:57 ◼   ► It is actually viable for playing some type of games, which is amazing.

00:16:01 ◼   ► We are sponsored this week by HelloFresh.

00:16:05 ◼   ► HelloFresh offers convenient,

00:16:07 ◼   ► easy and stress free meal kit delivery service.

00:16:11 ◼   ► They have convenient no contact delivery to your doorstep for easy home cooking

00:16:16 ◼   ► with your family. HelloFresh's recipes are easy to follow with simple steps and

00:16:20 ◼   ► pictures to guide you along the way.

00:16:22 ◼   ► I've personally used HelloFresh before and I love their recipe cards.

00:16:27 ◼   ► They really are very easy to follow.

00:16:28 ◼   ► They show you what things are supposed to look like at each step so you know

00:16:32 ◼   ► like how small do I chop this thing or what does it look like when it's cooking?

00:16:35 ◼   ► Like, you know, all those things that they show you.

00:16:37 ◼   ► HelloFresh cuts out stressful meal planning grocery store trips so you can enjoy

00:16:41 ◼   ► cooking and get dinner on the table in about 30 minutes.

00:16:45 ◼   ► And this comes at a great value.

00:16:46 ◼   ► You can save up to 40% when you use HelloFresh versus shopping at the grocery

00:16:50 ◼   ► store and feeding your whole family has never been easier with new lower prices

00:16:54 ◼   ► for larger box sizes. So more servings means more savings.

00:16:58 ◼   ► And this is a huge variety of delicious, nutritious food.

00:17:02 ◼   ► HelloFresh delivers fresh, high quality pre-portioned ingredients.

00:17:05 ◼   ► You can make meals with delicious and nutritious.

00:17:08 ◼   ► Over 90% of their ingredients are sourced directly from growers to ensure peak

00:17:12 ◼   ► flavor and ripeness and they offer more than 20 chef crafted delicious options

00:17:16 ◼   ► every week to help you break out of your recipe rut,

00:17:19 ◼   ► try new things and make any night feel special.

00:17:22 ◼   ► There's something everybody will enjoy, including 20 minute meals,

00:17:25 ◼   ► low calorie plans, vegetarian meals, kid approved recipes, and so much more.

00:17:30 ◼   ► And HelloFresh is flexible for your lifestyle.

00:17:33 ◼   ► You can easily change your delivery days or meal plan preferences,

00:17:36 ◼   ► skip a week whenever you need to write in their easy to use app and you can keep

00:17:40 ◼   ► your fridge stocked by adding extra meals or additional proteins,

00:17:43 ◼   ► quick meals like breakfast on the go or their 10 minute lunches and even desserts

00:17:46 ◼   ► to satisfy that sweet tooth.

00:17:48 ◼   ► Go to hellofresh.com/ATP90 and use code ATP90 to get $90 off,

00:17:53 ◼   ► including free shipping. That's hellofresh.com/ATP90,

00:17:59 ◼   ► code ATP90 to get $90 off, including free shipping.

00:18:03 ◼   ► Thank you so much to HelloFresh for sponsoring our show.

00:18:09 ◼   ► This is not something that I've dabbled with yet,

00:18:11 ◼   ► but I know a lot of people are very happy about this forthcoming change.

00:18:14 ◼   ► iOS 14.3 Beta 2 no longer opens the shortcuts app when launching an app from a

00:18:19 ◼   ► custom icon. So to back up a half step here, using shortcuts,

00:18:24 ◼   ► you can put a quote unquote app on your home screen with a custom icon that the

00:18:29 ◼   ► shortcut, all it does is launch the app.

00:18:33 ◼   ► And so people are using this to be able to make their own icons for existing

00:18:38 ◼   ► apps, both system and otherwise. And this is, you know,

00:18:41 ◼   ► part of the customization craze that, you know,

00:18:43 ◼   ► our dear friend David Smith has been riding with widgetsmith,

00:18:46 ◼   ► which you should check out if you haven't already.

00:18:48 ◼   ► So the way it used to work was, or way officially it still works,

00:18:52 ◼   ► is that when you launch one of these shortcuts from your home screen,

00:18:55 ◼   ► it will launch the shortcuts app for a flash and then punch you into this actual

00:19:00 ◼   ► thing you're trying to do. And apparently in Beta 2,

00:19:03 ◼   ► that's no longer the case. It will just immediately open the app.

00:19:05 ◼   ► And I believe it does one of the little banners at the top of the screen as well,

00:19:08 ◼   ► but it's, there's no like flash into the shortcuts app first,

00:19:12 ◼   ► which is excellent.

00:19:13 ◼   ► This makes me think about sort of the corporate mechanics inside Apple,

00:19:18 ◼   ► right? Because we know that, you know, where do shortcuts come from?

00:19:21 ◼   ► It was the workflow team that Apple sort of aqua hired and then they made

00:19:24 ◼   ► shortcuts for iOS. Okay. And you know,

00:19:27 ◼   ► especially given how, you know,

00:19:30 ◼   ► it wasn't very long ago when these were independent developers.

00:19:33 ◼   ► Now they're inside Apple. They see everyone doing this customization craze,

00:19:37 ◼   ► where they want to make custom icons for the applications.

00:19:39 ◼   ► And they do it by making a trivial shortcut and all it does is launch the apps

00:19:42 ◼   ► and it goes through the shortcut app and people use it anyway,

00:19:45 ◼   ► just because it's a cool thing to do, but it would be better if it didn't.

00:19:48 ◼   ► You know, that workflow development, the development team sees that and says,

00:19:51 ◼   ► Oh, well we can do that. Like we work inside Apple now. And we, you know,

00:19:55 ◼   ► we're the, we're the shortcuts team. It's pretty easy for us to say, Hey,

00:19:58 ◼   ► if you make a shortcut that just launches an app, don't bounce through shortcuts,

00:20:02 ◼   ► just launched the app directly. And so the implements didn't do it.

00:20:04 ◼   ► And this is a beta, so it hasn't shipped yet, but you know,

00:20:06 ◼   ► I can see that sort of feedback loop being simple and closed, right?

00:20:11 ◼   ► But on the other hand,

00:20:12 ◼   ► there are larger strategic implications of allowing

00:20:17 ◼   ► customers this level of customization, which based on our discussion of,

00:20:22 ◼   ► you know, the whole widgets thing back when we were talking about iOS 14,

00:20:25 ◼   ► we're all in favor of,

00:20:26 ◼   ► and we think it's a great idea to let people customize in this way. But Oh,

00:20:30 ◼   ► I have felt like that Apple accidentally backed into this by accidentally

00:20:35 ◼   ► providing a critical mass of customizability combined with, you know, apps like,

00:20:39 ◼   ► uh, underscores,

00:20:40 ◼   ► widgets Smith that a lot suddenly allowed people to express themselves in a

00:20:44 ◼   ► holistic way and theme there's, you know, home screens.

00:20:46 ◼   ► It doesn't look like anything that Apple planned and pitched and

00:20:51 ◼   ► promoted. They just kind of got it accidentally.

00:20:53 ◼   ► So this type of small changes,

00:20:55 ◼   ► the shortcuts team does an obvious thing based on customer feedback. Hey,

00:20:58 ◼   ► this is the thing people want and we can do it pretty easily.

00:21:00 ◼   ► So here you go is kind of a tacit endorsement of this

00:21:05 ◼   ► steaming craze without,

00:21:07 ◼   ► I feel like buy in from the larger org because the real full buy in way to do it

00:21:12 ◼   ► would be, Oh, now on the setting screen,

00:21:14 ◼   ► there's a place where you can customize every app icon or now on the setting

00:21:17 ◼   ► screen,

00:21:17 ◼   ► there's a themes section where you can download a theme packs from the theme

00:21:21 ◼   ► store. Like we know how I full throated sort of endorsement of this thing would

00:21:25 ◼   ► look from Apple, right? I mean,

00:21:28 ◼   ► how they did the iMessage store and I'll use that like a theme store or a watch

00:21:31 ◼   ► face store. Like we like, that's sort of the whole Apple buys in.

00:21:36 ◼   ► This is the shortcut team is filled with developers who want to do the right

00:21:39 ◼   ► thing. And so they have,

00:21:41 ◼   ► so I'm still watching this space to see how much Apple is willing to

00:21:46 ◼   ► embrace theming and customization on iOS.

00:21:49 ◼   ► Because I can tell you on the Mac, aside from free, your accent colors,

00:21:54 ◼   ► Apple has not fully endorsed theming and as mostly accidentally

00:21:59 ◼   ► shut down every application on Mac OS 10,

00:22:03 ◼   ► anyway,

00:22:03 ◼   ► that has allowed some kind of theming from candy bar to shape shifter to all

00:22:07 ◼   ► that other good stuff. It's not that they had it out for them.

00:22:10 ◼   ► It's just that they never considered that an important use case.

00:22:12 ◼   ► And they slowly semi accidentally semi on purpose broke it all until,

00:22:16 ◼   ► until where we are today, where no one really tries to do anything.

00:22:19 ◼   ► But iOS is going the other direction. So I'm still watching this space.

00:22:24 ◼   ► Yeah. Didn't wasn't there, I don't know,

00:22:26 ◼   ► like an iOS release or two ago a way that I think shortcuts could set like your

00:22:31 ◼   ► lock screen or your, your, your wallpaper on your home screen.

00:22:34 ◼   ► And then that like got pulled and then it came back and then it got pulled.

00:22:37 ◼   ► And from the outside,

00:22:38 ◼   ► it almost smelled like there was a cat and mouse going on between the, you know,

00:22:43 ◼   ► former the shortcuts team. What was it called? Workflow.

00:22:48 ◼   ► God, I just had a total blank, sorry.

00:22:50 ◼   ► The former workflow now shortcuts team and the rest of the iOS team at Apple

00:22:55 ◼   ► where, you know, a shortcut slash workflow was like, Oh yeah,

00:22:57 ◼   ► we can totally make it so you can change your home screen and your lock screen.

00:23:00 ◼   ► And then the rest of the iOS group was like, huh? What? No, no,

00:23:05 ◼   ► no, no, no. And it just went back and forth a few times.

00:23:08 ◼   ► I think it actually just came back recently, but yeah,

00:23:10 ◼   ► it is interesting to see a group of developers that,

00:23:14 ◼   ► that very much believed in the, you know,

00:23:16 ◼   ► go fast and break things mantra and how they're dealing with the large, you know,

00:23:21 ◼   ► aircraft carrier that is Apple.

00:23:23 ◼   ► I don't think they're going fast and breaking things.

00:23:25 ◼   ► They're just being as responsive as they would be as an independent company,

00:23:28 ◼   ► which is more responsive than Apple is. And by the way,

00:23:31 ◼   ► we're talking about teams being responsive. I will once again,

00:23:33 ◼   ► make my pitch to anybody. Apple who is listening.

00:23:35 ◼   ► There is a level of customization that iOS has allowed basically from day one,

00:23:39 ◼   ► which is the ability to rearrange icons in your home screen.

00:23:41 ◼   ► That interface continues to be a thorn in the world side. Please, please,

00:23:46 ◼   ► somebody, because third party developers can't do this.

00:23:48 ◼   ► If you let third party developers do it, it would be solved already,

00:23:50 ◼   ► but you haven't, you're keeping it for yourself.

00:23:52 ◼   ► So if you're going to keep it for yourself,

00:23:53 ◼   ► please give us a way to rearrange home screens and icons in a way that does

00:23:58 ◼   ► not drive us all up a wall and make us tear our hair out because the current

00:24:01 ◼   ► system is really, really bad.

00:24:03 ◼   ► And my home screens are a mess and I want to fix them and I can't please,

00:24:07 ◼   ► please Apple do something about this. Yeah.

00:24:10 ◼   ► I would like Apple to treat the home screen

00:24:15 ◼   ► arrangement as more valuable data than they appear to treat it right now.

00:24:20 ◼   ► Because right now I like, so you know, when I was going through, you know,

00:24:25 ◼   ► COVID last two weeks,

00:24:26 ◼   ► I decided I should add to my home screen a widget that can display the blood

00:24:31 ◼   ► oxygen level from my Apple watch, whatever the last measurement was.

00:24:35 ◼   ► I want that on a widget on my home screen so I can see it all the time.

00:24:37 ◼   ► Well, that's a good idea. Yeah, thanks. I thought so too.

00:24:39 ◼   ► And I did eventually find something that could do that. Um,

00:24:42 ◼   ► let me plug it real fast. It's called health view.

00:24:45 ◼   ► And it just displays like, you know, anything from health kit in a widget.

00:24:47 ◼   ► And when I first, like, I don't know anybody, myself included,

00:24:51 ◼   ► who has placed a widget on their home screen successfully without totally

00:24:57 ◼   ► screwing up a million other icons in the process. Yeah.

00:25:01 ◼   ► It's like setting off a bomb on your own screen. Yeah. And there is,

00:25:04 ◼   ► not only is there no undo,

00:25:06 ◼   ► but even if you just take whatever widget you had just bombed your home screen

00:25:10 ◼   ► with and just move it to a different page,

00:25:12 ◼   ► the icons don't go back to the way they were.

00:25:15 ◼   ► They go back to some other like shuffled random order. It's really weird.

00:25:19 ◼   ► It's literally like a bomb. It's like if you,

00:25:20 ◼   ► if you set off a bomb and then you, then you remove the bomb,

00:25:23 ◼   ► the bricks don't go back to where they are. They stay scattered to the four winds.

00:25:27 ◼   ► Yeah. Like they just collapsed back inward in random orders. Like I,

00:25:30 ◼   ► it's very strange. Like I had,

00:25:32 ◼   ► I had to look back at a screenshot that I had taken of my home screen a few days

00:25:36 ◼   ► earlier for other reasons. Like I had to look for that to know,

00:25:39 ◼   ► how did I have these icons arranged again?

00:25:41 ◼   ► That's the main reason I take screenshots of my home screen.

00:25:44 ◼   ► By the way, sometimes people say, Oh, show us your old home screen.

00:25:47 ◼   ► The only reason I have those is because they always take one if I'm smart before

00:25:51 ◼   ► I start trying to rearrange because I know it's going to be chaos. Yeah.

00:25:54 ◼   ► And it just, it seems like, it seems like maybe,

00:25:57 ◼   ► maybe whoever works on springboard at Apple just by the nature of working on

00:26:02 ◼   ► springboard, maybe they have such constant like, you know,

00:26:06 ◼   ► disturbances in their icon arrangements as part of that job that they just don't

00:26:10 ◼   ► realize how incredibly disruptive it is when it happens to most people.

00:26:15 ◼   ► Cause like, I know I'll keep the same icon arrangement for a lot of stuff,

00:26:18 ◼   ► like on my main home screen, I'll keep it for years.

00:26:20 ◼   ► And when that gets disrupted, it like you,

00:26:24 ◼   ► you can't find things where they are cause you,

00:26:26 ◼   ► you're so used to the spatial memory of where things were. And so like,

00:26:29 ◼   ► it's such a incredibly disruptive thing to have that be,

00:26:33 ◼   ► be messed up in a big way. And I don't know how you can possibly install a

00:26:38 ◼   ► widget on any home screen and not do that.

00:26:41 ◼   ► It's I do hope that Apple realizes how,

00:26:46 ◼   ► how bad this is sometime and actually improves that.

00:26:49 ◼   ► To be clear, the app library stuff and being able to hide home screens,

00:26:52 ◼   ► those are all good features. We're not saying get rid of those.

00:26:55 ◼   ► You have to have those too.

00:26:57 ◼   ► And also for the home screens that you want to actually have,

00:27:00 ◼   ► you have to be able to arrange things in a sane way.

00:27:02 ◼   ► I've been trying new techniques cause I've had to,

00:27:04 ◼   ► I've been trying to work on my later home screens cause they're like,

00:27:07 ◼   ► I said, my home screens are a mess and no technique I can find can get me

00:27:11 ◼   ► sanity. Like I've been,

00:27:13 ◼   ► I think the behavior has changed slightly recently because of what I've been

00:27:16 ◼   ► doing.

00:27:17 ◼   ► My main technique that I've used is the grab and hold and inflict with the other

00:27:19 ◼   ► finger to scroll. So it's like a two handed technique rather than pushing to the

00:27:23 ◼   ► screen edge, which is, you know, it's madness.

00:27:26 ◼   ► You'll never be successful that way. Right?

00:27:27 ◼   ► So you just grab one or more icons and then you swipe with the other finger

00:27:30 ◼   ► quickly.

00:27:31 ◼   ► The problem is no matter how fast you swipe momentarily,

00:27:35 ◼   ► iOS thinks you're on one of the other home screens and when you're on there

00:27:39 ◼   ► momentarily,

00:27:40 ◼   ► it tends to eject other icons like subatomic particles in a particle collider.

00:27:45 ◼   ► It ejects them outwards. Right. And you're like,

00:27:48 ◼   ► but I'm just passing through. I'm on my way to page five. Right. I don't,

00:27:52 ◼   ► I didn't want to disturb anything on page three, but I have.

00:27:55 ◼   ► And the new thing that I think in the last two iOS versions or whatever is I'll

00:28:00 ◼   ► get to page five, I'll find the place for it that I've, you know,

00:28:03 ◼   ► it's like preparing the way with my fridge again.

00:28:04 ◼   ► I will have prepared a spot for it. I'll get to page five,

00:28:07 ◼   ► I'll land the icons if I'm lucky.

00:28:08 ◼   ► Sometimes I'm unlucky and they just go somewhere and have no idea where. Um,

00:28:12 ◼   ► but then I'll go back and I'll swipe backwards and I'll see that icons were

00:28:16 ◼   ► ejected and somehow spawned a new page just for the ejected icons between what

00:28:21 ◼   ► was page three and page four.

00:28:23 ◼   ► And now there'll be a page with one or two icons on it. I'm like,

00:28:26 ◼   ► where did you come from? And I'll look to either side.

00:28:29 ◼   ► And sometimes either side is filled with icons. Like how is that possible?

00:28:32 ◼   ► Where did these come from? And how did you make a new page for yourself?

00:28:35 ◼   ► I was just passing through and I thought I successfully landed those two icons

00:28:39 ◼   ► on page five, but now I have an extra page with two icons on it. It's the worst.

00:28:43 ◼   ► It's getting worse. Like it's before it used to be, Oh, well like, you know,

00:28:47 ◼   ► I keep going from one edge to the other and it squirms out of the way and yada,

00:28:49 ◼   ► yada, but at least it was somewhat sensible.

00:28:51 ◼   ► Like there was just a big linear list that like a big snake and you just push

00:28:53 ◼   ► things, but now they spawn their own pages.

00:28:56 ◼   ► This like we need an interface to do this. That is not that a,

00:29:01 ◼   ► is undoable slash like sort of rearrange and then commit, right?

00:29:05 ◼   ► And then B provides a richer interface than our

00:29:11 ◼   ► fingers on the phone screen. If we could do it on a Mac,

00:29:14 ◼   ► if we could do it on an iPad,

00:29:15 ◼   ► or if we could do it on a phone and miniaturize size in a sort of trial mode

00:29:19 ◼   ► where you're just in this app that lets you rearrange home screens.

00:29:21 ◼   ► And only when you get it all the way you want,

00:29:23 ◼   ► do you say commit or you just bail and say, well,

00:29:25 ◼   ► nevermind or you leave it in progress or whatever. But like, like I said,

00:29:28 ◼   ► if this, if there was an API for doing this,

00:29:30 ◼   ► which I'm not really asking for this is a tall order for an API.

00:29:33 ◼   ► All I'm saying if there was a third party API for home screen management,

00:29:36 ◼   ► this problem will be solved a hundred ways because it's, you know,

00:29:39 ◼   ► third parties would make money doing it because there's a need and people want

00:29:42 ◼   ► to do it. And we're just mythos proof that there's an appetite for this type of

00:29:45 ◼   ► thing, but there's no third party API. Fine. It makes sense.

00:29:47 ◼   ► It should be a first party thing. Apple just do a better job at it.

00:29:50 ◼   ► That's all we're asking for.

00:29:52 ◼   ► You know, it's funny because just a couple of days ago I was talking to my dad

00:29:56 ◼   ► and he said, you know, can you still arrange your home screen and iTunes?

00:30:00 ◼   ► And I was like, well, it's not iTunes anymore. It's music. And no,

00:30:03 ◼   ► I don't think so. And you can't do it in the finder either. I don't believe.

00:30:06 ◼   ► In fact, I don't think that's been a thing for years.

00:30:08 ◼   ► You can do it in the Apple configurator too.

00:30:10 ◼   ► Supposedly I've tried took a run at it a few times and I this,

00:30:14 ◼   ► I think there's a way to do it,

00:30:16 ◼   ► but I think you kind of have to pretend that you're managing a bunch of people's

00:30:19 ◼   ► corporate iOS devices or something like it's, it was scary enough that I bailed

00:30:23 ◼   ► out every time I took a run at it. But before people send us his feedback,

00:30:27 ◼   ► apparently there is a way to use Apple configurator to, to somehow do this.

00:30:30 ◼   ► And Apple configurator too is an app that you run on your Mac.

00:30:32 ◼   ► So maybe there's a better, easier way to do it.

00:30:35 ◼   ► But iTunes was the best solution we've had so far.

00:30:38 ◼   ► And even that wasn't great because you would think that, Oh,

00:30:41 ◼   ► you've got a Mac screen and this huge, you know, huge iTunes window. Surely,

00:30:45 ◼   ► you can make this easy with a mouse cursor or whatever.

00:30:48 ◼   ► It was still weird and hard. It wasn't a good interface. If again,

00:30:51 ◼   ► if this was a third party API, there would the third part,

00:30:54 ◼   ► many people would make third party apps and the good ones would float to the top.

00:30:57 ◼   ► Instead, we just got the one implementation iTunes,

00:30:59 ◼   ► which was better than nothing.

00:31:00 ◼   ► Like that's why your dad was asking for it because it was better than just

00:31:03 ◼   ► trying to do with your finger, but it still wasn't good.

00:31:05 ◼   ► Then finally, one very quick piece of followup.

00:31:07 ◼   ► I was lamenting last week about GitHub's relationship with the U S immigration

00:31:12 ◼   ► and customs enforcement.

00:31:14 ◼   ► And it was pointed out to me that that isn't a direct relationship.

00:31:18 ◼   ► So what happened was a GitHub, or I'm sorry,

00:31:21 ◼   ► a third party reseller sold an on premises GitHub enterprise server,

00:31:26 ◼   ► like instance contract, whatever on behalf of GitHub.

00:31:30 ◼   ► But GitHub like didn't seek this business themselves.

00:31:32 ◼   ► There is a third party in between the two.

00:31:34 ◼   ► It doesn't make me feel that much better about it personally,

00:31:37 ◼   ► but you may have a different interpretation of it. And I will put it.

00:31:39 ◼   ► It's like money laundering for morals.

00:31:41 ◼   ► So I'll put a link in the show notes to their response about this,

00:31:46 ◼   ► which was from a late last, actually almost exactly a year ago.

00:31:50 ◼   ► So you can check it out if you're interested.

00:31:52 ◼   ► Also, I wanted to give a quick shout out to this wonderful. So, okay.

00:31:56 ◼   ► Earlier this evening, uh,

00:31:58 ◼   ► I had discovered throughout all of my new device setups that for the first time

00:32:03 ◼   ► ever, I had a sync problem with my contacts database.

00:32:06 ◼   ► For the first time ever,

00:32:08 ◼   ► I all of a sudden had duplicates of every single contact in my address book.

00:32:13 ◼   ► And I had heard Merlin talk about this before and back to work.

00:32:17 ◼   ► So I knew there were like apps out there that would help you find and clean up

00:32:20 ◼   ► duplicates. So I asked you in a Slack group, um, you know,

00:32:24 ◼   ► what these apps were called and what a good one was. And somebody who,

00:32:27 ◼   ► who knows a bit about Mac OS named, uh, John Surakusa

00:32:32 ◼   ► said,

00:32:33 ◼   ► did you try the menu command in the contact app to de duplicate?

00:32:38 ◼   ► I didn't even know this was here. John, when was this added?

00:32:41 ◼   ► It's been there for a long time. I asked that timidly,

00:32:44 ◼   ► cause a lot of people have contact problems and sometimes that D dupe command

00:32:49 ◼   ► doesn't do the job. And the reason they're asking is, Hey,

00:32:52 ◼   ► I tried the builtin D dupe command and it couldn't, it couldn't do what I wanted.

00:32:55 ◼   ► So now give me the third party like power app or whatever.

00:32:57 ◼   ► And I was hoping that maybe you didn't even know that was in there, but yeah,

00:33:00 ◼   ► it's been in there for awhile. I've used it a lot.

00:33:03 ◼   ► It's pretty good. Um, if you have a straight up duplication,

00:33:07 ◼   ► it will handle it. If instead you've done,

00:33:10 ◼   ► like I've done several times where you import a bunch of V cards that are from

00:33:13 ◼   ► an old backup or something and it really takes a little bit of doing,

00:33:16 ◼   ► it's a good first step. Um, I think it's been in there for many, many years,

00:33:21 ◼   ► but I couldn't pin it down to a particular release.

00:33:22 ◼   ► So this is under the card menu. So in the contacts app card menu,

00:33:26 ◼   ► look for duplicates. And it also had a feature,

00:33:29 ◼   ► which is a great little thing that I frequently need where if it has a,

00:33:35 ◼   ► if it has two names, you can have it optionally like merge the info.

00:33:39 ◼   ► Whenever it sees the same name in two different records.

00:33:41 ◼   ► And this happens all the time because of a relatively terrible design decision

00:33:46 ◼   ► in the messages app where you, and I think,

00:33:50 ◼   ► I think you see this a few other places like phone and stuff where you,

00:33:54 ◼   ► you get something from a new number and it gives you the option,

00:33:57 ◼   ► create a new contact or add to existing contact.

00:34:01 ◼   ► And the problem is, does anybody else face this problem?

00:34:04 ◼   ► I face this constantly. I never know. Do I have a contact for this person yet?

00:34:09 ◼   ► And so what I want is a name search box that I can type in their name and see

00:34:14 ◼   ► whether I have a contact for them yet.

00:34:18 ◼   ► And then if I do add this number to it and if I don't create a new one and the

00:34:23 ◼   ► interface doesn't let you do that, it just create new or add to existing.

00:34:27 ◼   ► And if you change your mind,

00:34:29 ◼   ► you got to like cancel out of that screen and then like rebuild the entire

00:34:32 ◼   ► sequence of taps or commands that got you there.

00:34:34 ◼   ► So I frequently would have like, you know,

00:34:38 ◼   ► two contacts with the same name that are the same person,

00:34:41 ◼   ► but one of them has an email address and one of them has a phone number.

00:34:43 ◼   ► And this command also fixes that. And I had some good luck with that as well.

00:34:48 ◼   ► So I ran this earlier tonight. I fixed my entire problem and yeah, thanks.

00:34:51 ◼   ► Contacts app in apparently Mavericks when this started.

00:34:55 ◼   ► That was a very good feature.

00:34:56 ◼   ► So the cynical message that I didn't write in the Slack to follow up is,

00:35:00 ◼   ► uh, and now enjoy doing that at least once a week for the rest of your life.

00:35:03 ◼   ► Because here's the thing, where did the duplicates come from?

00:35:07 ◼   ► Like you didn't make two of all your contacts.

00:35:09 ◼   ► Somehow they just appeared and not knowing why they appeared,

00:35:12 ◼   ► the fact that you can fix them, you're like, Oh great, I've solved that problem.

00:35:15 ◼   ► But then tomorrow you wake up and you have to all your contacts again.

00:35:18 ◼   ► And then you do find duplicates and fix it.

00:35:20 ◼   ► And then next week you see a do all your God. And so it's like,

00:35:23 ◼   ► I hope that doesn't happen to you. But like, but that is one of the sort of,

00:35:26 ◼   ► you know, when sync goes wrong and you don't know why,

00:35:29 ◼   ► it's great to be able to fix it in an automated way,

00:35:32 ◼   ► but it's still sort of shakes the foundation of your confidence in the,

00:35:36 ◼   ► in the system. Like how did this happen? And will it happen again?

00:35:40 ◼   ► Yeah, that's what I was so surprised to see it because really like the basics

00:35:44 ◼   ► of iCloud syncing, the, you know, contacts, calendars,

00:35:48 ◼   ► like that kind of stuff has always worked very well for me.

00:35:51 ◼   ► And I've heard of other people having issues. I've always,

00:35:54 ◼   ► I'm like a contact syncing unicorn, never had any problems.

00:35:57 ◼   ► And the amount of devices that I set up and,

00:36:01 ◼   ► and delete and you know,

00:36:02 ◼   ► and go through is higher than average.

00:36:05 ◼   ► And I've carried the same database through a much longer time span than probably

00:36:10 ◼   ► the average customer.

00:36:10 ◼   ► And so to have gone this long without having any sync hiccups at all,

00:36:15 ◼   ► I consider that a win. And when this happened, I was like, yeah,

00:36:19 ◼   ► of course this is going to happen. Like if this happens once in,

00:36:21 ◼   ► what is it in 15 years? Fine. That's, that's like mice or cockroaches.

00:36:26 ◼   ► Yeah. You don't just see one of them. Here's the speaking of your database.

00:36:29 ◼   ► You see one. It's okay. If you see more than one, you're screwed.

00:36:32 ◼   ► If you see one, there's a hundred, you don't see that's the whole thing. Um,

00:36:36 ◼   ► speaking of database,

00:36:37 ◼   ► have you made a backup of your contacts database recently slash ever? Yes.

00:36:41 ◼   ► Right before I did the deduping,

00:36:43 ◼   ► I did the export to the big contact archive thing.

00:36:45 ◼   ► That's a thing I would recommend people do anytime they find themselves diving

00:36:50 ◼   ► into the contacts app on their Mac.

00:36:51 ◼   ► Cause they're really gonna get some stuff straight.

00:36:53 ◼   ► Like say before doing your holiday cards, you're like, okay,

00:36:56 ◼   ► I'm going in and I'm gonna,

00:36:57 ◼   ► I'm gonna straighten out my addresses cause a bunch of people moved and yada,

00:37:00 ◼   ► yada stop.

00:37:02 ◼   ► Before you start getting in there and digging around,

00:37:04 ◼   ► make a complete backup of your contacts database.

00:37:07 ◼   ► The contacts app has a command to do this inside of it, uh,

00:37:11 ◼   ► and make a backup of it and don't just call it contacts backup,

00:37:14 ◼   ► put the date in the name,

00:37:15 ◼   ► put it in a folder somewhere and have multiple contacts backups.

00:37:18 ◼   ► So then if you do get bitten by an evil, you know,

00:37:22 ◼   ► contacts bug that keeps duplicating your things, you can always go scorched.

00:37:25 ◼   ► Or the beauty of contacts, unlike photos and other data that is really big,

00:37:29 ◼   ► contacts are relatively small.

00:37:31 ◼   ► You can make a complete backup of them and say, forget it. I give up,

00:37:35 ◼   ► wipe all your contacts, delete everything, and then just restore again,

00:37:39 ◼   ► from within the contacts app,

00:37:40 ◼   ► restore from your backup and get back to hopefully a good starting point.

00:37:43 ◼   ► Even doing that can be difficult because to really delete all of your contacts is

00:37:47 ◼   ► that surprisingly tricky because you'd have to delete it and allow the deletion

00:37:50 ◼   ► to sync to all your devices and make sure there's not one poison device screwing

00:37:53 ◼   ► everything up somewhere. And then when you're confident, okay,

00:37:57 ◼   ► the only contact in there is the me contact,

00:37:59 ◼   ► which is another tricky thing you've got to deal with.

00:38:01 ◼   ► How do you set up the me contact right now? I'm good.

00:38:04 ◼   ► Then you could restore from your backup, but if you don't have a backup,

00:38:06 ◼   ► you're kind of out of luck.

00:38:07 ◼   ► So please everybody make a backup of your contacts before you start screwing

00:38:10 ◼   ► with them.

00:38:11 ◼   ► Also, if you use the, um, you know, file export contacts, archive, uh,

00:38:15 ◼   ► version of this, it puts the date in the file name for you in the box.

00:38:20 ◼   ► So you don't even have to type in the date. It's right there.

00:38:22 ◼   ► Look at that there. I don't know when they started doing that,

00:38:24 ◼   ► but that's convenient. That's exactly what you want.

00:38:26 ◼   ► We are sponsored this week by flat file, the easy,

00:38:31 ◼   ► very well featured import button for your web app.

00:38:35 ◼   ► Think of the last time you imported a spreadsheet to an app.

00:38:38 ◼   ► Did it work correctly? The first time,

00:38:40 ◼   ► nearly everybody has dealt with formatting, messy CSVs or Excel files.

00:38:45 ◼   ► So the data can be imported correctly into an app. It's a huge pain.

00:38:49 ◼   ► Even engineers like us are not spared from this pain.

00:38:52 ◼   ► We are typically doomed to build our data parsers from scratch and usually not

00:38:57 ◼   ► even for the first time. I personally can relate to this.

00:39:00 ◼   ► I built so many CSV importers and it isn't just CSV upload.

00:39:05 ◼   ► It's things like header mapping and data validation or even a nice UI component,

00:39:09 ◼   ► which really adds to the complexity of this as exciting as it is to build

00:39:14 ◼   ► another custom importer compared to your core product features. Our,

00:39:18 ◼   ► our, our friends at flat file have finally made a good solution.

00:39:22 ◼   ► Flat file portal is the elegant import button offering an intuitive data import

00:39:27 ◼   ► experience for your app portal integrates with virtually any application and in a

00:39:33 ◼   ► matter of minutes can intelligently ingest, validate,

00:39:37 ◼   ► and transform incoming spreadsheet data.

00:39:39 ◼   ► So it's clean and ready to use in your backend.

00:39:42 ◼   ► If you're interested in testing out flat file portal in a production environment

00:39:46 ◼   ► or even playing with it, any code sandbox visit get dot flat file dot IO

00:39:51 ◼   ► slash ATP. Once again,

00:39:54 ◼   ► that's get dot flat file dot IO slash ATP for flat file portal,

00:39:59 ◼   ► the elegant import button for your app.

00:40:01 ◼   ► Thank you so much to flat file for sponsoring our show once again.

00:40:05 ◼   ► Okay. So for the last couple of weeks, if you've heard the bootleg,

00:40:12 ◼   ► which is our, you know, no holds barred, immediately released,

00:40:16 ◼   ► not that great sound quality, no ads,

00:40:18 ◼   ► but you get to hear all the garbage version of the show. I'm selling it super.

00:40:22 ◼   ► Really selling it.

00:40:23 ◼   ► Yeah. You can,

00:40:24 ◼   ► you can join us and listening to the bootleg or a ad free but nicely edited

00:40:29 ◼   ► version of the show by going to ATP dot FM slash join.

00:40:33 ◼   ► And for those of you who have already joined, thank you.

00:40:35 ◼   ► You've heard me trying desperately to bring up a topic for like two or three

00:40:40 ◼   ► weeks running that underscore my name is T said in the chat a couple of weeks

00:40:44 ◼   ► ago and they said after show requests for the week after next, which again,

00:40:48 ◼   ► it's like two or three weeks ago now, uh, two ends,

00:40:50 ◼   ► two Wednesday's from the time this was written is the 10th anniversary of the

00:40:53 ◼   ► first build and analyze episode.

00:40:55 ◼   ► It'd be great to hear some reflections on 10 years of podcasting,

00:40:57 ◼   ► how it's changed, what surprised you, et cetera.

00:41:00 ◼   ► I am so desperate to hear this,

00:41:03 ◼   ► not because I really have anything to contribute,

00:41:05 ◼   ► because I really want to hear what you guys have to say about this as the elder

00:41:10 ◼   ► statement statesmen of the three of us. You know,

00:41:12 ◼   ► what has podcasting been like for you for 10 years?

00:41:16 ◼   ► And since we started with build and analyze in this question,

00:41:18 ◼   ► let's start with Marco.

00:41:19 ◼   ► Well, first of all,

00:41:20 ◼   ► you do have a lot to contribute to this because this show is what,

00:41:23 ◼   ► seven years old.

00:41:24 ◼   ► That's a good point. I always, I feel like almost eight. Exactly.

00:41:27 ◼   ► It's like seven and a half years old. So you know, it really,

00:41:32 ◼   ► you have been here for most of this time.

00:41:34 ◼   ► You also were listening for this whole time. That's true too. So,

00:41:39 ◼   ► so you do have things that you should add to this.

00:41:41 ◼   ► So I am going to put you on the spot and have you add stuff to this later as

00:41:44 ◼   ► well. But I honestly, I, you know,

00:41:47 ◼   ► I don't have a lot to say about this right now. You know,

00:41:51 ◼   ► it's in many ways what we do as the,

00:41:56 ◼   ► the kind of podcasters we are, which that's a big asterisk,

00:41:59 ◼   ► but the kind of podcasters we are in many ways,

00:42:02 ◼   ► what we do hasn't really changed much during this time. You know,

00:42:05 ◼   ► we have our handful of totally non diverse white men,

00:42:11 ◼   ► talking with each other for a couple hours each week,

00:42:15 ◼   ► unscripted, you know, this, you know, we have like an outline of topics,

00:42:18 ◼   ► but we don't like script things that we're going to say. It's real.

00:42:22 ◼   ► It's a very,

00:42:23 ◼   ► very high quality implementation of very low production value.

00:42:28 ◼   ► So like if that makes sense,

00:42:30 ◼   ► like we try to sound as good as we can with sound quality and we put it in some

00:42:35 ◼   ► niceties that we didn't used to do things like chapter markers and you know,

00:42:39 ◼   ► we have our, our ad free members version as case we're just talking about and

00:42:42 ◼   ► stuff like that. It's like we have like production values in that sense. We,

00:42:46 ◼   ► you know, we do detailed show notes for each episode.

00:42:49 ◼   ► We do a little description.

00:42:50 ◼   ► So like the experience in a podcast app is pretty rich.

00:42:53 ◼   ► Our website exists and is functional. And so, you know,

00:42:56 ◼   ► it remains a product in our lineup. Yeah, no, but like there's,

00:43:00 ◼   ► there's permalink pages for each episode. That's not a given for podcasts.

00:43:04 ◼   ► Trust me, I know I make, I maintain an index of podcasts.

00:43:07 ◼   ► Like that's far from a given. You know,

00:43:09 ◼   ► our podcast is easily shareable because we don't do dynamic ad insertion and so

00:43:14 ◼   ► our timestamps matter and are stable. And so stuff like that. So like we,

00:43:18 ◼   ► we have a lot of production value on that side of things,

00:43:21 ◼   ► but we are not doing like a music bed under every segment and having scripted

00:43:26 ◼   ► segments and interviews where we go out on scene and talk to a person in Iowa

00:43:32 ◼   ► whenever, like we don't,

00:43:33 ◼   ► we don't do stuff like that the way like high production podcasts do.

00:43:37 ◼   ► Cause that's just a different style of show.

00:43:39 ◼   ► It requires a totally different skill set and a staff and everything else that

00:43:43 ◼   ► this is not what we do.

00:43:44 ◼   ► But what we do is a really good version of people talking to each other

00:43:50 ◼   ► unscripted about computers and Apple news each week and funded by a few

00:43:56 ◼   ► statically baked in host red ads spread throughout the episode.

00:44:02 ◼   ► And that basic format, we've been doing that for pretty much 10 years. You know,

00:44:07 ◼   ► we've refined the format over time,

00:44:09 ◼   ► but that basic thing we've been doing for pretty much the whole time and that

00:44:14 ◼   ► existed before us.

00:44:15 ◼   ► And there is no sign of that stopping anytime soon in the future either.

00:44:20 ◼   ► But that's,

00:44:24 ◼   ► that's just what we do in our little corner of podcasting.

00:44:26 ◼   ► What has mostly changed in the last 10 years is that while this is a very

00:44:33 ◼   ► different world, while this version of podcasting has definitely grown,

00:44:37 ◼   ► the rest of the podcasting world became way bigger and went in some pretty

00:44:43 ◼   ► different directions.

00:44:43 ◼   ► So our version of podcasting became a much smaller proportion of the podcasting

00:44:48 ◼   ► world as a whole. And some of this is great. You know, some of this,

00:44:51 ◼   ► I mentioned our diversity issues earlier as podcasting has gotten bigger,

00:44:56 ◼   ► diversity has dramatically improved and thank God it needed to,

00:44:59 ◼   ► it still needs to,

00:45:00 ◼   ► we're still nowhere near like being a representative sample of the world's

00:45:04 ◼   ► population, not even close.

00:45:06 ◼   ► But it is way better than it used to be and we're making progress in that area.

00:45:10 ◼   ► So that's good.

00:45:11 ◼   ► There are other areas of where podcasting has gone that might not be so great

00:45:16 ◼   ► for us. You know,

00:45:19 ◼   ► things like as the big money platforms move in and start locking stuff down and

00:45:24 ◼   ► making all these like exclusivity deals that take people out of podcasting and

00:45:29 ◼   ► lock their stuff into a certain app or something like that's, that's not great.

00:45:33 ◼   ► Like that, that actively threatens the world that we're in.

00:45:36 ◼   ► Not as much as people think, but it does threaten it. And also, you know,

00:45:40 ◼   ► I mentioned earlier that we are funded as we've been this entire time,

00:45:44 ◼   ► mostly by the sponsor reads that are in each episode.

00:45:47 ◼   ► And if the sponsor landscape dramatically changes in some way,

00:45:52 ◼   ► that could affect us too. So even if like the big business side of podcasting,

00:45:59 ◼   ► you know,

00:45:59 ◼   ► most of the time we can kind of coexist and we don't get any shutters away and

00:46:03 ◼   ► what they do is over there and we do is over here and it's fine.

00:46:05 ◼   ► But if they cause major changes to happen in the sponsor landscape,

00:46:11 ◼   ► that could affect us.

00:46:13 ◼   ► So that could be things like if all sponsors start requiring way more tracking

00:46:18 ◼   ► data than what we give them, you know, so far we have had to turn away,

00:46:24 ◼   ► I think one or two sponsors ever because we don't do demographic tracking.

00:46:29 ◼   ► But everyone else and they've required it and everyone else is like,

00:46:33 ◼   ► they ask and we say, we don't do that. And they say, okay.

00:46:35 ◼   ► And they just, they buy that anyway. Right.

00:46:38 ◼   ► We should tell you something about the value of demographic tracking. Anyway,

00:46:41 ◼   ► the, but like,

00:46:44 ◼   ► if the sponsor landscape shifts around in bigger ways where sponsors start

00:46:49 ◼   ► requiring for all the shows they sponsor the same kind of data or other things

00:46:58 ◼   ► that they get from like the really,

00:46:59 ◼   ► really big shows with their dynamic ad insertion platforms and like micro

00:47:03 ◼   ► targeting and tracking and all that stuff,

00:47:05 ◼   ► then that would affect us in a negative way.

00:47:08 ◼   ► But for the most part as the big world of podcasting has grown,

00:47:13 ◼   ► most of the way it is ways that it grows is just in having some really big shows.

00:47:19 ◼   ► And if a show is really big, that doesn't really negatively affect us, you know,

00:47:25 ◼   ► in almost any possible way. Again,

00:47:27 ◼   ► these kind of like macroeconomic factors could affect us, but they, so far,

00:47:31 ◼   ► they mostly haven't. So it's been overall good for us.

00:47:35 ◼   ► But interestingly, like the thing that we do,

00:47:38 ◼   ► like the kind of shows that we make, the kind of shows I listened to, frankly,

00:47:43 ◼   ► are mostly the same as they were 10 years ago. There's just,

00:47:47 ◼   ► there's more of them. There, there are more people in this world than there were.

00:47:51 ◼   ► Back then there are more listeners in this world.

00:47:54 ◼   ► The shows have all gotten better.

00:47:55 ◼   ► Like we've gotten better at doing this format and we've gotten just better at our

00:47:58 ◼   ► jobs and the equipment has gotten better. The software has gotten better. Our,

00:48:02 ◼   ► you know, our techniques and, and you know, that, that's all gotten better.

00:48:06 ◼   ► And hopefully we've gotten a little bit wiser in a decade.

00:48:09 ◼   ► But for the most part, the,

00:48:15 ◼   ► the core of what we do here is

00:48:20 ◼   ► basically the same but better now. And, and you know,

00:48:24 ◼   ► 10 years is a long time. I think it's a lot better. But overall,

00:48:28 ◼   ► we're still, you know,

00:48:30 ◼   ► hanging out talking with our friends once a week for a couple hours about

00:48:33 ◼   ► technology and Apple stuff. And I love that.

00:48:36 ◼   ► And I'm in no rush to change or stop that.

00:48:39 ◼   ► Because I still can't believe that like we get paid to do this.

00:48:44 ◼   ► Like that's, that's incredible to me. It's still incredible to me.

00:48:48 ◼   ► I said this before, so I won't go too far, too far into it,

00:48:49 ◼   ► but like I'm not even a good speaker. I've never been a good speaker.

00:48:55 ◼   ► I don't pronounce words right. And I stutter constantly.

00:48:59 ◼   ► And I don't plan what I'm saying. And I say, uh, or um, and like,

00:49:02 ◼   ► and constantly I know this cause I'm, I'm the editor of the show.

00:49:06 ◼   ► You would think like, you know,

00:49:09 ◼   ► I did used to think before podcasting that like I could never get a job as like,

00:49:13 ◼   ► like one of my like kid fantasy jobs was I wanted to be a radio DJ,

00:49:17 ◼   ► but I thought, well, I don't have the voice for it. I don't,

00:49:20 ◼   ► I don't speak well enough.

00:49:21 ◼   ► I don't have that kind of like low nineties radio DJ voice,

00:49:23 ◼   ► like all that stuff. So I figured, yeah, you know what? That's,

00:49:25 ◼   ► I guess I won't ever be able to do that.

00:49:26 ◼   ► It turns out being a radio DJ is a terrible job in practice and being a

00:49:32 ◼   ► podcaster is an awesome job in practice. And yeah,

00:49:35 ◼   ► I could never be a radio DJ. I'm not, I don't have the right traits to do that,

00:49:39 ◼   ► but I do apparently have the right traits to stumble into this.

00:49:43 ◼   ► Because the world of podcasting is different and more forgiving and more human

00:49:48 ◼   ► and less of that like polished, you know,

00:49:51 ◼   ► eighties artificial perfection kind of, kind of image. And so this,

00:49:55 ◼   ► like this I can do and it still shocks me that I can do this,

00:50:00 ◼   ► but you know, now like it's been 10 years and after this long,

00:50:05 ◼   ► like I'm not,

00:50:05 ◼   ► I used to be nervous before every single show for probably the first six or

00:50:09 ◼   ► seven years that's finally stopped. But yeah, otherwise like I,

00:50:14 ◼   ► I'm just very happy with what we do. I'm,

00:50:18 ◼   ► I'm incredibly happy that people still tune in and listen to this stuff because

00:50:23 ◼   ► to me it's just BSing with my friends about tech that we would have been talking

00:50:26 ◼   ► about anyway.

00:50:26 ◼   ► But we just happened to record it and do it in a more slightly more structured

00:50:30 ◼   ► way. And and that, you know,

00:50:33 ◼   ► that's how we started back then and that's still what we're doing now.

00:50:37 ◼   ► So I'm just really happy that we've kind of found our groove and that this is a

00:50:42 ◼   ► thing that continues to exist and that podcasting for,

00:50:47 ◼   ► for all of the effort that has been thrown at trying to ruin it.

00:50:51 ◼   ► No one has succeeded at ruining it. And I think it's,

00:50:55 ◼   ► I think it's proven to be pretty resilient to being ruined by many things about

00:51:00 ◼   ► its nature that are likely to stay that way for, for a while longer at least.

00:51:04 ◼   ► I hope so. Yeah. So far, I mean, you know, back when I started,

00:51:08 ◼   ► I wasn't as much in the business. I, you know,

00:51:11 ◼   ► I didn't make overcast until almost halfway through this period.

00:51:16 ◼   ► So, you know, I was only a podcaster before that.

00:51:19 ◼   ► I wasn't a podcast app maker for the whole time. So, you know,

00:51:24 ◼   ► that kind of changed my, my view of it in certain ways. But overall,

00:51:28 ◼   ► I just, I still love this world. I still like back when I,

00:51:32 ◼   ► when I announced overcast at XOXO, I think there was like 2013 or something.

00:51:37 ◼   ► I gave this comparison in my presentation about how like, you know,

00:51:42 ◼   ► the world of text on the web and big text was a world that I kind of was turned

00:51:47 ◼   ► off by.

00:51:48 ◼   ► But the world of big podcasting was just like everything I liked. It was,

00:51:52 ◼   ► it was all like good technology, good people, great audiences.

00:51:57 ◼   ► And I just, I loved how the world of podcasting worked.

00:52:01 ◼   ► And that was almost seven years ago now. Actually,

00:52:05 ◼   ► I might've been exactly seven years ago. And that's still the case.

00:52:09 ◼   ► Like I still look around the web and I still look at like the world of text on

00:52:14 ◼   ► the web, you know, all the big media articles and everything.

00:52:18 ◼   ► And I don't want to be a part of that world.

00:52:22 ◼   ► I was for a long time and it's,

00:52:24 ◼   ► it really took some bad turns over the, over these years. And it's not,

00:52:30 ◼   ► it's not a good place for a lot of people to be anymore.

00:52:33 ◼   ► Some people can still make, make a good time out of it,

00:52:35 ◼   ► but it's certainly not what it was.

00:52:39 ◼   ► Podcasting seems to just still be getting better. And you know,

00:52:42 ◼   ► as all this big money has moved in, it's certainly, you know,

00:52:45 ◼   ► there have been a little bit of kind of uneasy times here and there where we're

00:52:49 ◼   ► not like we see some massive amount of money being spent to try to ruin it.

00:52:53 ◼   ► And it's, you know, it's hard not to feel a little bit scared when that happens.

00:52:56 ◼   ► But so far like the,

00:52:58 ◼   ► those attempts mostly just bounce off of us or they are a bunch of the big shows

00:53:03 ◼   ► and the big ad networks and the big hosting platforms that host the big shows

00:53:08 ◼   ► with the big ad networks all like buying each other and doing things that affect

00:53:13 ◼   ► each other, but that don't affect us over here.

00:53:16 ◼   ► So as long as that dynamic continues to be the case, I'm,

00:53:19 ◼   ► I'm very optimistic and I love the world of podcasts and I, uh,

00:53:22 ◼   ► I'm very happy to be in it all this time and I'd like to stay here.

00:53:26 ◼   ► Concur. You know, I really want to hear what John has to say, but just briefly,

00:53:31 ◼   ► you know, I was looking as we record,

00:53:34 ◼   ► I get a pretty solid first draft of the show notes together and Marco will,

00:53:38 ◼   ► you know, edit things here and there.

00:53:40 ◼   ► And John will add things here and there and edit things occasionally.

00:53:43 ◼   ► But I was looking at the,

00:53:45 ◼   ► to get the link for the XOXO video that you just mentioned and you know,

00:53:49 ◼   ► I found it pretty quickly and I'm looking at this and it says, you know,

00:53:53 ◼   ► Marco Arment, the magazine slash Instapaper.

00:53:56 ◼   ► When he asked New York writer and programmer Marco Arment to speak at XOXO,

00:53:59 ◼   ► he was still developing Instapaper in the magazine and Tumblr was an independent

00:54:02 ◼   ► startup. Since then,

00:54:03 ◼   ► he sold the magazine to Glenn Fleischman Instapaper to Betaworks and Tumblr sold

00:54:06 ◼   ► to Yahoo and a billion dollar deal.

00:54:08 ◼   ► One of the best parts of independence is choosing what you work on.

00:54:11 ◼   ► And Marco's clearing his plate for something brand new.

00:54:13 ◼   ► It's funny to me that I'll read this and the magazine was like 30 years ago,

00:54:18 ◼   ► but ATP started last year. Like that's the way I feel. You know what I mean?

00:54:24 ◼   ► Like the magazine was so long ago to me.

00:54:27 ◼   ► It was forever ago and yes, conceptually, intellectually,

00:54:31 ◼   ► I understand that ATP was seven, you know, we started almost eight years ago now,

00:54:35 ◼   ► but it doesn't feel like it's been almost eight years,

00:54:38 ◼   ► which hopefully is a testament to the two of you and our awesome, you know,

00:54:42 ◼   ► our awesome listeners. But nevertheless,

00:54:44 ◼   ► it's just it's so funny how one thing can feel forever ago and another feels

00:54:48 ◼   ► like it was just yesterday. But anyway, John,

00:54:50 ◼   ► what are your thoughts on this as the other elder statesman?

00:54:53 ◼   ► Like when you need to,

00:54:54 ◼   ► I need to plan more surprise segments in the show to make Marco feel nervous

00:54:58 ◼   ► before we record again. I'm going to get back that nervous energy. Why?

00:55:02 ◼   ► You shouldn't be coming into the show complacent.

00:55:03 ◼   ► You don't know what's going to happen. Anything can happen.

00:55:05 ◼   ► You got to stop pre-flighting before we record.

00:55:07 ◼   ► It's time for Marco to get nervous again.

00:55:09 ◼   ► Yeah, I think, yeah,

00:55:11 ◼   ► I think Marco covered most of the industry stuff pretty well.

00:55:14 ◼   ► And I think I got most of the navel gazing/back padding out in the episode 400

00:55:19 ◼   ► spectacular. So the only thing I'll add here is,

00:55:23 ◼   ► some personal reflections on 10 years of podcasting.

00:55:26 ◼   ► Did I start hypercritical or maybe I started on the incomparable before

00:55:31 ◼   ► hypercritical. I don't remember. Like this is where you keep good records.

00:55:35 ◼   ► Cause no one else cares about it. Someone says, when did you start podcasting?

00:55:38 ◼   ► Honestly, I don't know. Like I can,

00:55:40 ◼   ► I can look it up and I can think I can determine that.

00:55:42 ◼   ► I think I was on the incomparable before the first episode of hypercritical,

00:55:45 ◼   ► but I might've been on another ad hoc podcast with episode with Dan Benjamin

00:55:50 ◼   ► before that.

00:55:51 ◼   ► You were probably on his interview show.

00:55:53 ◼   ► Yeah. What was that called? I forgot. Yeah. I don't know. We're all,

00:55:57 ◼   ► we're all too old. But all I'm saying is,

00:55:59 ◼   ► I think I started podcasting around the same time as Marco. So here, you know,

00:56:03 ◼   ► so I've been podcasting for a similar amount of time.

00:56:06 ◼   ► And the reason I am reflecting on this, I think about how,

00:56:12 ◼   ► how did I get into podcasting? Cause I had been, you know,

00:56:14 ◼   ► I've been a programmer since I left school since even before I left school.

00:56:18 ◼   ► That's been my profession. I still do it professionally.

00:56:20 ◼   ► And I had been, you know,

00:56:24 ◼   ► in the sort of Mac web world from the Usenet groups to eventual websites and web

00:56:30 ◼   ► forums. And eventually I started writing for Ars Technica and did all the Mac OS

00:56:34 ◼   ► 10 reviews. So I was, you know, I was in the tech world as well.

00:56:37 ◼   ► So I was a programmer and did tech writing, but podcasting,

00:56:42 ◼   ► it's another thing I can't really remember.

00:56:44 ◼   ► The first podcast I remember listening to though,

00:56:45 ◼   ► the first one that had made an impression on me was a Merle Aphrodis podcast

00:56:49 ◼   ► called I Should Be Writing.

00:56:51 ◼   ► As the tagline of the show went back when I started listening to it,

00:56:57 ◼   ► it was a podcast for wannabe fiction writers by a wannabe fiction writer.

00:57:02 ◼   ► At the time, Merle Aphrodis wanted to be a fiction writer and she was trying,

00:57:06 ◼   ► and she made a podcast where it was literally just her and a microphone.

00:57:11 ◼   ► And that was it. No music beds, no intro music, no theme song,

00:57:17 ◼   ► just Merle Aphrodis and a microphone.

00:57:19 ◼   ► And she would talk about the struggles of being a writer,

00:57:24 ◼   ► different things about being a writer, how to plan your stories, you know,

00:57:27 ◼   ► what to work on in your writing, the value of writer's workshops,

00:57:30 ◼   ► how to try to get an agent. And, you know,

00:57:32 ◼   ► she wasn't coming as an expert because she was a wannabe fiction writer.

00:57:36 ◼   ► She was working through these things in real time and podcasting about them.

00:57:39 ◼   ► Eventually the show tagline had to change because she became a published author.

00:57:43 ◼   ► She just kept doing the podcast, so on and so forth. But anyway,

00:57:46 ◼   ► I would listen to her in the car as an alternative to listening to NPR or

00:57:50 ◼   ► whatever I was listening to on my radio because, you know, I had music,

00:57:54 ◼   ► I think I had music on my iPod at this point, but I, you know,

00:57:57 ◼   ► my cars didn't have Bluetooth and so it was kind of hard to hook up.

00:57:59 ◼   ► So I would listen to NPR.

00:58:00 ◼   ► And when I didn't want to listen to that say during pledge week or whatever

00:58:03 ◼   ► pledge month, pledge decade, I was listening to, I should be writing.

00:58:07 ◼   ► I didn't want to be a fiction writer particularly, you know,

00:58:09 ◼   ► I enjoyed writing and I enjoyed hearing people talk about it,

00:58:12 ◼   ► but I just love listening to her talk about, I mean,

00:58:15 ◼   ► she's got a great voice for podcasts. It's very calming. It's very soothing.

00:58:18 ◼   ► She's, she was, you know, very sort of, uh,

00:58:23 ◼   ► thoughtful, uh, and, you know,

00:58:26 ◼   ► spoke about writing with a lot of feeling, uh, but with, you know,

00:58:31 ◼   ► a lot of good insights and, uh,

00:58:33 ◼   ► what sounded to me like good advice for all I knew about, about fiction writing.

00:58:38 ◼   ► Um, and so that was one aspect of it. And the second one was at this time,

00:58:42 ◼   ► I was going into the office like we all used to and working as a programmer.

00:58:46 ◼   ► And we, we had like, this was one of the smaller companies I was at.

00:58:49 ◼   ► There's like maybe one or 200 people by the end and started off significantly

00:58:54 ◼   ► smaller. And we had a lunch room and we'd all sit at the lunch table and eat,

00:58:57 ◼   ► whatever we packed as our lunch and just hang out and talk about things.

00:59:00 ◼   ► And inevitably we ended up talking about,

00:59:02 ◼   ► I'd talk with my friends about tech stuff. If we talk about anything,

00:59:05 ◼   ► it was a lot like a lunch room at school where maybe if you sat with your nerdy

00:59:10 ◼   ► friends, you got to talk about like D and D or cars or baseball or whatever it

00:59:14 ◼   ► was you were nerdy about. But if you were at a more sort of, you know,

00:59:18 ◼   ► general audience table, like on a field trip or something,

00:59:20 ◼   ► you just talk about whatever anyone wants to talk about.

00:59:22 ◼   ► And we'd have great conversations at this table about stuff like this.

00:59:26 ◼   ► And as you can imagine me being a big blabbermouth from hearing me on this

00:59:29 ◼   ► podcast,

00:59:30 ◼   ► sometimes I'd go off on a big rant about something and it doesn't take much,

00:59:33 ◼   ► you know,

00:59:33 ◼   ► it can be about refrigerators or computers or pizza or bagels or like,

00:59:38 ◼   ► you know, whatever it would be.

00:59:39 ◼   ► And sometimes I'd go off on one of those and I'd see that I had a little

00:59:44 ◼   ► audience of people who was listening to me do this.

00:59:48 ◼   ► And then I would quickly shut up because I would be embarrassed or whatever.

00:59:50 ◼   ► But I did take note that occasionally I could entertain an audience with the

00:59:55 ◼   ► thing that I do that I now do on podcasts. So these two things combined, Hey,

01:00:00 ◼   ► I like listening to Merle Lafferty talk about a topic that I'm not even that

01:00:04 ◼   ► interested in just because I find it entertaining and soothing and interesting.

01:00:09 ◼   ► Right. Uh,

01:00:11 ◼   ► and also I have things that I can talk about.

01:00:14 ◼   ► I'm writing articles on Ars Technica and people are reading those.

01:00:17 ◼   ► And I know a lot about a couple of specific tech topics.

01:00:21 ◼   ► Maybe I could combine those two things.

01:00:23 ◼   ► And it was a certain point where for doing both these things combining,

01:00:26 ◼   ► I said,

01:00:27 ◼   ► I think podcasting is a thing I might be able to do, right?

01:00:31 ◼   ► Because I'd heard people podcasting and I'd had experiences in real life where

01:00:35 ◼   ► occasionally people would be entertained by me talking about things.

01:00:39 ◼   ► And so that sort of culminated in, you know, however it happened,

01:00:43 ◼   ► that it probably knowing me,

01:00:44 ◼   ► what probably happened is Dan Benjamin probably reached out to me and said, Hey,

01:00:47 ◼   ► do you want to be in a podcast? Because like so many things in my life,

01:00:50 ◼   ► I need to be dragged kicking and screaming into things that are good for me.

01:00:53 ◼   ► Right. So I've always been very bad about sort of putting myself forward.

01:00:57 ◼   ► How it happened in Ars Technica.

01:00:58 ◼   ► I don't think I've told the story a million times or not,

01:01:01 ◼   ► but like I was active in the Ars Technica forums and Ken Fisher,

01:01:05 ◼   ► who founded Ars Technica, reached out to me and said, Hey,

01:01:08 ◼   ► would you like to write something for Ars Technica?

01:01:10 ◼   ► I didn't go banging on his door and say, I want to write articles for Ars Technica.

01:01:14 ◼   ► Every almost everything is good to happen to me.

01:01:16 ◼   ► All I have is that because someone else has dragged me into it,

01:01:18 ◼   ► which is why anyway, that's, that's just the way I am. I'm sorry.

01:01:22 ◼   ► I'm assuming that's what happened with podcasting as well.

01:01:25 ◼   ► But at that point I was primed to jump on that because I had already been

01:01:28 ◼   ► thinking about the idea that I think podcasting could be a thing that I could

01:01:33 ◼   ► do. And sure enough, 10 years later,

01:01:36 ◼   ► it seems like it is a thing that I'm glad I did.

01:01:39 ◼   ► Even if I had to be dragged into a kicking and screaming,

01:01:41 ◼   ► not really kicking and screaming, but I had to be,

01:01:43 ◼   ► I had to be like a vampire. I had to be invited in. Let's just put it that way.

01:01:46 ◼   ► And now I'm here and you can't get rid of me.

01:01:50 ◼   ► It's funny how different and yet similar all three of our paths into this were.

01:01:55 ◼   ► You know, for me, I didn't, I had been aware of podcasting,

01:02:01 ◼   ► but I don't think I'd ever listened to a podcast until this friend of mine from

01:02:06 ◼   ► when I was a child, literally a child,

01:02:09 ◼   ► who I think I had started to talk to again around this time.

01:02:13 ◼   ► But this guy, his name was Marco Arment,

01:02:15 ◼   ► and he started this podcast called Build and Analyze. And I thought, well,

01:02:18 ◼   ► you know, I know I knew Marco and I'm getting to know him at this point. Again,

01:02:22 ◼   ► I forget where this was in our like a grand reintroduction to each other,

01:02:27 ◼   ► but I believe I was instant messaging my friend Blista.

01:02:30 ◼   ► How can we forget? Um, but, uh,

01:02:35 ◼   ► one way or another, you know, this was around the time that Marco and I,

01:02:38 ◼   ► you know,

01:02:38 ◼   ► cause we were as close as two kids who saw each other once a year could be,

01:02:42 ◼   ► you know, when we were like 10 and then, you know, and not in an angry way,

01:02:46 ◼   ► we just fell out of contact for a long time.

01:02:47 ◼   ► And then around this time we started to talk again. Again,

01:02:51 ◼   ► I don't remember if we were talking already and then build and analyze started

01:02:53 ◼   ► or if, you know,

01:02:54 ◼   ► build and analyze was what spurred me into reaching out or whatever the case

01:02:58 ◼   ► may be. Maybe it was Marco that reached out. I don't know. But point is, you know,

01:03:00 ◼   ► we were, we were getting to know each other again in the sky, Marco Arment,

01:03:03 ◼   ► he was like, no, he's a friend, you know, he's doing a podcast.

01:03:05 ◼   ► I should check that out. And he does the same sort of work that I do for a living,

01:03:09 ◼   ► you know, in a different environment. You know,

01:03:11 ◼   ► I was all in on the windows stuff at the time, but you know,

01:03:14 ◼   ► it's still writing code so I should check that out. And you know,

01:03:17 ◼   ► as so many people have said to us, you know,

01:03:19 ◼   ► once you start listening to build and analyze, you know,

01:03:21 ◼   ► Marco's constantly name dropping hypercritical.

01:03:23 ◼   ► And then once you start listening to hypercritical,

01:03:25 ◼   ► you get constant name drops of the talk show.

01:03:27 ◼   ► And so it does not take long before every waking minute of your day is spent

01:03:32 ◼   ► listening to other people talk. And then, you know,

01:03:36 ◼   ► I would listen to the show to build and analyze.

01:03:39 ◼   ► I would listen live and I would be one of the jackals in the chat room.

01:03:41 ◼   ► And I remember even though Marco and I are friends and were at the time,

01:03:45 ◼   ► you know, every great once in a while he would mention my name or,

01:03:47 ◼   ► or even more impressively because I didn't know him,

01:03:49 ◼   ► John would mention my name and my,

01:03:51 ◼   ► my freaking heart would beat out of my chest. Cause like, Oh my God, Oh my God,

01:03:55 ◼   ► Oh my God. They, they know I exist. Like Mario, of course, Marco knows I exist.

01:03:58 ◼   ► Like we're friends. But nevertheless, when you hear it in the context of the show,

01:04:01 ◼   ► it's like, Oh my God, what's happening? And, uh, and so that was, you know,

01:04:05 ◼   ► circa 2010, 2011, something like that. And then I guess it was the end of 2012,

01:04:10 ◼   ► or thereabouts when I started needling Marco about doing a car show with me.

01:04:13 ◼   ► And I've told the story several times before, probably on this very program.

01:04:16 ◼   ► But you know, Marco had the genuinely brilliant idea of saying, well, you know,

01:04:20 ◼   ► I stopped building analyze and you know,

01:04:23 ◼   ► John Syracuse had just stopped hypercritical and he likes car stuff.

01:04:28 ◼   ► I wonder if the three of us could do something together. And,

01:04:31 ◼   ► and that's when neutral started and I still stand by neutral.

01:04:35 ◼   ► It was a disaster, but it was a beautiful, beautiful,

01:04:37 ◼   ► beautiful disaster and I loved it so much. And, and again, Marco,

01:04:41 ◼   ► in a stroke of brilliance realized, well,

01:04:44 ◼   ► when we stopped recording neutral and just started goofing off about the things

01:04:48 ◼   ► that, you know, allegedly we weren't there to talk about,

01:04:51 ◼   ► which always ended up being nerdy tech stuff. Well, Marco put,

01:04:54 ◼   ► what was it on sound cloud? The first few episodes went up. Is that right?

01:04:57 ◼   ► Yeah.

01:04:58 ◼   ► And then it turns out that we actually know things about technology and don't

01:05:02 ◼   ► really know squat about cars and you know, fast forward a couple of years,

01:05:06 ◼   ► three, four years and suddenly this is my job. Like it's, it's,

01:05:09 ◼   ► it's utterly bananas.

01:05:11 ◼   ► The path that I have taken from being one of those people in the chat room just

01:05:16 ◼   ► trying to get noticed. And I mean that both in a self deprecating and also a

01:05:20 ◼   ► literal way. Um, but now, you know, I'm,

01:05:23 ◼   ► I'm hosting what I like to think of at least for our, you know, little pond,

01:05:26 ◼   ► a relatively popular and important program. And it's,

01:05:29 ◼   ► it's amazing to me that this is how I can put food on the table and how I can

01:05:34 ◼   ► feed my family and how I can spend my time.

01:05:36 ◼   ► And the three of us are indescribably lucky that any of you are listening,

01:05:41 ◼   ► much less any of you who have spent money on either an advertiser or,

01:05:46 ◼   ► or membership or what have you. Um, it's wild. And it's,

01:05:50 ◼   ► it's scary to see these changes.

01:05:52 ◼   ► Scary for me cause I have very chicken little tendencies. Um,

01:05:56 ◼   ► it's scary for me to see all of this big money coming into big time podcasting.

01:06:00 ◼   ► And while I do think a rising tide raises all boats, you know,

01:06:04 ◼   ► and having cereal takeoff in such a big way, God have been poisoned.

01:06:08 ◼   ► I almost said so bigly, Oh God, is it over yet? Oh God, I know it's so bad. Uh,

01:06:13 ◼   ► having, having cereal takeoff in such a big way is helpful to all of us, right?

01:06:17 ◼   ► Because if, if,

01:06:18 ◼   ► if you're around the proverbial water cooler and everyone is saying you have to

01:06:23 ◼   ► listen to cereal, this podcast, it's amazing. Well, just like I said, you know,

01:06:27 ◼   ► cereal isn't about to name check the accidental tech podcast,

01:06:31 ◼   ► but if you find this thing that you enjoy and if you happen to enjoy Apple stuff,

01:06:36 ◼   ► I like to think that, you know,

01:06:37 ◼   ► ATP is one of a handful of shows that'll bubble to the top of the list of things

01:06:41 ◼   ► you should check out.

01:06:42 ◼   ► And so it may be that probably more listeners than we would expect came to us in

01:06:47 ◼   ► an indirect way from cereal. And it scares me to see what's going on with Spotify.

01:06:51 ◼   ► It scares me to see these exclusive, it scares me to see this,

01:06:54 ◼   ► this seemingly inevitable march toward the Facebook,

01:06:58 ◼   ► Facebook if vacation, Google of vacation, if you will, of, of big data,

01:07:03 ◼   ► of big data advertising and tracking and so on.

01:07:06 ◼   ► And obviously we have zero intention of ever doing that.

01:07:09 ◼   ► And that's part of the reason why membership is a thing.

01:07:11 ◼   ► But it scares me to think about that.

01:07:13 ◼   ► And I hope that it remains exactly as Marco describes where they're over there,

01:07:18 ◼   ► we're over here, never the two shall meet and everyone's happy.

01:07:21 ◼   ► But I don't know, I just,

01:07:24 ◼   ► I probably say this more often than I should to the point that I hope it's not

01:07:28 ◼   ► frustrating,

01:07:28 ◼   ► but it is important to me that listeners know that as best I can,

01:07:32 ◼   ► I recognize and we recognize how incredibly lucky we are that you'll spend a

01:07:37 ◼   ► couple hours of your week every week with us.

01:07:42 ◼   ► And that is an incredible,

01:07:44 ◼   ► incredible honor because one of the most precious things in the entire world,

01:07:48 ◼   ► golly, especially as we're learning in 2020,

01:07:50 ◼   ► one of the most precious things that we all have is time.

01:07:53 ◼   ► And to spend your time with us is such a compliment.

01:07:56 ◼   ► And I know all three of us are deeply appreciative and,

01:07:58 ◼   ► and it's extremely cool. And I know I speak for me,

01:08:02 ◼   ► and I'm pretty sure I speak for the other guys that, you know,

01:08:04 ◼   ► if you'll have us, we'll be doing this in another 10 years.

01:08:06 ◼   ► Please don't get sick of us cause we'd like to, we'd like to keep doing it.

01:08:11 ◼   ► Actually. And one, one last thing to add,

01:08:13 ◼   ► cause Casey you kind of touched on this,

01:08:14 ◼   ► like with the serialization of podcasting and you know,

01:08:18 ◼   ► all these big hits and the comedians who have all these massive audiences.

01:08:20 ◼   ► One major thing that has changed is that I no longer have to explain what

01:08:25 ◼   ► podcasting is to people. Like if you like, you know, 10 years ago,

01:08:29 ◼   ► you know, people would ask me what I do and oh yeah,

01:08:31 ◼   ► I make an app and I do some podcasts. They'd be like, what?

01:08:34 ◼   ► And even when I started overcast, cause I started overcast before serial,

01:08:39 ◼   ► not by much. It was like a year before something, but it was,

01:08:41 ◼   ► it was still before like that big wave started.

01:08:44 ◼   ► And so I've had to explain so many times to so many people like in the regular

01:08:49 ◼   ► world who are not nerds,

01:08:50 ◼   ► what podcasts are and I largely have stopped having to do that in the last few

01:08:55 ◼   ► years cause they're so big now,

01:08:57 ◼   ► such a large portion of at least Americans so far listen to podcasts.

01:09:02 ◼   ► It's still, podcasting is still very heavily skewed towards the English speaking

01:09:07 ◼   ► world. Um, very, very heavily like, you know, North America and you know,

01:09:11 ◼   ► the English speaking countries, Europe and you know, Australia, Canada,

01:09:15 ◼   ► still, you know, like it's,

01:09:16 ◼   ► it's still very much like a an English Anglo centric, what's the word for that?

01:09:21 ◼   ► An English centric thing. Um,

01:09:23 ◼   ► but among the people I interact with on a, you know,

01:09:27 ◼   ► around here in the English speaking North American world,

01:09:30 ◼   ► podcasting is so ubiquitous.

01:09:32 ◼   ► Podcast listening is so ubiquitous that I no longer have to explain what the heck

01:09:35 ◼   ► it is to almost every single person I meet. And that's, that's pretty great.

01:09:40 ◼   ► And that has only come because of those massive shows that attract those,

01:09:44 ◼   ► you know, giant audiences that are way, way bigger than ours. So, you know,

01:09:48 ◼   ► again, as long as,

01:09:49 ◼   ► as long as their world doesn't end up crushing our world accidentally,

01:09:54 ◼   ► then you know, I'm happy that they can exist and we can exist and everything's

01:09:58 ◼   ► good.

01:09:59 ◼   ► We are sponsored this week by purple. Now look,

01:10:03 ◼   ► you can throw some bedding on a bunch of different mattresses and sure they're

01:10:07 ◼   ► all going to look alike. The same goes for pillows,

01:10:09 ◼   ► but if you peel away the layers, look at what's inside.

01:10:12 ◼   ► You'll see they aren't all created equally.

01:10:15 ◼   ► And that's what makes every purple pillow and mattress unlike anything you've

01:10:20 ◼   ► ever slept on.

01:10:20 ◼   ► The purple grid is this technology that sets the purple mattress apart from

01:10:25 ◼   ► every other mattress.

01:10:27 ◼   ► It's a patented comfort technology that instantly adapts to your body's natural

01:10:31 ◼   ► shape and sleep style with over 1800 open air channels designed to neutralize

01:10:36 ◼   ► body heat.

01:10:38 ◼   ► Purple provides a cooling effect other mattresses can't replicate and this

01:10:42 ◼   ► cutting edge technology doesn't stop with the mattresses.

01:10:45 ◼   ► Every purple pillow is also engineered with the grid for total head and neck

01:10:50 ◼   ► support and absolute airflow. So you're always on the cool side of the pillow.

01:10:54 ◼   ► I love that.

01:10:55 ◼   ► Purple's proprietary technology has been innovating comfort for over 15 years

01:11:00 ◼   ► and you can see this. Look at their reviews. Look around the web.

01:11:05 ◼   ► You will see people back this up. People who have tried multiple mattresses,

01:11:10 ◼   ► they always rate purple really highly. I did the research.

01:11:14 ◼   ► You can do the research. You'll see the exact same thing.

01:11:17 ◼   ► You can try every purple product risk free with free shipping and free returns

01:11:22 ◼   ► and purple has financing available as low as 0% APR for qualified customers.

01:11:28 ◼   ► Experience the purple grid and you will sleep like never before.

01:11:33 ◼   ► Go to purple.com/ATP 10 and use promo code ATP 10 for a limited time.

01:11:38 ◼   ► You'll get 10% off any order of $200 or more.

01:11:42 ◼   ► That's purple.com/ATP 10 promo code ATP 10 for 10% off any order of $200 or more

01:11:47 ◼   ► terms apply. Thank you so much to purple for sponsoring our show.

01:11:53 ◼   ► So you guys have spent more time with your Mac books, air Mac book errors,

01:12:00 ◼   ► however we pronounce it. Uh, anything new you would like to share?

01:12:03 ◼   ► I have one new thing. I think someone asked about this.

01:12:06 ◼   ► It was going to be an ask ATV, but it's a shorty. Someone asked, uh,

01:12:09 ◼   ► is power nap still present on the M one max?

01:12:14 ◼   ► Uh, power nap is the feature that's been around for a while and Mac OS 10.

01:12:18 ◼   ► That basically says when you put your Mac to sleep,

01:12:20 ◼   ► occasionally it will wake up but not really wake up and do a bunch of things

01:12:26 ◼   ► mostly through Apple apps that you might want it to do.

01:12:28 ◼   ► Like it'll do a time machine backup, it'll check your mail or whatever.

01:12:32 ◼   ► But you know, as far as you can tell, your computer is still asleep.

01:12:35 ◼   ► Like it won't turn on the screen.

01:12:37 ◼   ► Some max at various times have tried not to turn on the fans or to turn them on

01:12:40 ◼   ► very low RPM and it will, you know,

01:12:43 ◼   ► basically do some activity when you think it's basically sleeping in sort of a

01:12:47 ◼   ► low power state and then go back to real for real. Is he sleep right?

01:12:50 ◼   ► One of the items on the slides that the presentation introducing the M one max

01:12:56 ◼   ► was like, what was like always on computer, always on processor.

01:13:00 ◼   ► Yeah, always on processor. Um, and if you have an iOS device,

01:13:04 ◼   ► iPhone or iPad or something, you know there's no power nap mode. Like yeah,

01:13:08 ◼   ► when you hit the quote unquote power button, the screen turns off. But yeah,

01:13:13 ◼   ► what determines whether your mail gets checked for is whether the mail

01:13:16 ◼   ► application has background processing turned on and settings, so on and so forth.

01:13:19 ◼   ► The idea being that you don't put your phone to sleep.

01:13:24 ◼   ► It's always kind of in this, in this same state.

01:13:27 ◼   ► It knows whether the screen is on or not,

01:13:29 ◼   ► and it knows whether an app should be allowed to run in the background and the

01:13:32 ◼   ► iOS versions that support that or not.

01:13:34 ◼   ► But there is no mode where it's like asleep and nothing is running to save power,

01:13:38 ◼   ► right? It's got an always on processor, I guess is what Apple's calling it.

01:13:41 ◼   ► Right? So the answer to this question on the M one max,

01:13:44 ◼   ► as far as I can tell any energy or saver slash battery preference pain,

01:13:47 ◼   ► there is nothing that mentions power nap the place where it should be.

01:13:50 ◼   ► There is just no checkbox for it because as far as I can tell,

01:13:53 ◼   ► it's not needed in the same way that it's not needed on iOS devices.

01:13:57 ◼   ► Now Mac iOS is not iOS and there's no sort of setting screen where you can

01:14:01 ◼   ► enable background processing or whatever. So I'm not entirely sure how,

01:14:05 ◼   ► what the interaction is between the OS and the applications,

01:14:08 ◼   ► because other than to know that,

01:14:10 ◼   ► that it seems like the Mac never goes to sleep sleep where nothing is happening

01:14:15 ◼   ► whatsoever. It must go into a low,

01:14:19 ◼   ► low power mode where maybe it just uses one of the efficiency cores or

01:14:22 ◼   ► something. And in that mode,

01:14:24 ◼   ► if mail is running or if there's a background demon that checks mail,

01:14:28 ◼   ► it seems like it will execute. Um, I don't know the technical details of it.

01:14:31 ◼   ► I'm just guessing,

01:14:32 ◼   ► but I can tell you that I could not find a power and act checkbox power nap

01:14:36 ◼   ► checkbox. So I think this is part of the always on processor thing,

01:14:40 ◼   ► that that is a complication that is no longer necessary.

01:14:44 ◼   ► Now that Apple is making the guts of these things.

01:14:46 ◼   ► Marco,

01:14:48 ◼   ► you tweeted earlier that your iMac pro has been a breast,

01:14:52 ◼   ► breath of fresh air to return to and you are so happy to have this actually fast

01:14:57 ◼   ► computer and not have to be slumming it on your Mac book air.

01:14:59 ◼   ► Did I read that right?

01:15:00 ◼   ► It certainly is taking many breaths of fresh air through its fans that I keep

01:15:05 ◼   ► hearing.

01:15:06 ◼   ► Is it really that much faster than that Mac book there?

01:15:09 ◼   ► It's, I mean, the reality is like the speed difference. Yes, it is faster.

01:15:13 ◼   ► Like the Mac book air is indeed faster.

01:15:16 ◼   ► Most of the time I'm not doing, I'm not like maxing it out constantly and,

01:15:20 ◼   ► and such that I would notice the speed difference. Honestly. Um,

01:15:24 ◼   ► I do notice the fan difference,

01:15:25 ◼   ► but I'm also just really happy to be back on my iMac pro after two weeks away

01:15:30 ◼   ► because it has all my stuff on it.

01:15:32 ◼   ► It's a giant screen and it has my nice split keyboard in front of it.

01:15:35 ◼   ► It's my main computer. You know,

01:15:37 ◼   ► when you're away from your main computer and you get back, it's nice.

01:15:39 ◼   ► And I was able to just plow through massive amounts of work today of just like,

01:15:44 ◼   ► you know, catch up admin work that I've been, that I've been, you know,

01:15:47 ◼   ► putting off or missing the ability to do. Um, so I've just been,

01:15:50 ◼   ► I've been getting a lot done on it, so I, I'm happy with this computer.

01:15:53 ◼   ► I do wish I could figure out the fan noise thing.

01:15:56 ◼   ► It does seem to be this intermittent thing probably caused by some dust wedges

01:16:00 ◼   ► and very hard to reach places that I won't be able to get out without like

01:16:03 ◼   ► disassembling the entire computer, which I'm not going to do. Um, so you know,

01:16:07 ◼   ► I do wish I could,

01:16:08 ◼   ► I could solve that problem because I know that when the iMac pro is working

01:16:11 ◼   ► normally, like when it was new,

01:16:13 ◼   ► I didn't have any fan noise like almost ever. It was,

01:16:17 ◼   ► it was nearly impossible to hear them. So I, you know,

01:16:20 ◼   ► I think it's just mostly cause it's old and full of dust, but like us, I, I,

01:16:24 ◼   ► but I am going to,

01:16:27 ◼   ► I'm very much looking forward to what a future iMac can be.

01:16:33 ◼   ► Like I don't think they need to do that much from what the Mac book air is now to

01:16:38 ◼   ► make larger computers that are, that are performance competitive.

01:16:41 ◼   ► There are some areas that I would like to mention quickly. Like, you know,

01:16:43 ◼   ► I think because we're seeing a lot,

01:16:46 ◼   ► a lot of speculation and worrying or,

01:16:50 ◼   ► or excitement about, you know, how do you take the M one and make bigger,

01:16:55 ◼   ► faster, more powerful, more expandable computers with it.

01:16:58 ◼   ► And we're also having a bunch of these new interviews that are being done by

01:17:04 ◼   ► Apple execs that are constantly like one of the biggest things everyone's talking

01:17:09 ◼   ► about, especially the Apple people is the unified memory architecture and quite

01:17:13 ◼   ► how much that matters. And so there is this massive question of like,

01:17:18 ◼   ► okay,

01:17:19 ◼   ► we know that they can make really incredible computers at the low end with this

01:17:24 ◼   ► chip.

01:17:24 ◼   ► We know that as long as they cap the Ram at 16 gigs for now and as long as they

01:17:29 ◼   ► don't need massive GPU power or more than four of these high performance cores,

01:17:36 ◼   ► we know they can make something amazing. But how do you scale that up to the

01:17:41 ◼   ► other models in the lineup?

01:17:42 ◼   ► The number one question I have is how do you scale the Ram?

01:17:48 ◼   ► That's, that's number one because the GPU is,

01:17:52 ◼   ► is almost as important as the Ram in terms of how do you make larger machines?

01:17:56 ◼   ► But the Ram I think is, is probably the harder question to answer.

01:18:02 ◼   ► You know, Apple engineered this giant Mac Pro case,

01:18:06 ◼   ► this whole new Mac Pro. And the, again,

01:18:10 ◼   ► when I talk to the Apple people in that, in that demo room after,

01:18:14 ◼   ► after the Mac Pro unveiling, I talked to a few different people there. And,

01:18:18 ◼   ► and you can also look at some of the comments that have been made publicly by

01:18:21 ◼   ► people like Craig Federighi, you know, like in,

01:18:24 ◼   ► in that machines PR release cycle that really make it seem like they did not

01:18:29 ◼   ► design that entire machine to be a one off. And so if you, if you think, okay,

01:18:35 ◼   ► well somehow the M1 Mac world is going to have to scale up through the rest of

01:18:40 ◼   ► the lineup or at least similar ish products in the rest of the lineup. You know,

01:18:46 ◼   ► I don't know if there's necessarily going to be an iMac Pro,

01:18:48 ◼   ► but I know there's going to be an iMac and I know there's going to be some kind

01:18:51 ◼   ► of Mac Pro and it's probably going to be very similar to the current one in,

01:18:55 ◼   ► you know, general dimensions and capabilities and stuff.

01:18:58 ◼   ► So how do you do that?

01:19:01 ◼   ► We've also seen in the teardowns of these new Macs,

01:19:05 ◼   ► the way they put the RAM quote on package is really,

01:19:10 ◼   ► you know, it's next to the processor. It's not part of the processor.

01:19:14 ◼   ► And we actually heard from a number of people who know more about chip design

01:19:18 ◼   ► and manufacturing than I do who said that it almost,

01:19:21 ◼   ► it would almost make no sense at all to ever move the RAM on die,

01:19:27 ◼   ► like where it's part of the same Silicon wafer and it gets punched out from the

01:19:31 ◼   ► same way for as the CPU.

01:19:33 ◼   ► Because apparently the way you make DRAM is so radically different than the way

01:19:38 ◼   ► you make logic chips at that Silicon level,

01:19:40 ◼   ► that it would be a ridiculous process and it would be inefficient in tons of

01:19:46 ◼   ► ways.

01:19:46 ◼   ► And it would be like really stupid to try to make DRAM and logic chips as part

01:19:51 ◼   ► of the exact same chip. So it's going to be a separate thing. And there's,

01:19:55 ◼   ► and there's, so there, there is already a separation between the CPU,

01:20:00 ◼   ► you know, or like the chip and the RAM.

01:20:02 ◼   ► There is some kind of interconnect between them and I don't know,

01:20:06 ◼   ► I don't know enough about modern memory bandwidth and interfaces to,

01:20:08 ◼   ► to know for sure.

01:20:09 ◼   ► Like how possible is it to just make that interconnect bigger and support four

01:20:14 ◼   ► RAM chips instead of two or eight or 16?

01:20:19 ◼   ► How much would that cost in terms of memory bandwidth or speed or,

01:20:25 ◼   ► you know, what,

01:20:26 ◼   ► would we be reasonably able to make some kind of like RAM in a socket that we

01:20:31 ◼   ► have like on the Mac pro or still even on the high end IMAX,

01:20:35 ◼   ► like can we have socketed RAM still in this new world or are the interfaces

01:20:40 ◼   ► between the pins in the socket and the CPU, are those too slow?

01:20:44 ◼   ► Would that cost too much performance?

01:20:46 ◼   ► So there's this huge open question of how the heck we scale these things.

01:20:50 ◼   ► You could also just take the current chip and just make it a lot bigger.

01:20:55 ◼   ► But then you'd run into some pretty significant cost issues with the

01:20:58 ◼   ► manufacturing and yields of that chip.

01:21:01 ◼   ► So I think the way they're probably going to have to scale it, you know,

01:21:06 ◼   ► we speculated a couple of weeks ago,

01:21:08 ◼   ► I said I think they're going to definitely move the RAM out to its own area of

01:21:13 ◼   ► the board and possibly also move the GPUs out.

01:21:17 ◼   ► And John said definitely not GPUs, I think, right?

01:21:21 ◼   ► I don't recall this conversation, but I can,

01:21:24 ◼   ► I can tell you what I think now for the Mac pro level machines you're thinking.

01:21:29 ◼   ► I'm thinking it's probably for everything above what we have now.

01:21:33 ◼   ► I think it's going to be for all of the laptops that might have a discrete GPU

01:21:37 ◼   ► in the past.

01:21:38 ◼   ► Oh yeah, no, no. I think, yeah, so this is coming back to me. Yeah, no,

01:21:41 ◼   ► I think there,

01:21:42 ◼   ► there is no way the GPU is moving out of the system on the chip for even their

01:21:45 ◼   ► biggest, meanest laptop. I feel like the 16 inch MacBook pro,

01:21:49 ◼   ► their top of the line laptop will have a quote unquote integrated GPU.

01:21:52 ◼   ► It will just be bigger. And I say that not be,

01:21:55 ◼   ► not because I think it's a limitation.

01:21:57 ◼   ► I say that because I think they can meet their performance requirements.

01:22:00 ◼   ► They can make a really, really good,

01:22:02 ◼   ► powerful GPU with it being integrated by just making it bigger and putting fans

01:22:07 ◼   ► on it.

01:22:07 ◼   ► I look at the size of this chip,

01:22:09 ◼   ► such a massive amount of the area on the chip is GPU area.

01:22:14 ◼   ► And another massive amount of it is the high performance CPU cores.

01:22:19 ◼   ► And so if we think like how does this chip scale up to the next level of

01:22:23 ◼   ► performance for the higher end machines? Do they add two more CPU,

01:22:26 ◼   ► two more of the high end CPU core maybe or four more even? Um, do they,

01:22:30 ◼   ► do they double the GPU core count so it has double the GPU performance?

01:22:34 ◼   ► Like that is roughly what we'd be talking about in terms of, you know,

01:22:38 ◼   ► the performance gaps between the model lines that we have had to date. Uh,

01:22:42 ◼   ► we don't know that that's a big assumption.

01:22:44 ◼   ► Maybe the next generation 16 inch Mac book pro isn't going to be twice as fast

01:22:49 ◼   ► as the air. It doesn't have to be technically like, you know,

01:22:54 ◼   ► Apple can design these things however they want basically.

01:22:57 ◼   ► But certainly the customer base has certain expectations and there,

01:23:00 ◼   ► there have things have been done a certain way for, for a long time.

01:23:03 ◼   ► And so we kind of expect that you would probably in the large laptop and in the

01:23:08 ◼   ► iMac and definitely the Mac pro,

01:23:13 ◼   ► we expect we're probably going to have much higher core counts for the CPU.

01:23:18 ◼   ► Uh, you know, eight core, 12 core, you know, for the high performance course,

01:23:21 ◼   ► I mean,

01:23:22 ◼   ► and you look at the GPU and you look at the performance of how it compares to

01:23:27 ◼   ► discreet GPUs today and how it might come, you know,

01:23:31 ◼   ► how you might be able to scale that up to achieve something like a 16 inch Mac

01:23:36 ◼   ► book pro or a high end iMac or a Mac pro customer would demand.

01:23:41 ◼   ► And I don't see how you keep this as one chip.

01:23:43 ◼   ► I don't see it scaling because you would need to make the chip so massive.

01:23:48 ◼   ► They wouldn't be able to manufacture any of them.

01:23:50 ◼   ► Like it would be ridiculous with, with yields and size and everything.

01:23:54 ◼   ► So I think there's gotta be a separation of probably both the Ram expanding so

01:23:59 ◼   ► it can accommodate way more Ram chips, not necessarily being socketed.

01:24:09 ◼   ► I think the, the idea of having, having the Ram be just in regular,

01:24:14 ◼   ► like Ram slots that you can pop out and you know,

01:24:17 ◼   ► put in a stick from crucial if you want to. I think those days are over.

01:24:20 ◼   ► Even with the Mac pro,

01:24:21 ◼   ► I don't see that coming back because what we're seeing is that when they make

01:24:25 ◼   ► the Ram integrated, truly integrated the way it is now, uh,

01:24:29 ◼   ► it can be way faster.

01:24:30 ◼   ► Like they have just massive amounts of bandwidth to the Ram and that's a big

01:24:34 ◼   ► part of why the M1 is so fast.

01:24:37 ◼   ► So I don't see them giving up the Ram speed necessarily,

01:24:42 ◼   ► but I,

01:24:43 ◼   ► I also don't see them sticking with the way it is now where it's just kind of

01:24:48 ◼   ► glued onto the package. Like as one, you know, very,

01:24:51 ◼   ► very small real estate thing. Uh,

01:24:54 ◼   ► I don't see that scaling to a Mac pro that supports hundreds of gigs of Ram.

01:24:59 ◼   ► I don't, I don't see how they can do that. They need more,

01:25:01 ◼   ► they just need more chips. Simple as that. Like they need more chips. Um,

01:25:05 ◼   ► they need more SSD chips too. Like to get larger storage capacities there.

01:25:09 ◼   ► Like they just, they need to scale these up.

01:25:11 ◼   ► And so the question is how did they scale these up? Um,

01:25:14 ◼   ► and you look at the real estate on the chip on the main chip and I don't see how

01:25:19 ◼   ► you meaningfully scale up CPU cores or GPU performance without splitting up that

01:25:25 ◼   ► chip also.

01:25:28 ◼   ► So I still think that the way this scales up is that the way it's going to

01:25:33 ◼   ► is the package gets split into everything that's in it now except the GPU,

01:25:38 ◼   ► which gets split off into its own thing and there's,

01:25:44 ◼   ► they can do whatever they want with how they talk to each other.

01:25:47 ◼   ► They can have some kind of massive hot wide bandwidth internet or interconnect

01:25:51 ◼   ► between the CPU, the GPU and the Ram. But I don't see, I mean,

01:25:56 ◼   ► they know way more about Chipperton than I do,

01:25:58 ◼   ► but I don't see how you double the core count and double the GPU size.

01:26:04 ◼   ► And not have some massive chip that it was very, very hard to manufacture.

01:26:07 ◼   ► And I also don't see how you really scale up to achieve their high end

01:26:13 ◼   ► performance needs for their higher end products without doing that.

01:26:17 ◼   ► Like without having twice the CPU cores and without having at least twice the GPU

01:26:22 ◼   ► area. This remains to be seen like how they're going to do this. But I also,

01:26:25 ◼   ► you look at how the M1 is turning out so far.

01:26:28 ◼   ► You look at what they're talking about, what they're really pushing.

01:26:32 ◼   ► And the unified memory architecture does not sound like it's going to be

01:26:35 ◼   ► optional in the future. It sounds like they're all going to have that.

01:26:38 ◼   ► So somehow they're going to keep a unified memory architecture,

01:26:41 ◼   ► but they're going to have to be able to support way more Ram than they have now,

01:26:46 ◼   ► more CPU cores and more GPU cores.

01:26:48 ◼   ► So I don't know how they do that without splitting this up.

01:26:51 ◼   ► People are getting hung up,

01:26:53 ◼   ► hung up a little bit on the unified memory architecture thing.

01:26:55 ◼   ► Like part of that is, you know, I think we said this before,

01:26:58 ◼   ► like this is it's a marketing term in some respects and that they're

01:27:03 ◼   ► trying to give a name to a thing that they're doing that lots of computers used

01:27:07 ◼   ► to do. Well, the computers used to do it to save money,

01:27:09 ◼   ► but Apple is not doing it to necessarily save money,

01:27:12 ◼   ► which is not having a dedicated pool of video memory. Now,

01:27:16 ◼   ► it was the cheap computers would do that because it's cheaper not to have a bunch

01:27:20 ◼   ► of video memory because there's a bunch of extra expensive chips you need to

01:27:23 ◼   ► have. So we can just use system memory. Max have done this.

01:27:26 ◼   ► You go back to like classic Mac OS or 68 K max. Yeah.

01:27:30 ◼   ► A lot of them had a quote unquote unified memory architecture.

01:27:32 ◼   ► That's all it means is that there's not separate VRAM, right? The, you know,

01:27:36 ◼   ► the iPhones and iPads and the arm based max are doing it for performance reasons.

01:27:41 ◼   ► Most of all, you know, part of it is cost and power savings on the mind,

01:27:44 ◼   ► you know, iOS devices, right?

01:27:46 ◼   ► But on the Mac for these similar devices that are also power constrained, um,

01:27:50 ◼   ► they're doing it just to, you know, to save the power,

01:27:53 ◼   ► but also because they can get really good performance. Like the,

01:27:55 ◼   ► the memory bandwidth of this on package Ram is like three times what it was on

01:28:00 ◼   ► the Intel machines that they replace. Right. So that's a benefit.

01:28:04 ◼   ► And also it's a benefit that you don't have to power the extra VRAM chips and

01:28:07 ◼   ► yada yada. Um, but like you said, Margot, if you, you know,

01:28:11 ◼   ► same thing with the unpackaged Ram and everything.

01:28:13 ◼   ► If you scale that up to the Mac pro the Mac pro,

01:28:15 ◼   ► you can put it 1.5 terabytes of Ram. Um, you, I mean,

01:28:20 ◼   ► you could make a gigantic package and have 1.5 terabytes of unpackaged Ram,

01:28:25 ◼   ► but that starts to look a little bit silly.

01:28:28 ◼   ► You'd have this little tiny system on a chip in the corner and then you'd have

01:28:31 ◼   ► Ram chip, Ram chip, Ram chip,

01:28:33 ◼   ► there's a rows and rows of them lined up probably lined up around it in a

01:28:37 ◼   ► circle. Um, you know, you could do it like if you, um,

01:28:41 ◼   ► if you look at some of the modern game consoles,

01:28:44 ◼   ► I always look at them a lot of people are asking like,

01:28:45 ◼   ► and you made the same point, like how big of a GPU can you put on dive?

01:28:49 ◼   ► Is the GPU is on dive, the CPU, GPU, neural engine.

01:28:52 ◼   ► That's all on the system on a chip dive and then next to it in the same package

01:28:56 ◼   ► are the Ram chips. How big can you double the size of the GPU on a die?

01:28:59 ◼   ► Absolutely. You can look at the Xbox series X or PlayStation five.

01:29:03 ◼   ► They have an AMD CPU and an a,

01:29:06 ◼   ► a pretty darn big AMD GPU all on the same die.

01:29:10 ◼   ► It's a bigger die. It takes more power than the M1 that's for sure,

01:29:14 ◼   ► but you absolutely can do it.

01:29:16 ◼   ► And those GPUs would more than satisfy the needs of a 16 inch MacBook pro.

01:29:21 ◼   ► Again, the 16 inch macro has never been like a gaming laptop.

01:29:24 ◼   ► We're not asking it to have the world's highest end GPU, but the,

01:29:27 ◼   ► the best discrete GPU Apple has ever shipped on a laptop.

01:29:31 ◼   ► I feel like Apple have no problem matching that performance with an on die GPU,

01:29:36 ◼   ► right? I Mac is potentially a different story depending on how they go in that

01:29:40 ◼   ► direction. And of course,

01:29:41 ◼   ► Mac pro is an entirely different story because the Mac pro can have four GPUs and

01:29:44 ◼   ► you're not, you're not going to, you know, or even more than that,

01:29:47 ◼   ► you're not going to fit that onto the die because that's where you start pushing

01:29:51 ◼   ► the limits of the die size. Um, the Ram question is interesting. If we can,

01:29:55 ◼   ► I'm looking at the clock,

01:29:56 ◼   ► I'm thinking we can segue this into the very first ask ATP question if that's

01:30:00 ◼   ► okay with totally nicely done. Um, cause that's what this,

01:30:05 ◼   ► this first question is about here.

01:30:06 ◼   ► I'll let Casey read it because he's good at reading questions,

01:30:09 ◼   ► but then I will attempt to answer it.

01:30:12 ◼   ► All right. So Chris, uh, Chafi coffee, coffee, uh,

01:30:16 ◼   ► Chris C writes the current M one chips have been system on a chip style chips

01:30:20 ◼   ► with the Ram right in the chip package.

01:30:22 ◼   ► I first thought that was that that would be fine for lower end chips and maybe

01:30:26 ◼   ► even all laptops, but for higher end systems like the I Mac pro and the Mac pro,

01:30:29 ◼   ► surely on chip Ram can't be the end state. This is what Marco was describing.

01:30:33 ◼   ► Then I thought about the memory hierarchy where computers really only operate on

01:30:37 ◼   ► registers and everything above that is just slower but larger forms of cash.

01:30:41 ◼   ► L one L two Ram, even, you know, SSD storage. So for some future Mac pro,

01:30:46 ◼   ► the ultimate Mac pro,

01:30:47 ◼   ► what if Apple keeps only eight or 16 gigs of fast Ram on chip,

01:30:50 ◼   ► maybe even calling it an L four cash and then allows off chip cash that just so

01:30:55 ◼   ► happens to be user serviceable D Ram chips.

01:30:58 ◼   ► I think this would give Apple a huge performance boost for the memory intensive

01:31:01 ◼   ► tasks that people put hundreds of gigabytes of Ram into a Mac pro for while

01:31:05 ◼   ► allowing them to keep a lower number of variants that they need to design in

01:31:08 ◼   ► stock.

01:31:08 ◼   ► All right, so here, this question is basically about the cash hierarchy,

01:31:13 ◼   ► the memory hierarchy and computers. So a brief sort of, uh, you know,

01:31:17 ◼   ► introduction to this concept, you know, memory, uh,

01:31:22 ◼   ► it's just a place to store data,

01:31:24 ◼   ► but there's memories of different speeds and different latencies inside your

01:31:28 ◼   ► computer. Um, when your CPU or GPU or whatever is doing some kind of processing,

01:31:32 ◼   ► the fastest sort of closest,

01:31:34 ◼   ► lowest latency memory or registers and they hold small amounts of data,

01:31:37 ◼   ► one or two numbers, five numbers, whatever,

01:31:39 ◼   ► depending how big the registers are. And there's, you know, 10,

01:31:44 ◼   ► 20 of those depending on the execution unit. So those are registers very,

01:31:47 ◼   ► very tight, just one, you know, 10 or 21 number slots or something like that.

01:31:50 ◼   ► Right. Above that you have what they call level one or L one cash,

01:31:54 ◼   ► which is on, you know, also on the CPU, you know, the,

01:31:57 ◼   ► the CPU died just like the registers. Um,

01:32:01 ◼   ► and that is a smaller amount,

01:32:03 ◼   ► some amount of kilobytes or maybe a couple of megabytes or whatever,

01:32:06 ◼   ► depending of memory. What makes that memory fast?

01:32:10 ◼   ► Registers are barely like memory. It's like, you know, like a mail slot.

01:32:13 ◼   ► You can put one or two little things in there.

01:32:15 ◼   ► SIMD registers get weird cause they're a little bit wider, but they're small,

01:32:18 ◼   ► right? L one cache is a pool of memory, but it's a pool, a tiny pool in,

01:32:22 ◼   ► you know, in the grand scheme of things.

01:32:24 ◼   ► Why don't computers just make all their memory L one cache if L one cache is the

01:32:29 ◼   ► fastest memory and it's right next to the computer, why don't they just,

01:32:32 ◼   ► why not make everything out of the stuff that the black box is made of?

01:32:35 ◼   ► I make the whole plan out of the black box, whatever. Um, it's the same answer,

01:32:38 ◼   ► right? So L one cache, first of all,

01:32:42 ◼   ► you can't put 1.5 terabytes on die cause your die can't be that big, right?

01:32:47 ◼   ► So that's reason number one.

01:32:47 ◼   ► Number two is when you make the memory that makes up L one cache,

01:32:52 ◼   ► it's a different kind of memory than DRAM.

01:32:55 ◼   ► Usually it's SD Ram or something similar S Ram S Ram. There you go.

01:32:58 ◼   ► The different,

01:32:59 ◼   ► the main difference in terms of economics that you care about is whereas in D

01:33:04 ◼   ► Ram,

01:33:05 ◼   ► we try to use as few transistors as possible to store a bit of memory.

01:33:10 ◼   ► SRAM will use like six or eight transistor to store one bit of memory.

01:33:14 ◼   ► It's, it's complicated. You can look into it and see,

01:33:16 ◼   ► but basically what it boils down to is it costs way more money and takes up way

01:33:21 ◼   ► more space to store one bit in SRAM than it does in D Ram. Right?

01:33:24 ◼   ► So it's a cost speed trade off. So that's your L one.

01:33:27 ◼   ► L two is also on the die and even larger pool of memory that is more distant in

01:33:33 ◼   ► terms of latency and slower than the L one, right?

01:33:36 ◼   ► So there's your cash hierarchy registers, L one, L two,

01:33:39 ◼   ► sometimes you'll have an L three and you could keep going.

01:33:41 ◼   ► What level and the whole, you know, however many levels of, of a cash you want to

01:33:46 ◼   ► have all about to memory and you can visualize the hard drive being another pool

01:33:51 ◼   ► of caching for Ram, just even slower and more distant. Uh,

01:33:55 ◼   ► the key point of the cash hierarchy is the sort of closer you get to being L one

01:34:01 ◼   ► or, you know, our registers or whatever, the more expensive it is,

01:34:05 ◼   ► the lower latency,

01:34:06 ◼   ► meaning how long does it take for me to get anything into or out of it and the

01:34:09 ◼   ► more expensive it is. Uh,

01:34:11 ◼   ► and to have a useful cash hierarchy,

01:34:16 ◼   ► you need gaps between them.

01:34:18 ◼   ► You need your L one to be slower, cheaper,

01:34:21 ◼   ► and more distant than registers. Otherwise, why wouldn't you just use registers?

01:34:25 ◼   ► You need L two to be slower, cheaper, and more distant than L one, right?

01:34:29 ◼   ► You need Ram to be slower, cheaper, and more distant than L two.

01:34:33 ◼   ► That's the point of the hierarchy.

01:34:35 ◼   ► If L one was 2% faster than L two,

01:34:39 ◼   ► you're not, you know,

01:34:41 ◼   ► doing something useful because having that cash hierarchy has overhead.

01:34:45 ◼   ► The way cash hierarchies work is when you need some piece of memory,

01:34:50 ◼   ► you can say, okay, well let me check. Is it in a register? Well,

01:34:52 ◼   ► this is not how the world works, but just imagine you can check whether it's in

01:34:54 ◼   ► a register. If it's not in a register, you can look in L one,

01:34:58 ◼   ► but it might not be in L one. If it's not in L one, you can check L two,

01:35:00 ◼   ► but it might not be in L two. If it's not L two, you can check in Ram.

01:35:03 ◼   ► All of that checking takes time and adds complexity.

01:35:07 ◼   ► And then you have cash invalidation,

01:35:08 ◼   ► which is one of the two hard problems in computer science.

01:35:11 ◼   ► Naming be the other one. I'm not going to do the joke. Um,

01:35:13 ◼   ► then you have to invalidate that memory.

01:35:19 ◼   ► If something changes something in the memory, Oh, but I have a cash copy,

01:35:22 ◼   ► copy of that in L two and possibly a cash copy in L one.

01:35:24 ◼   ► You need to invalidate those caches or have some sort of cash coherency

01:35:28 ◼   ► algorithm,

01:35:28 ◼   ► especially when you have multiple cores sharing different pools of memory to

01:35:32 ◼   ► make sure everyone's all on the same page in some well-defined way. So there are,

01:35:35 ◼   ► so there's a complexity to a cash hierarchy and to,

01:35:39 ◼   ► for that complexity to be useful,

01:35:41 ◼   ► you need the cash hierarchy to be distant from each other.

01:35:45 ◼   ► You need them to be different.

01:35:46 ◼   ► You need them to have different performance characteristics.

01:35:48 ◼   ► You need the more distant ones to be bigger and cheaper to make up for the fact

01:35:52 ◼   ► that they're slower. You don't want a cash hierarchy where each level of cash is

01:35:56 ◼   ► like a tiny percentage different than the previous one.

01:35:59 ◼   ► So getting back to this question, finally,

01:36:01 ◼   ► what about a Mac that has like an M one or whatever style chip where there is on

01:36:07 ◼   ► package Ram that counts as kind of like an L four cash and then the big pool of

01:36:12 ◼   ► 1.5 terabytes of D Ram,

01:36:14 ◼   ► like in the Mac pro in slots on the motherboard that's slower, right?

01:36:19 ◼   ► So I just got done saying a little bit earlier that according to people who have

01:36:24 ◼   ► measured this,

01:36:26 ◼   ► the memory bandwidth of the on package Ram on M one chips is about three times as

01:36:31 ◼   ► fast as it was on the Intel laptops that they're replacing in terms of bandwidth

01:36:37 ◼   ► like whatever it was like 60 gigabytes per second or something or gigabits.

01:36:40 ◼   ► I forget, but whatever it is, it's around three times as fast,

01:36:42 ◼   ► three times as fast or three times as much bandwidth or whatever I think is

01:36:49 ◼   ► pushing the limits on an acceptable difference in performance for a cash

01:36:54 ◼   ► hierarchy, right? It's not 1%,

01:36:58 ◼   ► but it's also not an order of magnitude and it's also not two orders of

01:37:01 ◼   ► magnitude, right? So I question whether,

01:37:05 ◼   ► oh, first of all, even if you just accept all those numbers and say, well,

01:37:09 ◼   ► D Rams say the DRAM was like the speed of the Intel D Ram,

01:37:11 ◼   ► which I don't accept as a given, but say it was,

01:37:13 ◼   ► you've got a cash hierarchy where there's a three X difference between, you know,

01:37:18 ◼   ► the L four and then the D Ram. I feel like that may not be worth,

01:37:22 ◼   ► the juice may not be worth that squeeze, right?

01:37:25 ◼   ► That you'd want a bigger gap in the cash hierarchy.

01:37:28 ◼   ► And the second thing is getting back to Marco's question,

01:37:31 ◼   ► I don't think there's any reason that you can't have an extremely high bandwidth,

01:37:36 ◼   ► high performance memory bus that is,

01:37:39 ◼   ► that talks to Ram that is not on the package just because the Intel Mac book pros

01:37:45 ◼   ► had Ram that was three times as slow as the M ones did.

01:37:48 ◼   ► Doesn't mean that the $6,000 high end Mac pro can't have an extremely high

01:37:54 ◼   ► bandwidth memory bus. And I haven't looked up the numbers,

01:37:57 ◼   ► but for all I know the current Mac pros have a higher bandwidth memory bus than

01:38:00 ◼   ► the Intel MacBook, which wouldn't surprise me or whatever. Again,

01:38:03 ◼   ► I point to game consoles that have had similar problems where they need to have

01:38:07 ◼   ► a large pool of memory to feed that big honkin' GPU that's in their system on a

01:38:11 ◼   ► chip or whatever. And a lot of them do actually arrange on the motherboard,

01:38:16 ◼   ► a bunch of Ram chips in a circular pattern.

01:38:18 ◼   ► So they're equal distant from the big giant SOC to feed it. That's not,

01:38:22 ◼   ► you know, there's no dims, right?

01:38:24 ◼   ► Because it's a game console and it's like 500 bucks or whatever,

01:38:26 ◼   ► but you can have, or just look at any GPU,

01:38:29 ◼   ► you can have very high bandwidth pools of memory.

01:38:32 ◼   ► They're not literally on the same package, but are on the same board.

01:38:36 ◼   ► Look at a high end GPU, look at their pools of multiple gigabytes of memory.

01:38:40 ◼   ► You can buy GPU's with like 16, 32 gigs of memory.

01:38:43 ◼   ► That memory has a huge extremely wide, sometimes they're like five,

01:38:47 ◼   ► 12 bits wide buses to and from the GPU.

01:38:51 ◼   ► Right? So I'm thinking that

01:38:54 ◼   ► almost all of the benefit you get of having the,

01:38:58 ◼   ► the sort of on package memory chips you can get by having an extremely expensive,

01:39:03 ◼   ► extremely wide, slightly more distant as in not on the same package,

01:39:06 ◼   ► but still on the same board bus on your $6,000 plus the Mac

01:39:11 ◼   ► pro. So when I'm picturing a Mac pro in my head, I like,

01:39:15 ◼   ► the only reason I can imagine doing this sort of L four based approach where

01:39:19 ◼   ► there's,

01:39:19 ◼   ► there's the very fast Ram that's close to it in the big distant pool is if Apple

01:39:24 ◼   ► couldn't justify the expense of doing a custom solution,

01:39:26 ◼   ► because what Chris describes would work and be

01:39:31 ◼   ► feasible.

01:39:32 ◼   ► I still think the gap in performance wouldn't be big enough to justify having a

01:39:36 ◼   ► separate pool of Ram, but you could do it.

01:39:40 ◼   ► And the only reason you do it is like, well,

01:39:41 ◼   ► we didn't want to make a whole new chip that doesn't have,

01:39:44 ◼   ► it has an entirely different interface to memory. So we'll just take what we have,

01:39:48 ◼   ► which is whatever the biggest chip we had in our 16 inch Mac book pro,

01:39:52 ◼   ► and then tack on a bunch more memory and everything.

01:39:54 ◼   ► But I have a feeling that's not what Apple would do. Again,

01:39:57 ◼   ► going back to Marco's impression that he got from hearing people talk about the

01:40:00 ◼   ► Mac pro that this beautiful giant case over here,

01:40:05 ◼   ► like was not made in a world where Apple didn't know it was making our max like

01:40:09 ◼   ► they knew, right?

01:40:10 ◼   ► So there has to be a plan for that and Apple is not shy about charging a lot of

01:40:14 ◼   ► money if they, if they want to do it.

01:40:16 ◼   ► So I feel like they should end up with a solution that is properly fit

01:40:20 ◼   ► for a computer of this size. It doesn't have to be quite as modular.

01:40:24 ◼   ► It may not use standardized parts,

01:40:26 ◼   ► but a solution that gives you huge amounts of Ram that are also

01:40:31 ◼   ► very, very fast is definitely feasible.

01:40:34 ◼   ► And it doesn't require any sort of gymnastics.

01:40:36 ◼   ► The GPU question is a little bit different because Apple has in the past made

01:40:41 ◼   ► computers that have a quote unquote integrated GPU and then in more distant, uh,

01:40:45 ◼   ► but more powerful discrete GPUs.

01:40:48 ◼   ► And they could do that again because the iOS certainly supports it,

01:40:51 ◼   ► but they could also just excise the GPU entirely from the system on a chip and

01:40:54 ◼   ► put it entirely external on the, on a machine like the Mac pro,

01:40:57 ◼   ► because what is all that space in the Mac pro for if not to put cards that

01:41:02 ◼   ► potentially have multiple GPU's on them and tons of Ram and all the other stuff.

01:41:06 ◼   ► Um, I wanted to look up the die, the relative die sizes,

01:41:10 ◼   ► and I couldn't do it in time. Maybe I'll do it for next week.

01:41:12 ◼   ► But I was trying to look up, you know, how much bigger can the die get? The M one.

01:41:17 ◼   ► Yeah. I tried looking this up actually. Do you have a,

01:41:20 ◼   ► I found the Intel die size, but I couldn't find the M one die size.

01:41:23 ◼   ► So my impression is looking at the, I fix it thing is the M one die.

01:41:28 ◼   ► Like it's not anywhere near the limits of what is reasonable to put on a

01:41:33 ◼   ► single die.

01:41:34 ◼   ► So if I imagine it in a Mac pro type system where you've removed the Ram from the

01:41:39 ◼   ► package and also remove the GPU, there's plenty. And, and again,

01:41:43 ◼   ► look at the cooling solution that the Mac pro case has available to it.

01:41:47 ◼   ► It has so much available cooling for a chip that so far barely needs a wimpy

01:41:52 ◼   ► laptop fan. You could make,

01:41:55 ◼   ► I feel like you can make the M one four times as big ditch the,

01:41:58 ◼   ► the Ram off the package and dip digit GPU off the package and still not put a

01:42:02 ◼   ► dent in this thing, CPU cooling capacity. Right?

01:42:05 ◼   ► So I think lots of things are possible

01:42:09 ◼   ► depending on how far Apple is willing to go.

01:42:12 ◼   ► And the Mac pro is the machine that has the least

01:42:16 ◼   ► ambiguity about what they're going to do,

01:42:18 ◼   ► because we know his whole purpose in life is to have huge capacity.

01:42:22 ◼   ► So whatever it takes to get huge capacity, Apple will do that.

01:42:25 ◼   ► Right? The real machines in my question are what do they do with the iMac?

01:42:28 ◼   ► Because the iMac, you could squint and say, well,

01:42:30 ◼   ► whatever you did for the 16 inch Mac pro do it for the iMac. But then again,

01:42:33 ◼   ► what about the iMac pro? Well, then maybe we can do something in between.

01:42:36 ◼   ► It's those types of machines where I'm interested in how they're going to make

01:42:39 ◼   ► that trade off between just do what you did in the laptops,

01:42:43 ◼   ► but more and faster.

01:42:44 ◼   ► You could end up running into limits there because if you want a GPU that is

01:42:49 ◼   ► bigger than you could fit in any laptop, iMac pro can support that.

01:42:52 ◼   ► It's got the cooling solution for it, but you do eventually run into the limit of

01:42:57 ◼   ► the quote unquote integrated GPU. You know, as far as we know,

01:43:00 ◼   ► like what's the biggest GPU we've ever seen shoved onto a system on a chip.

01:43:03 ◼   ► It's probably this current generation of consoles and it's pretty darn big,

01:43:07 ◼   ► but you can get bigger and faster GPUs discreet, no problem whatsoever.

01:43:11 ◼   ► So that's, that's really real question is,

01:43:14 ◼   ► and that's just a trade off that Apple is going to make.

01:43:16 ◼   ► But I think on the top end Mac pro,

01:43:20 ◼   ► there is no reason that they can't have discreet GPUs,

01:43:23 ◼   ► discreet Ram and still have amazing performance. It's just a,

01:43:27 ◼   ► it's just a question of money. How do you make a very high bandwidth,

01:43:31 ◼   ► a high speed Ram interfaces?

01:43:34 ◼   ► You give the chip more power and you spend more money on intertext and you

01:43:38 ◼   ► charge more money for the product. And Apple is really good at that.

01:43:40 ◼   ► So I have some confidence that they will,

01:43:43 ◼   ► they will sort this out for the Mac pro, especially if you consider, well,

01:43:47 ◼   ► I mean, again, don't try to associate price with parts in the Mac pro.

01:43:50 ◼   ► It won't add up for you,

01:43:51 ◼   ► but those zions they buy from Intel are ridiculously expensive and they don't

01:43:56 ◼   ► have any GPU in them that Apple is using in their Mac pros.

01:43:59 ◼   ► They don't have a neural engine inside them that have any of that stuff in there.

01:44:02 ◼   ► Right. And they cost so much money. Apple, you know, we'll get,

01:44:06 ◼   ► Apple will get its M whatevers for the Mac pro.

01:44:08 ◼   ► They'll get them at costs because they make them themselves, right?

01:44:10 ◼   ► They just have to pay Taiwan semiconductor margin or whatever.

01:44:13 ◼   ► So I'm hoping that whatever money they save by making the system on a chip

01:44:18 ◼   ► themselves to the Mac pro, they can spend that on more exotic,

01:44:22 ◼   ► faster memory interfaces and so on.

01:44:25 ◼   ► And it will end up being a really good machine.

01:44:27 ◼   ► Yeah. You've kind of convinced me a little bit because I was, I looked up the,

01:44:31 ◼   ► the dye size numbers and I don't think anyone has measured the M one dye size

01:44:36 ◼   ► yet, but that's what I couldn't find. Yeah.

01:44:38 ◼   ► Like the eight 12 X for reference is about 128 square millimeters.

01:44:42 ◼   ► And to put that in context, like the high Xeon core count,

01:44:45 ◼   ► like the 18 core Xeons are in the like 500 square square millimeter range.

01:44:50 ◼   ► So there's certainly a long way to go between, you know,

01:44:57 ◼   ► eight 12 X size, which is probably a little bit smaller than the M one,

01:44:59 ◼   ► but probably not that much smaller. You know,

01:45:02 ◼   ► at 128 ish and the Xeons at like 500 ish.

01:45:06 ◼   ► And remember they don't like, it's not like you take the M one and say, well,

01:45:09 ◼   ► we'll just stamp out four of those because if you're multiplying the cores,

01:45:12 ◼   ► the CPU cores are not a hundred percent of the chip.

01:45:15 ◼   ► You're not adding a second neural engine, right?

01:45:17 ◼   ► You're not adding a second image processing engine.

01:45:19 ◼   ► Maybe you make those a little bit bigger,

01:45:21 ◼   ► but you're just multiplying the core.

01:45:23 ◼   ► So multiplying the cores does not multiply the size of the, you know,

01:45:27 ◼   ► the dye area by the same amount. So if you had four times the dye area,

01:45:31 ◼   ► you could,

01:45:32 ◼   ► you can get more than four times the cores or you can get twice as many cores and

01:45:36 ◼   ► twice as many GPU cores and still not be at a four X scaling.

01:45:39 ◼   ► So there's breathing room there. And mostly what I think about is again,

01:45:42 ◼   ► the cooling capacity of this gigantic computer, look at what it's cooling. Now,

01:45:46 ◼   ► look how many Watts does the on takes,

01:45:48 ◼   ► what how many Watts is the 28 core Xeon take to try many,

01:45:52 ◼   ► try to use that power budget,

01:45:54 ◼   ► try to use that power budget with Apple's processor.

01:45:57 ◼   ► How big would you have to make this die?

01:45:59 ◼   ► How high would you have to clock it to absorb that power budget?

01:46:03 ◼   ► And that's assuming that Apple even wants to absorb that power budget.

01:46:06 ◼   ► Like there's so much headroom that they can basically like, I was,

01:46:09 ◼   ► I was thinking this idly the other day is like,

01:46:11 ◼   ► if you took the M1 and you just put it straight into the Mac pros chassis and

01:46:16 ◼   ► just overclocked it until you couldn't cool it anymore,

01:46:19 ◼   ► how fast could you make it go? And it's not,

01:46:21 ◼   ► I know cooling is not the only limiting factor on clocking.

01:46:23 ◼   ► Eventually you're running to design constraints,

01:46:25 ◼   ► having to do with pipeline depth and so on and so forth. But like,

01:46:27 ◼   ► I feel like they have a lot of headroom to make faster computers once you can

01:46:33 ◼   ► plug them in and put a huge fan on top of it.

01:46:36 ◼   ► And that's what makes me fairly confident that technologically speaking,

01:46:39 ◼   ► there's nothing stopping them from making an amazing Mac pro.

01:46:41 ◼   ► The only thing that would be stopping them is like budgetary considerations.

01:46:45 ◼   ► And we've talked to those before,

01:46:47 ◼   ► like how exotic does Apple want to make things for the Mac so far?

01:46:51 ◼   ► The answer is not that exotic because if you squint these,

01:46:54 ◼   ► they look a lot like an a 14 X,

01:46:56 ◼   ► which is the exactly the appropriate thing to do for the machines they've

01:46:59 ◼   ► introduced.

01:46:59 ◼   ► But Apple has not yet proven that they're willing to invest a huge amount of

01:47:04 ◼   ► money to make a design that is radically different than, uh,

01:47:08 ◼   ► than what they use their iOS devices.

01:47:10 ◼   ► There have been a couple of interviews where Apple people have been insisting,

01:47:12 ◼   ► no, no, this isn't just an eight 14 X. You see,

01:47:14 ◼   ► we had to do X and Y and Z for the Mac.

01:47:16 ◼   ► And mostly they refer to things like expected texture formats for the Mac

01:47:20 ◼   ► operating system and stuff like that. And that's all true.

01:47:23 ◼   ► Like it's not just a straight up,

01:47:25 ◼   ► like we just took an eight 14 X and put it into a Mac. But the,

01:47:28 ◼   ► but the tweaks are minor compared to the idea of like we made a whole new

01:47:32 ◼   ► processor core or, you know, we, we invested, uh,

01:47:36 ◼   ► in getting the GPU, uh, you know, the whole discrete GPU interface.

01:47:41 ◼   ► And by the way, you can have a discrete GPU with still,

01:47:44 ◼   ► still have a unified memory architecture.

01:47:45 ◼   ► You can have the Ram off chip and still have a unified memory architecture.

01:47:48 ◼   ► Cause unified memory architecture just means no separate VRAM,

01:47:51 ◼   ► the GPU and the CPU all at the same pool of memory. You can,

01:47:54 ◼   ► there's no reason you can't continue to do that no matter where you put the GPU,

01:47:57 ◼   ► no matter where you put the Ram.

01:47:58 ◼   ► It just gets much more difficult and expensive and takes an investment because

01:48:02 ◼   ► iOS devices and the M1 work, nothing like that. So you'd have to invest in saying,

01:48:07 ◼   ► okay, I'm going to make a chip that supports this, you know,

01:48:10 ◼   ► a huge number of PCI express lanes and has more Thunderbolt connectivity and can

01:48:14 ◼   ► drive umpteen 6k displays and knows how to talk to the external

01:48:19 ◼   ► discrete Apple GPU and shares its fast Ram pool with the CPU.

01:48:23 ◼   ► And it's all soldered to the board and you have to pick when you buy,

01:48:25 ◼   ► whether you want the one terabyte configuration or the five 12 or the 64,

01:48:29 ◼   ► you know, whatever they come up with. Right. Um, or they could just down the capacity.

01:48:33 ◼   ► Like it's not outside the realm of possibility that the first ARM based Mac pro

01:48:36 ◼   ► does not support 1.4 terabytes of Ram.

01:48:38 ◼   ► And Apple just has a story for you about how it doesn't need it.

01:48:40 ◼   ► I can see that happening too because it's all soldered to the board and the

01:48:44 ◼   ► biggest configure they have is a 768 and yada yada. Like,

01:48:47 ◼   ► but that would still be an amazing machine. So yeah.

01:48:50 ◼   ► And as we all suspect,

01:48:52 ◼   ► like that will probably be the last machine they introduced. Uh,

01:48:56 ◼   ► and in the meantime, we will get to see how much of an appetite Apple has to,

01:48:59 ◼   ► to crank this thing up and to make more exact designs with things like the iMac,

01:49:04 ◼   ► which I think is the first point where we could reasonably assume that the

01:49:07 ◼   ► machine could support something other than an M one style thing

01:49:12 ◼   ► where everything is in the same place. It's just bigger.

01:49:14 ◼   ► I think they're going to do that all the way through the whole laptops.

01:49:16 ◼   ► Then the iMac is the first and maybe just the iMac pro if they have one is the

01:49:20 ◼   ► first opportunity where they probably should consider doing something

01:49:25 ◼   ► other than exactly the same thing they've done for the laptops. So bigger.

01:49:28 ◼   ► Marco, how, how much Ram did your prior Mac book,

01:49:34 ◼   ► I guess your Mac book pro, how much Ram does that have? Is it 32 gigs? Nope.

01:49:38 ◼   ► 16. I've been getting 16 for awhile. Oh, okay. So wait,

01:49:42 ◼   ► I think something actually don't know. Wait a minute.

01:49:44 ◼   ► Whatever the stock configuration is at the slightly higher than base level.

01:49:51 ◼   ► Let me see. I don't even move on. I'll answer it in a second. All right.

01:49:55 ◼   ► So the reason I bring this up is because I,

01:49:58 ◼   ► how do I not know?

01:49:59 ◼   ► Because it ultimately doesn't really matter. I agree that,

01:50:04 ◼   ► especially the Mac pro probably needs a

01:50:08 ◼   ► crap ton. Okay.

01:50:10 ◼   ► It probably needs a metric crap ton of Ram for certain use cases. You know,

01:50:14 ◼   ► not most people, certainly not the three of us, but in certain use cases,

01:50:18 ◼   ► you probably do want to have a metric crap ton of Ram in the Mac pro,

01:50:22 ◼   ► but for everyone except those, you know, odd balls,

01:50:26 ◼   ► I keep thinking about how I heard a lot of people like me

01:50:32 ◼   ► complaining about 16 gigs of Ram as the maximum in these new M one max,

01:50:36 ◼   ► but I haven't heard anyone who has had one and used one

01:50:42 ◼   ► really complain about it not having enough Ram,

01:50:45 ◼   ► which has many words to say,

01:50:48 ◼   ► I don't know if we should really judge them the same, these M one max or is

01:50:52 ◼   ► Apple Silicon max.

01:50:53 ◼   ► I don't know if we should judge them the same way we judge Intel max.

01:50:56 ◼   ► Cause my, my 13 inch MacBook pros, 32 gigs of Ram.

01:50:59 ◼   ► And I think that that's probably as low as I would want to go on an Intel

01:51:02 ◼   ► machine in 2020 or 2021 is we're almost there,

01:51:04 ◼   ► but on a Apple Silicon Mac,

01:51:07 ◼   ► I would probably entertain 16 now that I know a little more.

01:51:12 ◼   ► So I'm going to push back on this.

01:51:13 ◼   ► I still see this conversation happening and I tried to wave people off from her

01:51:16 ◼   ► last time and I'll try again.

01:51:18 ◼   ► The architecture and the unified memory thing does not make Mac iOS

01:51:24 ◼   ► use less memory. Oh no, totally. Now the code page,

01:51:29 ◼   ► you know, the,

01:51:29 ◼   ► the code segment of executables may be different size on arm versus

01:51:34 ◼   ► x86.

01:51:35 ◼   ► Although I wouldn't necessarily imagine that it will be smaller on arm because

01:51:38 ◼   ► x86 has variable with instructions and arm I believe does not.

01:51:41 ◼   ► So that actually can make the arm executables bigger. But anyway, um, well,

01:51:44 ◼   ► there is, there is the texture format thing that,

01:51:46 ◼   ► that often needs like multiple copies of video memory,

01:51:51 ◼   ► uh, on Intel machines. But you know, that's, it's still not going to be everything.

01:51:55 ◼   ► I'm not saying they're exactly the same cause they're not,

01:51:58 ◼   ► but they vary in ways that are not like, Oh, well I could get away with, with a,

01:52:02 ◼   ► uh, an eight gig arm Mac, but I would require a 16 gig Intel. No,

01:52:06 ◼   ► there is nothing with the architecture.

01:52:08 ◼   ► There's nothing with anything that has changed in the hardware or the operating

01:52:12 ◼   ► system that makes that a reality. What has changed is, you know,

01:52:16 ◼   ► three times the memory bandwidth and just faster overall. Um, and so,

01:52:21 ◼   ► and you know, the SSD and the air got twice as fast.

01:52:24 ◼   ► All that can contribute to make swapping more tolerable. But in the end,

01:52:28 ◼   ► if you need to have some thing that's bigger than eight gigabytes in,

01:52:32 ◼   ► in working memory at the same time and on,

01:52:34 ◼   ► you get a machine that they can't support that and it has to swap,

01:52:37 ◼   ► it's going to be slower,

01:52:38 ◼   ► slower in a way that you will notice and feel sad about.

01:52:41 ◼   ► It's just that most people don't have those kinds of workloads.

01:52:43 ◼   ► The closest people have is they open lots and lots of apps and the ones they

01:52:47 ◼   ► don't use get paged out and they switched to them and they get paged back in.

01:52:49 ◼   ► Right. But once they're in there, they're working okay. Right. So there is,

01:52:53 ◼   ► there's nothing about like, you know, the,

01:52:57 ◼   ► the changed arm that is making it so suddenly you can get away with half as much

01:53:01 ◼   ► Ram.

01:53:01 ◼   ► And I would tell people that the trend always goes that you need more Ram in the

01:53:06 ◼   ► future, not less.

01:53:07 ◼   ► So don't certainly don't skimp on Ram by getting an arm based Mac that has less

01:53:11 ◼   ► Ram than you, than you're currently using that you think you need.