514: My Immense Softness


00:00:00   All right, I added some brief to follow up that should only take a moment the Apple dropping webkit

00:00:06   I think it's worth you trying to do the preflight. Is that what's happening here? Yes. What are you doing? Oh, that's a that's a power move

00:00:12   No, I'm putting a stop to that. Okay. Sorry dad. I do the preflight. What's next Marco writing the show notes?

00:00:20   Now I'm now the official the official preflight will begin

00:00:28   (clears throat)

00:00:30   (beeping)

00:00:32   - Get the bad feeling all of our good content

00:00:33   happened before we went live.

00:00:35   - Don't worry, I'm using some of that for the pre-show

00:00:37   in the release version.

00:00:39   - All right, good deal.

00:00:39   - Yeah, bootleg people, go listen to the release version

00:00:41   of like the first five minutes

00:00:42   and you'll hear everything that just happened.

00:00:44   Not even, 45 seconds.

00:00:46   Yeah, listen to the modem sound.

00:00:47   - How many people, I guess in our audience it's fine,

00:00:50   but we always refer to it as the modem sound.

00:00:53   Just the number of people who know what that is

00:00:54   or will recognize it when they hear it

00:00:56   or know what we're talking about just shrinks every day.

00:00:58   - Yep. - Yeah, but I think among our,

00:01:00   I think it's shrinking in the world,

00:01:02   but I think in our audience it's staying about the same,

00:01:04   possibly even growing.

00:01:06   - We need some more period piece movies

00:01:08   that capture that little slice of time

00:01:11   when the internet was a thing,

00:01:13   but most people were getting onto it with modems.

00:01:15   Like that was not a very long period of time,

00:01:17   but for the people who lived through it,

00:01:18   it was significant and no one has really done

00:01:20   like a period piece, I mean, setting aside like WarGames

00:01:22   and the Acoustic Coupler and all that,

00:01:24   I'm talking about the like, you know,

00:01:26   2,400 bought and on.

00:01:28   Normal people are on the internet,

00:01:29   but they're using modems to do an era.

00:01:31   - The problem is when you make a movie about some time,

00:01:34   you know, three or four decades ago, whenever,

00:01:36   like, first of all, like, you know,

00:01:37   you gotta figure like, are the mid 90s or early 90s

00:01:41   when this was really like becoming a big thing,

00:01:44   is that time even cool enough to make a movie about yet?

00:01:46   And at some point, it might be, right?

00:01:48   - Oh, for sure.

00:01:49   It's definitely, because I feel like the 80s are fading

00:01:52   as we '80s kids get too old to care anymore,

00:01:55   and I feel like '90s is very in.

00:01:57   - Probably, however, the part of culture

00:02:00   that is represented when you do a look-back movie

00:02:02   by a few decades is what the cool kids were doing

00:02:04   at that time, and I assure you, no one hearing that sound

00:02:07   was a cool kid, 'cause I was that person,

00:02:09   and there is no way I was a cool kid.

00:02:11   - It doesn't have to be a cool kid's thing.

00:02:12   Again, I will point to "War Games."

00:02:14   Sometimes you make a movie intentionally

00:02:15   about the nerds or whatever.

00:02:16   - Yeah, but they're never the nerds.

00:02:18   They're never the, it's always like,

00:02:20   oh, you take the kid's glasses off,

00:02:21   and all of a sudden he's hot.

00:02:22   Like it's, you know, it's never, it's never the nerds.

00:02:25   - What movies are you watching?

00:02:27   - She's All That.

00:02:28   Actually isn't there a He's All That now,

00:02:29   but that's brand new.

00:02:30   I just read, and I'm never gonna be able to find it,

00:02:33   but I just read in the last like 48 hours,

00:02:34   live streams, Jon, live streams.

00:02:36   Something about how, and I'm gonna butcher it,

00:02:38   but it was, the premise was like, the movie Superbad,

00:02:42   which probably does not hold up at all,

00:02:44   but my recollection of it is very fond.

00:02:47   Superbad was in the brief window of time

00:02:49   right before just everyone was carrying phones and text messaging and so on and

00:02:53   so forth. And, and the, the, the, the,

00:02:57   the thought was that this was the last movie of, of its ilk,

00:03:02   right? The last time you can have a movie where not being able to communicate

00:03:06   is the crux of the movie. And.

00:03:10   Oh no, they do that all the time now.

00:03:11   They just have increasingly ridiculous contrivances to make that the case.

00:03:14   Oh, I've got no signal. My phone is gone. Or there's an EMP or whatever.

00:03:18   It's always some reason.

00:03:19   Yeah, now it's just all zombies. The zombies took out the cell towers first for some reason.

00:03:23   Right, exactly. But you know what I'm saying, that, you know, that alleges this person on,

00:03:28   like, Twitter, Mastodon or whatever.

00:03:29   I was watching some 80s anime with my son over his break, and it's a movie set in, I

00:03:36   mean, the future of the 80s, so, like, they think it starts, I don't know, in the 2000s

00:03:40   or something, but anyway, everything's all futuristic, right? There's giant robots, you

00:03:44   you know, the whole nine yards,

00:03:46   and they have really cool futuristic-looking phones,

00:03:49   but they are in phone booths.

00:03:52   So whenever people need to communicate,

00:03:53   they go to a phone booth and use this amazing

00:03:55   futuristic phone that like, you know,

00:03:57   scans their retina or responds to voice commands,

00:03:59   but it's in a booth and they're huge.

00:04:01   Like at their desk, they have a really cool-looking

00:04:04   space futuri kind of phone,

00:04:06   and it's like the size of a shoebox on their desk

00:04:08   with a cord connected to it, right?

00:04:10   'Cause they can imagine the future,

00:04:11   They can imagine the future with giant robots, right?

00:04:14   And amazing technology and space travel

00:04:16   and all sorts of things and aliens,

00:04:18   but everyone having a phone, no.

00:04:19   So there are plots that involve someone trying

00:04:22   to call someone and they can't get them

00:04:24   because they're not by a pay phone or at their desk.

00:04:26   - Those were the days.

00:04:28   I just want an entire documentary on the battle

00:04:31   between X2 and K56 Flex, that's what I want.

00:04:35   - There wasn't much of a battle.

00:04:36   Like X2 kicked their butts, like that was the battle.

00:04:39   Like X2 won by a lot.

00:04:41   - And also the people had the K56 flex things,

00:04:44   it's not like they couldn't, like ISP supported that enough

00:04:47   like you know if you bought it and you could use it,

00:04:49   it's not like it was a bad purchase,

00:04:50   it wasn't like you got, what was it called,

00:04:52   like Divix Drive or something.

00:04:54   - Oh yeah. (laughs)

00:04:55   - Well no, it was a bad purchase after a while.

00:04:57   Initially, 'cause the thing is like,

00:05:00   on the receiving end, until there was that like V92

00:05:03   that unified them, on the receiving end,

00:05:05   they had to choose like certain phone lines would be X2

00:05:09   and the other phone lines would be K56 Flex,

00:05:11   and you'd have to call the right one

00:05:13   to match the format of your modem.

00:05:14   And since X2 was way more popular from the get-go,

00:05:18   what ended up happening was the ISPs and BBSs

00:05:22   would have way more X2 lines than K56 Flex lines.

00:05:25   - By then, your modem probably broke anyway,

00:05:27   so get a new one, don't worry about it.

00:05:30   - We are brought to you this week by Nebula,

00:05:34   a streaming service created and owned

00:05:35   by some of the internet's most thoughtful creators,

00:05:38   offering full length and ad free videos

00:05:41   earlier than anywhere else as well as Nebula exclusives.

00:05:44   These are fully produced Nebula originals

00:05:47   like Anita Sarkeesian's That Time When or Impact,

00:05:50   Rene Ritchie's Retrospective on the Launch of the iPhone.

00:05:53   They also have Nebula classes featuring all kinds

00:05:56   of online courses led by creators,

00:05:58   including an introduction to programming called

00:05:59   What is Code from creator and NYU professor Daniel Schiffman

00:06:03   and tons of bonus and exclusive content.

00:06:06   You know, I first heard about Nebula

00:06:08   because I noticed that a large portion of the YouTubers

00:06:12   that I would watch, at the end they would say,

00:06:14   "Hey, go find me on Nebula."

00:06:15   I make content there and all my best stuff goes there.

00:06:18   It's there first, it's there ad-free.

00:06:20   And this is people like Wendover Productions,

00:06:22   and Half as Interesting of course,

00:06:24   Georgia Dow, Real Engineering, Practical Engineering,

00:06:27   kind of all the stuff that I love to watch on YouTube,

00:06:30   really, they all were on Nebula.

00:06:32   And so I went there and it's great

00:06:34   because you get all this bonus content,

00:06:36   You get, of course, everything's ad-free.

00:06:38   It's exactly what you'd want as a nerd

00:06:40   out of a video site, a player experience.

00:06:42   They're on all your devices and everything.

00:06:44   It was all these creators I was already into

00:06:46   and then I got to discover new creators

00:06:48   that were similar quality and producing longer things.

00:06:53   It's a great service.

00:06:54   I strongly recommend, if you're nerdy enough

00:06:57   to listen to our show, if you don't already know

00:06:59   these producers, these content creators,

00:07:01   you're gonna love their stuff

00:07:02   and Nebula is a great way to get it.

00:07:04   It's a great deal.

00:07:05   So you can go there right now.

00:07:06   it's normally, I mean, look, normally it's a good deal.

00:07:08   Five bucks a month or 50 bucks a year.

00:07:11   But we have a special, you can get a year of Nebula

00:07:13   for just 40 bucks at nebula.tv/atp.

00:07:18   So once again, normally five bucks a month

00:07:20   or 50 bucks a year, that's a great price,

00:07:21   but it's even less now, 40 bucks for a whole year of Nebula

00:07:25   at nebula.tv/atp.

00:07:28   Thank you so much to Nebula for sponsoring our show.

00:07:30   - I have a little bit of follow up

00:07:35   with regard to my Docker adventures.

00:07:37   If you recall, last episode in the after show,

00:07:39   I was talking about how I couldn't get Docker

00:07:40   working properly on my Mac Mini.

00:07:42   A lot of people reached out to say,

00:07:44   "Hey, it works on my desk."

00:07:45   And I'm glad, I really am,

00:07:47   but it doesn't work on my desk.

00:07:48   And so a handful of people,

00:07:50   and I don't have names in front of me, I'm sorry,

00:07:51   but a handful of people recommended Colima,

00:07:53   which is short for containers on Linux on Mac.

00:07:58   And so the idea is it's what John was suggesting.

00:08:01   It's like a Linux VM in which you run Docker

00:08:04   And so as far as Docker's concerned, you're--

00:08:06   - I mean, that's what Docker desktop is too, right?

00:08:08   I was suggesting you use VMware,

00:08:10   like just run something that doesn't know anything

00:08:12   about Docker, but just says,

00:08:12   "Hey, I'm going to run Linux on your Mac."

00:08:14   And then within that Linux VM, run Docker.

00:08:17   - And that's fair, but one way,

00:08:19   I honestly, I don't care enough to understand

00:08:22   the hyper specifics about it,

00:08:23   but suffice to say, it makes,

00:08:25   it's an alternate way of running Docker.

00:08:27   And I had a bunch of problems with that

00:08:29   with regard to network shares that are unimportant,

00:08:31   but as of just a couple of hours ago,

00:08:33   I think I have that all squared away now

00:08:35   and that seems to be working.

00:08:36   So now I have temporarily powered off my Docker containers

00:08:40   on my Raspberry Pi.

00:08:41   I have migrated everything to the Mac mini

00:08:44   and I'm gonna see how that goes for the next few days.

00:08:46   And I think and hope it's working

00:08:48   and hopefully that'll go pretty well.

00:08:51   Now the motivation for doing this is,

00:08:53   even though the Raspberry Pi seems to be handling

00:08:56   all four of the containers, no problem.

00:08:59   And I'm quite surprised by that

00:09:01   'cause I think it's a two gig RAM Raspberry Pi.

00:09:04   And I was saying, you know,

00:09:05   I really want an eight gig Raspberry Pi.

00:09:07   I don't think I need it, I really don't.

00:09:08   But nevertheless, there's not a lot of space

00:09:10   in that Raspberry Pi.

00:09:11   I don't wanna like thrash the SD card

00:09:13   if one of these containers is downloading something.

00:09:16   And so I wanted to move it to the Mac Mini.

00:09:19   And in my initial tests on the Mac Mini,

00:09:22   if I have one of these containers download something

00:09:24   on the Raspberry Pi, it goes at like 15,

00:09:26   maybe 20 megabytes a second.

00:09:27   On the Mac Mini, I was kissing 90 megabytes a second.

00:09:30   that's bytes, not bits, mind you.

00:09:32   So I was basically maxing out my gigabit ether internet line

00:09:36   which is pretty freaking cool.

00:09:38   So that's part of the reason I wanted to go to Mac mini.

00:09:41   And so far, knock on wood,

00:09:43   everything seems to be working now.

00:09:44   So if you're in a similar boat, I'm so sorry,

00:09:46   but Colima, and we'll put a link in the show notes.

00:09:49   Tell me, Jon, Apple's considering not requiring

00:09:52   all of us to use WebKit, what?

00:09:54   - Anytime Apple consider, quote unquote,

00:09:56   considers anything like that,

00:09:57   it's probably because the government is making them.

00:09:59   So this is part of the Germin story about the Digital Markets Act and he claims that

00:10:05   that's one of the things that's on the table to comply with the DMA is stopping the requirement

00:10:12   that every web browser that is on iOS must use Apple's WebKit engine.

00:10:16   If you don't know, you can get Chrome for your phone but Chrome is running WebKit under

00:10:21   the covers, not Google's fork of WebKit that is called Blink.

00:10:25   It used to be that both Chrome and Safari were built on WebKit, but those days are gone

00:10:30   now when Google went off and forked it.

00:10:33   Anyway, this is something that I'm sure browser makers have wanted for a while.

00:10:38   We've talked in the past about all the various security reasons why Apple wouldn't want this,

00:10:43   but there are other reasons why Apple wouldn't want this.

00:10:44   Setting aside, anything having to do with security and just-in-time compilers and all

00:10:48   sorts of exploits that can happen in web browsers, this is sort of the thing that a lot of developers

00:10:54   who were angry about, web developers were angry about

00:10:57   when Apple was too slow in their opinion

00:11:00   in supporting various web standards on iOS.

00:11:04   - Which by the way, that was always like

00:11:07   a really BSC classification of what was actually happening.

00:11:11   Which was basically it's like, you know,

00:11:12   the web people who didn't like the take off of apps

00:11:16   would propose a quote standard saying,

00:11:18   all right, now all web browsers have to have,

00:11:21   you know, insert app capability X notifications,

00:11:24   background execution, workers, all this stuff that Apple was

00:11:28   like, we don't agree that's what everybody should have,

00:11:31   because that's going to make everything worse on our phones.

00:11:34   And so it was way more complicated

00:11:37   than Apple being quote, "slow."

00:11:39   Well, it's some things that were slow,

00:11:41   like even just simple stuff like features they did support,

00:11:43   but they weren't up to date with the latest changes in them.

00:11:46   It wasn't always a feature that Apple disagreed with.

00:11:49   But that's changed a lot lately.

00:11:50   I think Safari and the WebKit engine

00:11:52   has really picked up in web standards.

00:11:54   But the reason it's relevant is what you were getting at,

00:11:57   that if you are dissatisfied with the one and only app

00:12:00   store that exists on the phone, Apple has always

00:12:04   had a way for you to sort of, well,

00:12:06   why don't you just ship it as a web app?

00:12:07   Make it a web page.

00:12:08   You can even put web apps on your home screen.

00:12:10   That's a feature that Apple introduced.

00:12:12   I think it was-- was it part of the suite solution,

00:12:14   or did it come later?

00:12:15   I forget.

00:12:15   Yeah, it was really early on.

00:12:17   I think that was iOS 2.0, wasn't it?

00:12:19   Even before 2.0.

00:12:20   as well as features that people don't even know exist

00:12:24   and may even be using without knowing it.

00:12:26   So it's kind of, you know,

00:12:27   it's almost like a different Apple made this,

00:12:28   like, oh, you mean I can have an icon on my home screen

00:12:30   that really just launches a webpage,

00:12:32   but I don't know that 'cause it just looks, anyway.

00:12:34   That feature still exists.

00:12:35   Apple could take that away at any time if they wanted.

00:12:37   But anyway, if you're dissatisfied with the App Store,

00:12:40   but there's some feature that you would like to have

00:12:42   that you can't implement in WebKit, essentially,

00:12:46   you're like, man, if only I had Chrome,

00:12:48   the actual Chrome, you know, Blink-based Chrome

00:12:50   on iOS, I could do this other thing with my app.

00:12:53   And setting aside things that integrate with the OS,

00:12:55   like, oh, I want to have push notifications, stuff like that,

00:12:57   just stuff that happens just in that web page,

00:13:00   something having to do with service workers,

00:13:02   or I mean, I guess this impacts the OS as well,

00:13:05   but like our IndexedDB, or just there's some web feature

00:13:08   that Apple doesn't have yet that you think, if I had this,

00:13:10   I could make a web-based app that would be really great.

00:13:15   If Apple lifts this restriction and says browser engines

00:13:18   the WebKit are allowed, it is plausible that a motivated group could decide, "Okay, we

00:13:24   don't have to wait for Apple to implement these new features, these new WebGL features

00:13:29   or something that we really need for our cool web-based application," because Blink has

00:13:33   already implemented it, and so we'll use the Blink-based Google Chrome on iOS to run our

00:13:39   cool application.

00:13:40   Like, it's kind of a -- it's not a backdoor, it's not even a side door, it's just sort

00:13:44   of like a sideshow.

00:13:46   Being divorced from WebKit, saying,

00:13:49   "We're not dependent on Apple implementing these features,"

00:13:52   we're dependent on Google now instead, right?

00:13:54   But say you are Google, for instance,

00:13:56   it really would help them

00:13:57   if they had a web-based application

00:13:58   that they wanted to do cooler stuff with.

00:14:01   So I don't know if this will happen,

00:14:03   but I think it's probably a good thing,

00:14:05   and I think it's kind of long since past the time

00:14:08   that Apple can say, "Nope, it's just too unsafe.

00:14:11   "We can't handle it."

00:14:11   It's like, "Find a way, Apple.

00:14:12   "You found a way to do third-party keyboards

00:14:14   "for crying out loud.

00:14:15   about things that are potentially unsafe.

00:14:16   It's how many years into the iPhone.

00:14:18   We're not asking on year two of the iPhone

00:14:20   to allow third-party browser engines,

00:14:21   but 15 years in or whatever we are now,

00:14:24   I think it's the technical reasons

00:14:26   that that is difficult to remain,

00:14:28   but it's not, there shouldn't be barriers.

00:14:29   I think it is time for third-party browsing engines

00:14:32   to be available, and I actually hope it does happen,

00:14:34   because that's a pretty clean thing to happen.

00:14:36   It's like a yes/no, it's a policy decision.

00:14:38   Apple doesn't have to do anything.

00:14:39   They don't have to waste any time on this

00:14:42   other than sort of making sure everything is secured.

00:14:44   And, you know, frankly speaking, it's not like there are a million web browser engines

00:14:49   out there waiting to burst through the door. There's a small number. There's Firefox,

00:14:52   which whatever their engine is called these days, and there's Blink, which is a WebKit

00:14:55   fork. And then there's little hobby projects. So it would be nice to see those other engines

00:15:00   on your iPads and iPhones.

00:15:03   Well, but there is one complex part of it, though, that I think is a significantly complex

00:15:07   part, is that this is not just a policy decision. There's a technical side of this, which

00:15:11   is that WebKit has privileged execution abilities

00:15:15   that other apps on iOS don't have,

00:15:18   where it can do things like just-in-time compilation

00:15:21   of JavaScript and then execute that at full speed.

00:15:23   And I forget the details of exactly how all this works,

00:15:26   but basically, other apps can't do that.

00:15:28   Safari does it.

00:15:30   Like UI WebView and WKWebView,

00:15:32   or UI WebView does not do it.

00:15:34   WKWebView, I think, does, 'cause it can be out of process.

00:15:36   Anyway, whole thing.

00:15:36   But the Safari rendering engine on iOS

00:15:40   has lower level access to how it can compile

00:15:44   and execute JavaScript.

00:15:45   So JavaScript will run faster on that

00:15:48   than it would run on any third party app

00:15:50   that doesn't have that ability,

00:15:51   which is all third party apps.

00:15:52   So even if they change the policy

00:15:55   that says you can't have other web browser engines,

00:15:57   if they actually allow Blink or whatever else in an app,

00:16:02   but don't grant it this ability,

00:16:03   and there's lots of reasons why they wouldn't,

00:16:05   'cause I can't imagine Apple wanting to open up

00:16:07   that security can of worms

00:16:09   with just letting Google's code do it on their devices.

00:16:12   There's no way in heck they're gonna allow that.

00:16:14   So the result we're gonna get if they do this

00:16:17   is browsers and other products that are based

00:16:20   on those other browsing engines

00:16:22   will have slower JavaScript speeds than Safari.

00:16:24   And then people will say that's unfair.

00:16:26   But there actually is this pretty significant

00:16:28   security reason why not to do that.

00:16:30   So I think it wouldn't surprise me

00:16:33   if this doesn't go as well as people plan.

00:16:35   And there's a good reason for that

00:16:36   that I think Apple, Apple's going to take a lot of hits for

00:16:40   because people won't fully understand the reason,

00:16:42   but there is a good reason why they wouldn't want to let

00:16:46   any other browser engines perform as quickly

00:16:48   as Safari does, you know, on the backend.

00:16:51   - I mean, I understand the reason,

00:16:52   that's what I was getting at with the security concerns,

00:16:53   but I think at this point, you can't just keep saying,

00:16:56   oh, it's impossible to allow third parties

00:16:57   to have fast JavaScript engines.

00:16:58   Like, fine, it's impossible for the first decade and a half,

00:17:01   eventually you gotta figure it out, right?

00:17:02   It's like, no, it'll never be safe.

00:17:04   Look, if it's safe enough for Apple to do it,

00:17:06   Apple can find a way to allow the two third parties

00:17:09   I'm talking about that have engines

00:17:10   figure out a way to do it safely.

00:17:12   Even if it is simply like, hey, it's safe for us to do it

00:17:14   and occasionally we have exploits,

00:17:16   we're going to carefully vet what you submit

00:17:18   and you're gonna have exploits occasionally too

00:17:20   and you'll patch them and that'll be that.

00:17:21   Like that's what I'm saying.

00:17:23   Like same thing with third party keyboards.

00:17:24   Incredibly dangerous to allow that.

00:17:26   Apple found a way to do it safely.

00:17:27   I think there is a way to do this safely

00:17:30   and it does involve not allowing anyone else to use a JIT,

00:17:33   although I have seen some stuff recently,

00:17:35   Maybe it was in Gecko saying they found a way

00:17:37   to make the JavaScript engine fast without doing JIT.

00:17:40   I don't know.

00:17:41   Anyway, I think they can find a way to make this work

00:17:44   because it's not like they have to let

00:17:46   a million different things,

00:17:47   like they don't have to make a blanket policy.

00:17:48   It's like now just-in-time compilation of anything

00:17:50   is free for all, anyone can do it.

00:17:51   Put it in your casino games for children, we don't care.

00:17:54   It's like no, you know who you're dealing with.

00:17:55   You're dealing with Firefox and you're dealing with Google

00:17:58   and those submissions can get some extra attention

00:18:01   and you can figure out a way to allow them

00:18:03   to have alternate browser engines with this extra,

00:18:06   you know, either they can decide to do it

00:18:07   without just-in-time compilation.

00:18:10   And, again, I think there are ways

00:18:12   to make that reasonably fast.

00:18:13   Or if you really wanted to do a JIT,

00:18:15   find a way to make yourself okay with that in the same way

00:18:20   that you're okay with Safari doing it,

00:18:21   because it's not like Apple's humans are magic

00:18:23   and don't make bugs that cause security problems in WebKit,

00:18:27   and Google just can't do that.

00:18:28   Like, make it work.

00:18:30   - All right, we got some absolutely phenomenal feedback

00:18:34   like a month ago, I don't know, it was a while ago.

00:18:37   And I loved it, but we didn't have a chance

00:18:40   to talk about it for a few weeks

00:18:42   'cause other things were going on.

00:18:43   And I'd like to talk about it now.

00:18:45   It's a little bit on the longer side,

00:18:46   but I think it's worth it.

00:18:47   So Chris writes, "So I guess it's my turn

00:18:49   "to add credence to the theory

00:18:50   "that there is someone in the ATP audience

00:18:52   "with specialized knowledge about any niche topic

00:18:54   "that you stumble into.

00:18:55   "This example, inertial celestial navigation

00:18:58   "or more generally GPS-free navigation

00:19:00   for military applications.

00:19:02   Chris says, "I was a US Navy submarine warfare officer

00:19:05   and have worked extensively with submarine-launched weapons,

00:19:07   including the Trident II D5 missile system.

00:19:09   The SR-71," which is, I think, where this all came from,

00:19:12   as Marco was saying.

00:19:13   Yep, from Blackbird.

00:19:14   Your thing was the Blackbird.

00:19:16   "The SR-71 has an astro-inertial navigation system.

00:19:18   In other words, it had a celestial navigation system,

00:19:21   but it didn't work independently from the inertial navigation

00:19:23   system of the INS.

00:19:25   Many platforms, like planes, submarines, ships, land

00:19:27   vehicles, et cetera, use INSes.

00:19:29   These systems measure acceleration over time and integrate acceleration to calculate velocity,

00:19:34   changes in position, attitude, heading, and other navigational properties.

00:19:38   As John alluded, integration in the calculus sense of acceleration also means the integration

00:19:42   of errors associated with acceleration measurements due to things like ocean currents, atmospheric

00:19:45   prevailing winds, and measurement errors.

00:19:48   The rate that errors accumulate depend on a large number of factors and are probably

00:19:51   beyond the limits of this email.

00:19:53   But as an example, total error includes not only the acceleration sensing error, but errors

00:19:57   caused by moving across the curvature of the Earth, either underwater or in near space,

00:20:00   which is the so-called Schuyler oscillations. And we'll put a link in the show notes.

00:20:04   Over time, inertial systems need to be told where they are because

00:20:08   in position uncertainty tends to grow. So the Position Navigation and Timing,

00:20:12   PNT, game is about reducing position error through better inertial measurement

00:20:16   or limiting and correcting the error that grows from the INS you're using.

00:20:20   Engineers have lots of clever schemes to do this, such as looking at the stars. The SR-71

00:20:23   use stars to estimate position to correct or limit the errors of the INS. Together, they were better

00:20:28   than either component. According to the Aviation Geek Club, their position error had about a 300

00:20:35   foot radius. That's pretty good. And there's a link that Chris provided to the Aviation Geek

00:20:41   Club, which explains a lot of this. And I thought an interesting segment of that link. The ANS works,

00:20:48   this is the navigation system, works by tracking at least two stars at a time listed in an onboard

00:20:52   catalog and with the aid of a chronometer calculates a fix of the sr-71 over the ground

00:20:56   it was programmed before each flight in the aircraft's primary alignment and the flight

00:20:59   plan was recorded on a punch tape that told the aircraft where to go when in turn when to turn

00:21:04   and when to turn the sensors on and off the stars are sighted through a special quartz window located

00:21:09   behind the cockpit and there was a special star tracker that could see stars even in the daylight

00:21:14   how freaking cool it was an interesting story about this uh referencing that the punch tape or

00:21:20   or whatever, I don't know if it's one of the links

00:21:22   we'll put in the show notes, again, live streams,

00:21:24   I don't know where I saw it, maybe it was in a podcast.

00:21:27   So they had two catalogs of star position information,

00:21:31   one for the Northern Hemisphere and one for the Southern,

00:21:33   presumably due to memory limitations

00:21:35   and whatever ancient computer from the 60s

00:21:37   is in this thing, right?

00:21:38   But they couldn't fit the whole world, sorry, right?

00:21:40   And one of the SR-71 pilots, I think it was a podcast,

00:21:43   was telling a story about a mission he had

00:21:45   that flew from the Northern Hemisphere into the Southern,

00:21:48   and he was just like landing just over the equator

00:21:50   in the southern, but they hadn't loaded

00:21:52   the southern hemisphere stars in the thing.

00:21:55   He just had the northern one,

00:21:56   so as soon as he crossed over the equator,

00:21:58   the plane didn't know where it was anymore,

00:22:00   and it was a little bit of a freak out.

00:22:02   Some technology, man.

00:22:04   We can't fit all those stars

00:22:05   and you just get one hemisphere,

00:22:06   and if you fly over it, oh well.

00:22:08   He said that he managed to land fine anyway

00:22:11   because the plain old radio radar type stuff

00:22:15   that leads you to the airport he was landing in,

00:22:18   when you're flying at like 80,000 feet,

00:22:20   you get really good range on that radar.

00:22:22   You can see long distances.

00:22:23   So as soon as he crossed over the equator,

00:22:26   he basically had line of sight practically

00:22:28   onto the place where he was landing

00:22:30   because he was flying so high.

00:22:32   - That's awesome.

00:22:33   - That is absolutely bananas.

00:22:34   And then building on this,

00:22:36   I don't remember where this link came from,

00:22:37   but it turns out that Turnabout is fair play.

00:22:40   So this Twitter user, John McElhone, put a short thread up

00:22:44   and it describes, there's apparently a new US Air Force stealth aircraft

00:22:49   and the US government took a picture of it at night with stars in the

00:22:54   background and posted it on, I don't know, like social media or official forums,

00:22:58   I'm not sure exactly what the genesis of this image was or where it was posted,

00:23:01   but John McElhone

00:23:03   took, via this Twitter thread, took us through

00:23:06   figuring out exactly when and where that picture was taken based on the stars in

00:23:10   the background of the picture, which I think is the coolest thing.

00:23:14   So here the United States government rolls out this top secret. Well,

00:23:18   I guess it's not top secret since they're posting a picture about it,

00:23:20   but you know what I mean? Like they're, they're announcing this new aircraft.

00:23:23   They take a picture of it at night. They have the stars in the background.

00:23:26   I wouldn't have thought anything of it and I,

00:23:28   they didn't think anything of it.

00:23:29   And John McElhone was able to pinpoint within I think a couple of hours and

00:23:34   basically which air force base this thing had its picture taken on.

00:23:38   This is like a 10 tweet thread and it is well worth the two minutes of your time

00:23:42   to read it. It is so cool.

00:23:44   It reminds me of those stalkerous threads where someone will post a photo and then the

00:23:47   rest of the internet will figure out where on earth that photo was taken.

00:23:50   Oh yeah, there's a subreddit for that but I forget what it's called.

00:23:53   That's just called Reddit.

00:23:54   Well, fair, fair.

00:23:55   Based on like the background sidewalk or the house or like it's not even like road signs

00:24:00   they can just tell like you know where that is just because either someone is familiar

00:24:04   with that area or has seen it before or they'll narrow it down based on Google Earth stuff.

00:24:09   If we lived in the fantasy world of the movies where all large institutions are smart about

00:24:14   than individual people, which we don't, the government would have put in that star field

00:24:21   in the background with entirely fake stars to make people think it was on an Air Force

00:24:24   base in Arizona or wherever when really it was in the secret Air Force base that we don't

00:24:28   know about.

00:24:29   Right, right, right.

00:24:30   Anyway, I just thought that was super cool.

00:24:32   And then we also got, and Marco, I think you're best equipped to summarize it, but we got

00:24:36   a lot of feedback about moisturizers, including several people who are like either doctors

00:24:41   or residents that are in dermatology.

00:24:44   So apparently there is nothing that our show cannot reach

00:24:48   and I'm very proud of this,

00:24:49   probably unreasonably proud of this.

00:24:51   - Yeah, so last week's pre-show,

00:24:53   I was talking about my dry hands

00:24:55   and especially the skin around my fingernails

00:24:57   always get all cracked and cut

00:24:58   and I'd have to wear band-aids and it sucks.

00:25:00   And so we heard from so many people,

00:25:05   many of whom either were dermatologists themselves

00:25:09   or they were married to a dermatologist

00:25:11   or they knew a dermatologist,

00:25:13   or they went to a dermatologist,

00:25:15   or they saw a dermatologist once from across the room.

00:25:18   Like, there were so many people with so many connections

00:25:22   to various parts of dermatology expertise.

00:25:25   And everyone kind of agreed that, okay,

00:25:27   to answer my question, are moisturizers real

00:25:30   and do they work?

00:25:31   Turns out, yes.

00:25:32   In my defense, a lot of medicine,

00:25:33   especially over-the-counter stuff,

00:25:35   is not super real and doesn't super work.

00:25:37   So, you know, I think that was a valid question to ask.

00:25:40   Far and away, there were a few consistently

00:25:43   recommended items.

00:25:44   The first most recommended one by far

00:25:47   is something called either O'Keeffe's or O'Keeffe's,

00:25:50   I don't know how this is pronounced.

00:25:52   Anyway, O'Keeffe's Working Hands.

00:25:54   Now, go search Amazon for O'Keeffe's Working Hands.

00:25:58   I have never in all of my life searching Amazon

00:26:02   seen a product with such incredibly high reviews

00:26:06   and so many reviews.

00:26:08   - Oh wow, yeah.

00:26:09   80,000 reviews, 4.7.

00:26:11   34,000 reviews, 4.7.

00:26:13   5,000 reviews, 4.8.

00:26:15   5,000 reviews, 4.7.

00:26:17   This is the most highly rated product

00:26:19   I think I've ever seen on Amazon

00:26:20   that I've ever searched for.

00:26:22   So I thought, all right, I'll give this a shot.

00:26:25   So I got some of that.

00:26:27   I also, dermatologists and various dermatology friends

00:26:31   also frequently said, Casey was right,

00:26:34   that Cetaphil is a really good all-arounder.

00:26:38   for more kind of advanced needs,

00:26:41   like what I was having with my hands in the winter.

00:26:43   It wasn't quite strong enough,

00:26:45   and what they recommended was either the high-end creams

00:26:49   from CeraVe and from Usurin, and I got all three.

00:26:53   So I have O'Keeffe's Working Hands,

00:26:55   Usurin Advanced Repair Hand Cream,

00:26:58   and CeraVe Moisturizing Cream,

00:27:00   and I have to say, I have had significant progress

00:27:04   in one, not even one week, like in five days

00:27:07   since I've had these.

00:27:08   So, yeah, gotta say, y'all were right, thank you.

00:27:13   And so far, the O'Keeffe's working hands

00:27:18   seems to be the strongest in terms of forming a barrier.

00:27:21   'Cause it's basic, it's almost like wax.

00:27:23   Like it dries, it's very waxy, and it works pretty well.

00:27:28   I think the one that makes me the softest feeling afterwards,

00:27:32   which I think is better for daytime use maybe,

00:27:34   'cause then I get to enjoy my immense softness,

00:27:36   is the CeraVe moisturizing cream.

00:27:39   Somebody, a few people said there's a difference

00:27:40   between cream and lotion.

00:27:42   Nobody was clear on what that difference is

00:27:43   or how much it matters, so I got cream for the CeraVe.

00:27:46   It's great, I feel ridiculously soft and it's amazing.

00:27:49   And finally, many people wrote in to suggest the technique,

00:27:53   'cause I was complaining about getting it,

00:27:55   my hands all greasy and then being able to,

00:27:57   you know, being unable to operate doorknobs.

00:27:58   Many people wrote in to suggest

00:28:00   that what you're supposed to do really is apply the cream

00:28:03   to the back of one of your hands

00:28:06   and then rub the two backs of your hands together

00:28:09   to get it all over the back and like the top of your fingers

00:28:11   that way you never actually get it on your palms.

00:28:15   That's very good advice, I've been doing that as well.

00:28:17   I do wanna get it kind of around the fingernails

00:28:19   'cause like the little cuticle areas

00:28:20   always get all messed up too.

00:28:22   So I'm trying to get it all around there too.

00:28:24   So I get it a little bit on my hands

00:28:25   but certainly I'm not getting on the palms anymore.

00:28:26   So anyway, that's good.

00:28:28   Strongly recommend the O'Keeffe's Working Hands

00:28:31   and the CeraVe Moisturizing Cream

00:28:32   and I'll put links in the show notes

00:28:33   to those products on Amazon.

00:28:35   but it's incredible, I've never seen reviews this good

00:28:39   for the O'Keeves.

00:28:40   - Seeing all those people suggesting

00:28:42   with the rubbing the back of the hands together,

00:28:44   it makes me think about rubbing my two kneecaps together,

00:28:46   like the back of my hands are bony.

00:28:48   How do you, rubbing two bony parts of my body together,

00:28:51   I feel like I wouldn't get the moisturizer

00:28:52   in all the spots.

00:28:53   - Don't form a fist, just keep your hands flat.

00:28:55   - No, I understand, I guess I don't have enough

00:28:57   back of hand fat or something.

00:28:59   - Well you gotta work on that.

00:29:01   - It's just, I got bony hands, anyway.

00:29:04   - Keep your hands some pizza and beer,

00:29:05   I don't know what else.

00:29:06   - Yeah.

00:29:07   (laughing)

00:29:08   And the other suggestion I saw a lot of people had

00:29:10   was not just wearing it at nighttime, but wearing gloves.

00:29:12   In fact, when I did the search for the O'Keeffe's

00:29:14   working hands on Amazon, there were also glove results.

00:29:17   - Oh yes.

00:29:18   - So basically said, put the moisturizer on

00:29:19   and then put, lots of people had different suggestions

00:29:21   for gloves, put gloves on top of your moisturized hands

00:29:23   and then go to sleep to sort of lock in the moisture

00:29:27   at nighttime so it doesn't evaporate out of your hands

00:29:29   or whatever.

00:29:30   And it just, I don't know, I like it,

00:29:31   for someone who gets so bundled up when I go to sleep,

00:29:33   it still seems weird to me to wear gloves in Tibet.

00:29:36   - Yeah, I agree with that.

00:29:37   - If any of us would, it's you though.

00:29:38   - That's also true.

00:29:40   - You're basically wearing an entire snow suit

00:29:42   when you go to bed every night.

00:29:43   - I know, but not gloves, that's weird.

00:29:45   Anyway, and then the thing is, underneath the gloves

00:29:48   is your slimy lotioned hands, not, ugh,

00:29:50   give me a definite ick factor.

00:29:52   - Yeah, I might try that if I like,

00:29:54   'cause on Amazon there's packs of basic cotton gloves

00:29:57   that are made basically for this purpose

00:29:59   where they're kind of semi-disposable.

00:30:01   - Or other people suggested socks too,

00:30:04   which is even funnier.

00:30:04   You can do little sock puppets in bed.

00:30:06   Oh, goodnight, goodnight.

00:30:07   (laughing)

00:30:09   - Yeah, and people also suggested like, you know,

00:30:11   wearing rubber gloves if you have to wash dishes and stuff.

00:30:14   I haven't quite gone that far yet, but I have.

00:30:16   Like, it's been kind of crushing my soul.

00:30:19   I've been delaying doing dishes throughout the day.

00:30:23   So instead of washing dishes constantly

00:30:25   as they're being used to keep a clean kitchen,

00:30:27   now I like will let them accumulate for a couple hours

00:30:30   and then do them and it's killing me.

00:30:31   I am not good at leaving dishes behind in the kitchen.

00:30:35   Like I always have to have a clean kitchen.

00:30:37   And it's, that's very difficult for me.

00:30:39   That's the hardest part.

00:30:40   I'd rather have greasy hands all day

00:30:42   than have dirty dishes in the kitchen.

00:30:43   But it's been crushing me, but that also has been helping

00:30:47   to reduce the frequency of that.

00:30:49   - Did you ever see the Palmolive ads

00:30:50   that Merlin and I talk about?

00:30:51   - I don't think so.

00:30:52   - Softens your hands while you do dishes,

00:30:54   that was the pitch.

00:30:55   - Oh yes, that's right.

00:30:57   - It was an ad campaign where they show various people

00:30:58   dipping their hands into it and it's like, oh, you know,

00:31:01   like they're at a spa, remember we talked about

00:31:02   the directives, the commercials like they're,

00:31:04   they're getting their nails done or something

00:31:05   and their hands are soaking in a bowl of stuff,

00:31:07   which is apparently a thing that you do

00:31:09   when you get a manicure, I don't know,

00:31:10   'cause I've never gotten one.

00:31:11   And they're talking about dish soap

00:31:13   and the big surprise is they talk about pulling off

00:31:16   is you're soaking in it, you didn't even know it,

00:31:17   but your hands are in dish soap, like,

00:31:19   because the dish soap is so therapeutic

00:31:21   and softening to your hands that you could, anyway.

00:31:24   Their slogan was, "Softens your hands while you do dishes."

00:31:26   I bet it probably doesn't,

00:31:28   'cause it's probably soap, but yeah, wear gloves.

00:31:30   - Yeah, that doesn't seem like it could possibly work at all.

00:31:32   - I mean, it was the '70s.

00:31:34   - We are brought to you this week by Memberful.

00:31:39   Monetize your passion with membership.

00:31:42   Memberful lets you build sustainable recurring revenue

00:31:46   by selling memberships to your audience.

00:31:48   And Memberful is the easiest way to sell memberships

00:31:50   as you buy some of the biggest creators on the web,

00:31:52   because among many other reasons,

00:31:54   it integrates with the tools and platforms you already use.

00:31:57   you don't have to rebuild your whole tech stack.

00:31:59   It works with things like WordPress, MailChimp,

00:32:02   Discord, and so many more.

00:32:04   And they offer a great checkout experience,

00:32:06   great tools to sell your memberships,

00:32:09   everything you need to run a membership program,

00:32:10   including custom branding, all your stuff there,

00:32:13   so people don't feel like they're buying from Memberful,

00:32:15   they feel like they're buying from you.

00:32:17   And in fact, the payments even go directly

00:32:19   to your own Stripe account.

00:32:21   So they really, Memberful, they're a tool,

00:32:23   and they kind of step back from the limelight

00:32:25   to let you and your brand and your membership

00:32:27   really be what's facing customers.

00:32:29   And you retain full control over the whole thing.

00:32:32   And then they have a great checkout experience

00:32:34   for your customers with Apple Pay support,

00:32:36   free trial support, gift subscription support,

00:32:39   so many more features.

00:32:41   Memberful, they want to align their incentives with you.

00:32:44   They let you retain control and ownership

00:32:46   with the audience and your brand and your membership.

00:32:49   And they offer great support to help you

00:32:51   if you need to optimize stuff

00:32:52   or you just need help getting started,

00:32:54   Great support there at Memberful.

00:32:55   They have a world-class support team.

00:32:57   So see for yourself, you can get started for free

00:32:59   with no credit card required,

00:33:01   and see what Memberful can do for you.

00:33:03   It's a great way to launch a membership program.

00:33:06   There's almost no work to do,

00:33:07   and trust me, that has a lot of value.

00:33:10   And there's a reason why this is used

00:33:12   by some of the biggest creators on the web.

00:33:13   It's just a great platform.

00:33:14   See for yourself at Memberful.com/ATP.

00:33:18   Get started today for free with no credit card required.

00:33:22   Get that free trial going.

00:33:23   and once again, memberful.com/ATP.

00:33:27   Go there today.

00:33:28   Thank you so much to Memberful for sponsoring our show.

00:33:31   - All right, do you need me for this

00:33:36   or should I just bugger off?

00:33:38   - You need to know about the Mac Pro

00:33:40   and so does the rest of the world.

00:33:41   - Do I?

00:33:42   - You've had such a reprieve here

00:33:43   because it was like the two year transition

00:33:45   and we did that item a couple of shows ago.

00:33:48   Apple's entered year three of its two year transition to ARM

00:33:52   in a way that no one cares about it.

00:33:53   And the Mac mini, they still sell the Intel Mac mini

00:33:56   and the Mac Pro that no one cares about is still Intel.

00:33:59   As far as the world is concerned,

00:34:01   Apple transitions and it's fine.

00:34:02   That's why nobody really cares.

00:34:03   'Cause like, well, they transitioned the Macs

00:34:05   that actually matter, but there are these odd holes.

00:34:08   So this is the latest news story,

00:34:10   latest rumors from Bloomberg, I'm assuming Gherman

00:34:13   about the new Mac Pro.

00:34:16   And the story here is that,

00:34:19   he goes through a little bit of the history of like,

00:34:20   Oh, originally they planned to do basically

00:34:23   like two M1 Ultra stuck together.

00:34:25   We talked about this in the show many times in the past.

00:34:28   Can they get like four M1 Maxes stuck together

00:34:32   into this gigantic chip with like 40 cores?

00:34:34   - Yeah, that was a Jade 4C, right?

00:34:37   - Yeah. - Yeah.

00:34:38   - Yeah, and it seems like that just didn't happen

00:34:40   and isn't going to happen

00:34:41   'cause they're onto the M2 ones and you know.

00:34:43   I mean, maybe it could have come out like, you know,

00:34:46   could have come out towards the end of this year,

00:34:47   but it didn't.

00:34:48   So it seems like that's all off the table, right?

00:34:50   So now it's like Apple changed their mind and now the Mac Pro is going to be M2 based.

00:34:55   And so the obvious expectation was it would be like four M2 Macs stuck together in a big

00:35:03   square thing.

00:35:04   So it would be even better, right?

00:35:06   And the rumor is now that the big giant four M2 things stuck together Mac Pro was scrapped.

00:35:13   And the only thing you're going to get in the Mac Pro is going to be kind of like an

00:35:16   M2 Ultra, right?

00:35:18   So it's two M2 Mac-ish things stuck together.

00:35:21   They're not doing the big 401.

00:35:23   - Which to be clear is basically the same class of chip

00:35:26   that's available in the Mac Studio.

00:35:28   - Exactly, and presumably they would put this

00:35:29   in the Mac Studio, right?

00:35:30   If they upgrade the Mac Studio to M2,

00:35:32   it would get the M2 Ultra, and so would the Mac Pro.

00:35:36   And so this would be 24 CPU cores, 76 graphic cores,

00:35:40   and the rumor is up to 192 gigs of memory, right?

00:35:43   So, you know, a good chip, right?

00:35:45   But the M2 Extreme, like there would have been four of them,

00:35:48   just double all that it would have had 48 CPU cores 152 graphics cores and they didn't

00:35:53   say anything about the RAM and so the the story is that this is according directly to

00:35:58   the story not a surprise the company made the decision because of both the complexity

00:36:00   and cost of producing a processor that is essentially four M2 max chips fused together

00:36:05   it will also help Apple and partner Taiwan semiconductor manufacturing company save chip

00:36:09   production resources for higher volume machines basically the idea is if this thing costs

00:36:14   too much money, it takes up valuable fab space and wafers, and we're not going to sell a

00:36:20   lot of them anyway.

00:36:21   The suggested rumored price is $10,000 without any other upgrades for that big giant chip,

00:36:27   which kind of makes some sense, but not that much sense, because it seems weird that it

00:36:33   would be more expensive than the 2019 Mac Pro.

00:36:37   You could say, "Yeah, that's because this is a system on a chip and it's got everything

00:36:39   on one big chip, and that chip is expensive," but that hasn't been the case with the other

00:36:43   computers they've made. Combining the GPU and the CPU into a single SoC did not make the

00:36:49   MacStudio tremendously more expensive. Actually there are economies there of not having separate

00:36:54   components, not having to give a profit margin to AMD or wherever you're buying GPUs from.

00:36:58   But anyway, the idea that this is too expensive, too hard to make, and not a lot of people buy it,

00:37:06   I see where they're coming from with that, but I have some issues with it. The issues I've always

00:37:11   had with this discussion, the open question, which we won't know the answer to, and it

00:37:15   seems increasingly likely the answer is going to be disappointing.

00:37:18   If Apple does not allow third-party GPUs, how many times have we just talked about this?

00:37:23   Apple just freaking releases computers, so we know what you're going to do.

00:37:26   If they don't allow third-party GPUs, the discussion has always been, "Okay, but given

00:37:30   the rumors of this Jade 4C thing or whatever, they can put enough GPU on the SoC to be equivalent

00:37:37   to a current highest of high-end single GPU from a third party and that is that

00:37:44   was true back when the J4C rumors were out and if you look at the current stuff

00:37:47   it seems plausible right but if they that was assuming they would do the the

00:37:52   4x thing but if they'd only do the 2x thing it starts looking a little bit

00:37:56   more grim so I'd ran a bunch of numbers this is using Geekbench's metal

00:38:01   performance score which again like benchmarks are silly but and I didn't do

00:38:05   like a gaming benchmark, because I'm saying like,

00:38:08   let's treat this the way Apple treats it.

00:38:09   Like Apple cares about metal, they use it in their stuff.

00:38:12   And so the metal score is a reflection of how well

00:38:16   could Apple use these GPU resources using its own API?

00:38:20   Whatever, again, how much screens you give to Geekbench,

00:38:22   it gives a good back of the envelope estimate.

00:38:26   So let's take for starters, the AMD Radeon W6900X,

00:38:35   which is like the best single GPU you can get

00:38:38   in a 2019 Mac Pro from Apple.

00:38:41   That score is 166,000 on the metal Geekbench score.

00:38:46   If you had two of them,

00:38:48   so if you bought a fancy new Mac Pro in 2019 Mac Pro

00:38:52   and put two of those cards in it, easy W.

00:38:54   You have about 335,000 on your metal score, right?

00:38:57   AMD has just come out with their next line of cards,

00:39:01   which just increased the first number.

00:39:03   So instead of 6,900, it's 7,900.

00:39:05   So the AMD Radeon RX 7900 XTX,

00:39:10   gotta love that because it's different than the XT,

00:39:11   which is not the good one, XTX is the good one.

00:39:13   Anyway, the XTX is more than double the speed of the 6900.

00:39:18   So it gets around maybe 350,000.

00:39:20   I couldn't get an actual metal score for this,

00:39:22   but looking at game benchmarks and stuff,

00:39:24   it very easily doubles the frame rate

00:39:26   of the previous card, which is very impressive.

00:39:27   It's very close to matching Nvidia's best card.

00:39:30   So it's 350,000 for a single one of those.

00:39:32   And if you got two of those, that would be 700,000.

00:39:34   These are the numbers we're talking about

00:39:35   for the metal scores, right?

00:39:37   Let's look at what the M1 Ultra does with 64 GPU cores today.

00:39:42   So again, the best single card you could get

00:39:45   on the market today is about 350,000.

00:39:46   The best one you can get from Apple is 166.

00:39:49   The M1 Ultra is 94.

00:39:51   So not really in the ballpark of even

00:39:54   of the previous generation high-end card.

00:39:57   Certainly nowhere near the current generation

00:39:59   high-end cards.

00:40:00   Again, setting aside Nvidia entirely, right?

00:40:03   The rumored M2 Ultra, which is just a 2x M2 thing,

00:40:07   has 76 GPU cores, and assuming they scaled linearly,

00:40:10   which they probably want that,

00:40:10   probably be better than that,

00:40:11   but assuming they scaled linearly,

00:40:13   that's like 112,000.

00:40:15   Still not in the ballpark of a 6900,

00:40:18   let alone a 7900 from AMD.

00:40:21   Let's look at the M2 Extreme,

00:40:23   the one that would have been like four M2 Maxis

00:40:27   stuck together.

00:40:28   That is 225-ish thousand.

00:40:31   That's in shooting distance, spitting distance,

00:40:34   whatever they're saying is.

00:40:35   That's close to 350,000,

00:40:38   but still nothing to really write home about.

00:40:40   And I ran these numbers on my own Mac Pro

00:40:42   just for the heck of it.

00:40:43   I have this hodgepodge of old video cards in here.

00:40:48   I've got my W57X, and I've got a Pro Vega II,

00:40:53   and those two combined give me 170,000 on MetalScore.

00:40:59   So my two GPUs combined do about as much as one

00:41:03   of the best GPUs you can get from Apple.

00:41:06   And still considerably more than the rumored M2 Ultra.

00:41:10   The rumored M2 Ultra will get 112, my Mac Pro gets 170.

00:41:14   That leads to the question I always say,

00:41:16   how are you going to sell a Mac Pro

00:41:19   to people who bought your previous Mac Pro

00:41:22   when the previous one could just be filled

00:41:24   with so much more GPU power?

00:41:26   unless you also provide a way to add third-party GPUs.

00:41:30   Now, you don't have this problem.

00:41:31   If this is true, if these rumors are true

00:41:33   and they're not sending the Forex thing,

00:41:35   I think that's a reasonable thing to do

00:41:38   as long as you are not constrained on GPU

00:41:43   by what they decide to put in the SOCs.

00:41:46   That's always kind of been the worry about this.

00:41:47   It's like, all right, if I want to let a GPU,

00:41:50   I have to get all that CPU because it just comes with it

00:41:53   'cause they come in these little units.

00:41:54   You get the GPU cores, the CPU cores,

00:41:56   the neural engine, you know, it's a system on a chip,

00:41:59   you get all of it, and they've been doing it

00:42:00   with little building blocks.

00:42:01   If you just want a machine with a reasonable CPU

00:42:04   and then a huge amount of GPU, you can't get it.

00:42:07   Every time you want more GPU, it comes with a whole bunch

00:42:09   of other stuff that you may not be able to use.

00:42:12   - And vice versa, by the way.

00:42:14   I love having a whole bunch of CPU power,

00:42:16   I barely need GPU power, but those things

00:42:18   are tied together now.

00:42:19   - Yeah, and so for a modular Mac Pro system,

00:42:23   that poses some of a problem if you're also not going to offer slots.

00:42:28   And so that has been the question we've been working with.

00:42:30   How are you going to offer an expandable machine when you're putting so much more on the SoC?

00:42:34   And I really do think that in the, especially for the top end macro,

00:42:37   if they had gone with the quad arrangement, you could just say,

00:42:40   "Look, I know it's not as flexible as you want it to be, but this machine is so expensive anyway.

00:42:45   Tough luck. If you don't want all that GPU, you're getting it whether you want it or not.

00:42:49   If you don't want all the CPU, you're getting it whether you want it or not.

00:42:51   And by the way, you're paying for it in both cases.

00:42:52   And if there's any market that is able to eat this and be like, all right, we'll just deal with it

00:42:56   it's the Mac Pro because it costs so much money to begin with and

00:42:59   Money is less of an object to these people

00:43:01   It is in quote-unquote inefficient for them to be buying stuff

00:43:05   They don't need but oh, well it is I think it hurts more in your case Marco where you're like

00:43:09   Look, this is supposed to be your mainstream product. I don't want to have to buy a bunch of GPU cores

00:43:12   I don't need to make my compiles go faster. But you know, the prices are low enough, you know

00:43:16   You're not talking tens of thousands of dollars that it seems reasonable

00:43:19   but it will be disappointing if they ship a Mac Pro

00:43:23   and it's basically a Mac Studio that holds more RAM

00:43:26   and bigger SSDs that can't have as much GPU power

00:43:29   as my 2019 Mac Pro has.

00:43:32   - Yeah, this is why I still am really, really curious

00:43:37   to see what the heck they do to the Mac Pro

00:43:39   'cause it seems like, you know,

00:43:41   if you look at what I think is most likely to happen,

00:43:44   it's basically the Mac Studio.

00:43:46   Like that's like, you know,

00:43:47   what I have been guessing for a while,

00:43:49   like once we saw the direction of Apple Silicon,

00:43:52   I'd been guessing for a long time,

00:43:53   okay, well they're gonna be done with card slots,

00:43:54   no more card slots, definitely no more third party GPUs,

00:43:57   and they're just gonna multiply this chip

00:43:59   until they have enough GPU power to be competitive.

00:44:01   But that leaves so many open questions,

00:44:03   like as you're saying, they're just,

00:44:05   they're not in the ballpark of GPU performance

00:44:08   once you are competing with multiple slots,

00:44:11   each of which can hold the highest end NVIDIA or ATI,

00:44:13   or excuse me, AMD GPUs.

00:44:15   That isn't because Apple's GPUs are bad,

00:44:18   it's that you're dealing with massively different

00:44:21   power envelopes, hugely different.

00:44:24   The reason why Apple's GPUs are so good

00:44:29   for mobile devices and laptops and small desktops

00:44:32   is because they're so incredibly low power consumption

00:44:36   for the amount of computational power they offer.

00:44:39   They're very efficient.

00:44:40   But when you're looking at these big beefy cards

00:44:43   from Nvidia and AMD, they're like themselves

00:44:47   consuming hundreds of watts and have their own tremendous

00:44:50   heat sinks and fans and power supply,

00:44:53   just to power these giant things,

00:44:55   'cause they are, for many people,

00:44:56   that is the most important component on their computer,

00:44:59   so they're willing to spend that power

00:45:01   and heat and space and everything else.

00:45:03   But the way Apple designs their GPUs,

00:45:05   they just are not made for that kind of power consumption

00:45:09   and performance class at that high end.

00:45:12   And because they have this unified memory architecture

00:45:16   where everything is together as one big thing,

00:45:19   I don't see them ever making anything

00:45:21   that is going to be in that class,

00:45:24   unless they go for something radically different

00:45:27   than what they've made so far with the M line,

00:45:29   which would be, for instance, like a slotted GPU

00:45:33   that has its own memory and is not part

00:45:36   of the unified memory architecture

00:45:37   of the rest of the system and everything else.

00:45:39   And they can do that if they want to.

00:45:41   There's nothing saying they can't do that,

00:45:43   But I really decreasingly see the likelihood

00:45:47   of them doing that for one product, the Mac Pro.

00:45:51   And that's why my theory from the beginning here

00:45:54   has been there isn't going to be an M series Mac Pro

00:45:59   that uses third party GPUs or even has slots for other GPUs

00:46:04   and possibly has slots for anything.

00:46:06   We'll see about that.

00:46:07   But yeah, I just, I wonder too,

00:46:11   maybe what they're seeing in the market

00:46:14   is that maybe they actually finally don't have enough reason

00:46:18   to address this market.

00:46:20   Maybe they're okay saying, you know what,

00:46:23   the performance level that we offer in the Mac Studio

00:46:27   is enough for our highest end users and that's fine.

00:46:31   I don't necessarily agree with that,

00:46:33   but maybe that's what they're seeing, I don't know.

00:46:35   And certainly as Apple Silicon has gotten so, so good

00:46:39   and these new Macs that use it are so good

00:46:41   so fast and so powerful, many people, myself included,

00:46:45   and maybe this is just me seeing my own needs

00:46:47   and overestimating other people's resemblance to them,

00:46:50   but many people have been able to go down a level

00:46:54   in product line in terms of what kind of power

00:46:57   they quote need for their work.

00:46:59   I used to get the most powerful desktops,

00:47:02   then I was okay with iMacs, then I was okay,

00:47:04   now I'm okay with a laptop that happens to have

00:47:07   the M1 Macs and all this crazy RAM and ridiculous

00:47:11   SSD speeds and huge CPU power, you know,

00:47:15   it's all on a laptop and that's enough

00:47:16   for my needs right now.

00:47:18   Maybe enough people have had their needs shift

00:47:22   down the lineup in terms of like being able

00:47:25   to get away with or even be very happy with something like

00:47:29   a MacBook Pro or a Mac Studio that maybe this high end

00:47:33   market is not enough for them anymore to justify this.

00:47:36   And maybe, you know, at the same time,

00:47:39   I think a lot of high-end video users

00:47:41   have moved away from Apple in general,

00:47:43   like over the last decade or so.

00:47:45   So there's shifts in the marketplace there as well.

00:47:48   So I don't know, it seems like most of the demand these days

00:47:53   for all that massive GPU grunt

00:47:56   is in doing things like AI training,

00:48:00   where you actually are kinda not really,

00:48:02   you're not really doing that on a Mac full of,

00:48:06   full of Mac video cards, necessarily.

00:48:08   - They have dedicated hardware for that.

00:48:10   If you needed the big power on that,

00:48:12   you'd use one of those TPU things or whatever,

00:48:14   like actual silicon made for that.

00:48:16   - Right, but there's never been a CUDA story

00:48:18   on the Mac and stuff like that.

00:48:19   So there's all these areas of high end computing

00:48:21   that the Mac either has kind of lost over time

00:48:24   or was never really in to begin with.

00:48:27   And so I wonder, look, they've been getting away just fine

00:48:32   for the last couple of years not having

00:48:36   a new high-end Mac Pro, not having a Mac Pro

00:48:39   that used the modern processors,

00:48:42   and it's been pretty much fine.

00:48:45   I don't know a lot of people who are clamoring

00:48:49   for an Apple Silicon-based Mac Pro

00:48:52   that also has a bunch of GPU card slots.

00:48:56   I'm sure there are those people,

00:48:57   but are there enough to justify

00:49:00   this hugely different architecture

00:49:02   from all the rest of their Macs that use the same chip?

00:49:06   Probably not.

00:49:08   - I mean, this story is saying that they can't even

00:49:10   justify making a 4X cookie, you know,

00:49:13   not as easy as cookie cutter,

00:49:15   but like the whole idea of the double and quad

00:49:19   is let's reuse our investment in the computers

00:49:21   that actually sell, like the notebooks.

00:49:23   They were never making a custom custom thing.

00:49:25   I said, we can't just clean sheet,

00:49:28   make a Mac Pro processor, right?

00:49:29   We have to start with processors we use on our laptops.

00:49:32   Can we do a bunch of work to weave them together

00:49:35   to make a big one?

00:49:35   And they're saying, we can't even justify that

00:49:38   in the Forex case.

00:49:39   And that's not the custom one, right?

00:49:41   'Cause they're like, it's just too expensive,

00:49:43   too big or whatever.

00:49:43   And by the way, what you said before about like the GPUs,

00:49:46   high-end GPUs using hundreds of watts by themselves,

00:49:48   if you put four of these things together like this,

00:49:51   that whole arrangement will take hundreds of watts.

00:49:53   Like they can, they are within the same power envelope

00:49:57   as a single top-end GPU plus CPU combination, right?

00:50:01   So it's not like the architecture

00:50:03   can't scale up to that level.

00:50:05   It can, and it would.

00:50:06   If they put four of them together,

00:50:07   it would take hundreds of watts.

00:50:09   It would be, you'd crank it up,

00:50:10   and it would be a heck of a thing, right?

00:50:13   And it would require very fancy custom cooling,

00:50:15   and the interconnect would be very impressive or whatever.

00:50:18   But they're saying, "No, we can't even justify

00:50:20   "the investment in that,

00:50:22   "'cause we don't think it's worthwhile,

00:50:23   "and we don't have enough people who are gonna buy it,

00:50:25   "and it's gonna take up."

00:50:26   because you can imagine how big that would be.

00:50:28   The advantage of having even a Xeon CPU

00:50:31   plus a separate GPU is that you don't have to put it all

00:50:33   in one package, right?

00:50:36   It's not one die, but one package, right?

00:50:38   This would be, what, four dies, four M2 maxes

00:50:41   in one massive package arrangement?

00:50:43   That is a complicated beast to make, right?

00:50:46   Much more complicated than getting an Intel CPU

00:50:49   and then an Nvidia GPU and a PCI bus between them.

00:50:53   And so they can't justify the investment in that.

00:50:56   I think like an M2 Ultra,

00:51:00   put the same thing in the studio

00:51:01   and basically make the Mac Pro be a Mac studio

00:51:03   but with more expandability.

00:51:05   That will cover most of their needs.

00:51:08   But even for something as simple as the three games

00:51:13   that are optimized per year for Metal,

00:51:15   like whatever it was recently,

00:51:16   they did like a Resident Evil port,

00:51:18   the latest Resident Evil,

00:51:19   two years after it comes out on every other platform

00:51:21   they had a Mac port of it.

00:51:22   And by all reports, the Mac port is amazing.

00:51:24   It takes amazing advantage of Apple's metal architecture,

00:51:27   and it runs at really good frame rates

00:51:30   at very low temperatures

00:51:32   and without a huge amount of fan noise,

00:51:34   as it is an impressive game, right?

00:51:36   It's not like those use cases don't exist,

00:51:39   but it seems like the balance that Apple has struck,

00:51:42   like when they make these system-mounted chips,

00:51:44   they say, "How many CPU cores?

00:51:45   "How many neural engines?

00:51:46   "How many H.264 and H.265 decoder encoders?

00:51:50   "How many ProRes encoders?"

00:51:51   sort of the recipe for the SOC,

00:51:53   their recipes have been very sort of balanced.

00:51:56   Not too much CPU, not too much GPU,

00:51:59   just kind of a nice arrangement of stuff

00:52:01   that really hits the sweet spot for,

00:52:03   let's say, high-end notebook, right?

00:52:05   It has most, your use case aside, Mark,

00:52:08   where you're like, I don't need all the GPU cores,

00:52:10   it's a pretty good balance for most things

00:52:12   that most people want to do.

00:52:14   If they can never move out of that sweet spot,

00:52:18   they're never going to get like the low-end thing

00:52:21   that changes the balance heavily in favor of CPU

00:52:23   or runs headless or something, right?

00:52:26   With not a lot of GPU grant.

00:52:27   Or the high-end one, where you take a small number

00:52:29   of CPU cores and much, much more GPU, right?

00:52:33   They don't seem so far willing to go in those directions.

00:52:36   So if they don't go in those directions

00:52:38   and they're not willing to even take forward

00:52:39   their sort of well-balanced little things

00:52:41   and put them together, they're kind of boxing themselves

00:52:44   out of sort of top-end GPU performance.

00:52:48   Not even top-end, but sort of just high-end, right?

00:52:50   And it's not like, it's very often in their demonstrations,

00:52:54   they will show a piece of software

00:52:56   or even just simple 3D stuff where,

00:52:58   especially with like 6K screens and stuff,

00:53:00   you can get some fairly trivial,

00:53:02   like scene kit demo application that can pull down,

00:53:07   pull the frame rates down on a powerful Mac

00:53:10   because you don't have enough GPU grunt

00:53:12   to do all those pixel shaders on a 6K screen.

00:53:15   And that seems like, it kind of seems like,

00:53:18   I was gonna try and use a car analogy,

00:53:19   but I can't think of a more recent one,

00:53:20   but it's like, I'll use an even more obscure analogy.

00:53:23   When the Power Mac G4 had a front side bus

00:53:26   that was hilariously slower than the CPU,

00:53:29   I forget what it was, but the CPU was like

00:53:31   at four or 500 megahertz and the front side bus

00:53:33   is like at a hundred megahertz

00:53:34   and it was just so RAM starved and so unbalanced.

00:53:38   If you're shipping a screen with 6K pixels,

00:53:41   but you can't drive a fancy looking 3D scene

00:53:44   that uses all those pixels, even on your most expensive Mac

00:53:47   because you just don't have enough GPU cores

00:53:50   by the way you no longer have slots, that's not a good look.

00:53:53   So I feel like if you're going to extend the Mac line up to a high end, and you should,

00:53:59   you can have something.

00:54:01   Even if it's really expensive and nobody buys it, there has to be some kind of offering

00:54:04   there because right now there still are some customers and some applications that use it.

00:54:08   I forget what the app that was being demoed when the Mac Pro was introduced, but you remember

00:54:12   there was a Pixar thing showing a big 3D scene?

00:54:15   - Yeah, it showed, when we went, when we were at WWDC,

00:54:18   and we were able to go to that like,

00:54:20   not really hands-on, but kind of hands-on area,

00:54:22   there was somebody from Pixar showing the Mac Pro,

00:54:25   and they showed the entirety of the town

00:54:28   that most of Toy Story 4 took place in,

00:54:31   like the fair and everything.

00:54:32   And you could zoom in and out, pan around,

00:54:35   and it was the entire fair, like that entire fair

00:54:38   and the town and everything associated with it

00:54:39   was being rendered in real time.

00:54:42   And it wasn't, you know, perfect cinematic quality,

00:54:44   but it was stunningly good.

00:54:45   And you could zip around that thing like it was nothing.

00:54:48   It was amazing.

00:54:49   - All right, my recollection of that is

00:54:51   you could still make the thing chug

00:54:52   if you looked at the right thing.

00:54:54   Like the frame rates weren't, you know,

00:54:56   they weren't even up to 60 frames per second all the time,

00:54:58   let alone higher refresh rates

00:55:00   that presumably future Macs will have, right?

00:55:03   So I don't think it's the end of the world

00:55:05   if they just make an M2 Ultra.

00:55:07   And in fact, if they made that extreme quad one,

00:55:10   it's not like I would buy it

00:55:10   'cause I think it would just be too expensive,

00:55:13   but it's a shame that if they don't have any other GPU answer

00:55:18   that the 2019 Mac Pro will still be the GPU king

00:55:22   simply because you can shove more GPU in it, right?

00:55:25   And it'll be super sad if, like I said,

00:55:27   the M2 Extreme, they rumored,

00:55:31   has a similar amount of, or not the M2 Extreme,

00:55:34   the M2 Ultra, like the 2X one,

00:55:37   has a lower metal score than my current Mac Pro,

00:55:42   my current one.

00:55:43   And granted, MetalScore is like when you're combining

00:55:45   both the GPUs, so it's not like I'd be able to run a game

00:55:47   at higher frame rates, but I could buy a Radeon Pro 6900

00:55:52   from Apple for a hugely ridiculous price

00:55:56   that would also be faster than the rumored M2 Ultra, right?

00:56:00   That's not great, that's like last generation card

00:56:03   and I can put it in my Mac, just one of them,

00:56:05   and be faster than the M2 Ultra,

00:56:07   so I don't know what their pitch would be.

00:56:09   It would be like, oh, we've got the M2 Mac Studio,

00:56:11   which is great and it's small and blah, blah, blah.

00:56:13   And we've got the Mac Pro,

00:56:15   which is just like the Mac Studio, but it holds more RAM.

00:56:18   I don't know what the pitch is,

00:56:21   if it just holds more RAM and more CPU.

00:56:23   Like those are the two, and those are good,

00:56:25   and people might need those, but I don't know.

00:56:28   The mystery deepens here.

00:56:30   I really feel like Apple could solve their problems

00:56:33   by supporting third-party GPUs in card slots,

00:56:35   which is what everybody else does.

00:56:37   And yeah, you could still ship a really good one on the SoC,

00:56:40   But for the people who want the third-party ones,

00:56:42   it would be nice if they're there,

00:56:43   but we've talked to this to death in the past.

00:56:46   Shows it just doesn't make sense

00:56:47   from the unified memory architecture thing, right?

00:56:50   That's why we keep going around in this circle,

00:56:52   'cause it doesn't make any sense.

00:56:54   It doesn't make any sense to offer card slots,

00:56:57   and now with this rumor, it almost doesn't make sense

00:56:59   not to offer card slots.

00:57:00   - So I wanna pull on that a little bit.

00:57:03   I am talking well outside my depth,

00:57:06   and I don't particularly care about the Mac Pro

00:57:08   in the grand scheme of things,

00:57:09   But from my recollection, when we were initially pitched

00:57:12   on the M1, which was what, WWDC 2020, I think?

00:57:17   Something like that?

00:57:19   We were told that, hey, the thing that makes this amazing

00:57:22   is the unified memory architecture.

00:57:24   And the fact that the standard vanilla RAM

00:57:28   is shared with the video cards.

00:57:30   So a lot of the time that you would be waiting

00:57:32   for things to be copied about and moved around and whatnot,

00:57:35   there is no waiting because it's all the same batch.

00:57:38   It's all the same batch of RAM.

00:57:39   It's a unified memory architecture.

00:57:42   - And also that memory happens to be really fast

00:57:44   'cause it's also really, really close to the CPU.

00:57:46   - Right.

00:57:47   So if a lot, probably not all,

00:57:51   because I know the SSDs are also extremely fast

00:57:53   in these things, there's a lot of other stuff around this.

00:57:55   But I thought that one of the big things

00:57:57   about the M1 architecture and M2 now

00:57:58   and so on and so forth,

00:57:59   is that the unified memory architecture

00:58:02   is a very kind of wild way of making a CPU/GPU combination

00:58:07   that not a lot of people are doing,

00:58:09   and Apple did it so well, and it is so damn fast,

00:58:11   that that's where so much of the speed comes from,

00:58:14   and efficiency, but particularly speed.

00:58:16   So, and you've kind of alluded to this, Jon,

00:58:17   but I wanted to ask you directly,

00:58:20   doesn't having a card with a GPU,

00:58:24   doesn't that take away a lot of the benefit,

00:58:26   all of the benefit of a unified memory architecture?

00:58:28   'Cause it ain't unified anymore.

00:58:30   So like, would it even be worthwhile?

00:58:33   Is that what we really want?

00:58:35   I don't know.

00:58:36   - Well, I think it depends on the workload to some degree.

00:58:38   If you think about how this kind of system would be built

00:58:43   if there is a theoretical new Mac Pro

00:58:47   that has card slots that can support GPUs,

00:58:50   I think all three of those are questionable.

00:58:54   Like whether there even will be a Mac Pro

00:58:56   and whether it will have card slots

00:58:58   and whether those card slots will support GPUs.

00:59:00   But assume we have all three of those.

00:59:02   - Apple said we have Mac Pro.

00:59:04   They just, remember they said,

00:59:05   where we'll talk about that another day.

00:59:06   Like they publicly announced it,

00:59:08   so I really feel like it's still coming.

00:59:11   Something's coming.

00:59:12   - I have one, I have two, maybe one word for you.

00:59:15   - Air power, yeah I know.

00:59:16   - Air power.

00:59:17   - Yeah, I know, well.

00:59:17   - Yeah, we'll see.

00:59:18   But anyway, so if you look at like,

00:59:20   already we can make some assumptions.

00:59:22   Like there would be, presumably,

00:59:25   some kind of M2 or M3, Ultra Max Extreme,

00:59:28   whatever it would be, a whole bunch of M CPUs glued together

00:59:32   as the main processors of the whole thing.

00:59:35   If they're gonna have socketed memory,

00:59:38   then maybe the chip would have some kind of massive

00:59:42   like L2 or L3 cache, but for the most part,

00:59:46   I think main memory would just be its own bank of stuff.

00:59:50   Because otherwise, they're never gonna hit the 1.5 terabytes

00:59:54   that the current 2019 Mac Pro offers.

00:59:56   - They're just, yeah, they're just not gonna hit that

00:59:59   on the SoC, and I feel like they're,

01:00:02   it's easier for them to give up on the 1.5 terabytes of RAM,

01:00:05   which was mostly just like a silly bragging point,

01:00:07   than it is for them to give up on the GPUs,

01:00:10   because they're top-end GPU, you can just buy it,

01:00:14   and it gives you a metal score that's higher

01:00:16   than the rumored M2 Ultra thing.

01:00:19   But how many people ever buy with 1.5 terabytes of RAM?

01:00:22   So the tiered RAM, and there were rumors of that,

01:00:24   seems less likely to me than slotted GPUs.

01:00:28   - Yeah, and also, you look at the way

01:00:31   that PCs are structured now.

01:00:34   Like the architecture already supports GPUs

01:00:37   having not only access to system RAM

01:00:40   through various like fast methods

01:00:42   and you know dedicated buses and things like that,

01:00:44   but also they have access to their own giant pool of RAM

01:00:47   each and workloads and schedulers and everything

01:00:51   are already designed to distribute work to multiple GPUs,

01:00:56   each of which has its own pool of RAM.

01:00:58   And you know that's what John's Mac Pro is doing.

01:01:00   That's how it can achieve these high scores.

01:01:02   So it is possible to do all of that

01:01:06   with the Apple Silicon architecture, I'm sure.

01:01:10   It would take work and there would be different trade-offs.

01:01:14   Having it unified, you probably have lower latencies

01:01:17   in certain ways, you might have higher bandwidths

01:01:18   in certain areas, almost certainly.

01:01:21   So chances are if there was such a Mac Pro

01:01:27   that had only slotted GPUs and only socketed RAM,

01:01:32   maybe they would actually be slower

01:01:34   in certain microbenchmarks because they wouldn't,

01:01:37   the RAM would be a little bit higher latency

01:01:39   or a little bit lower bandwidth or whatever it would be,

01:01:41   but that when doing these massive jobs

01:01:44   that people would buy these machines for,

01:01:46   overall they would be able to outperform

01:01:49   the MacStudio and the MacBook Pro and things like that

01:01:51   because they would be able to dispatch all this work around

01:01:55   and the increased latency would be made up for

01:01:59   by just the massive amount of parallel power

01:02:02   and total resources you'd have at your disposal.

01:02:05   So it doesn't necessarily say that it would be guaranteed

01:02:09   slower or worse, it would just be a different trade off.

01:02:12   - Yeah, mostly what you're getting

01:02:13   with the unified memory architecture,

01:02:14   like it's an efficiency choice.

01:02:15   And I don't mean efficiency in terms of power usage,

01:02:17   oh, that is definitely there as well.

01:02:19   I mean in terms of how much stuff do you have to spend money

01:02:23   on to put into your computer.

01:02:24   So with separate GPU slots,

01:02:26   like these GPU come with like 32 gigs of RAM.

01:02:28   That used to be the maximum you could even put

01:02:30   in an entire machine with an M1 SOC before the Mac Studio

01:02:35   or whatever broke through that barrier was.

01:02:37   So you have to get so much more stuff.

01:02:39   So you get this video, top-end video card,

01:02:42   which has huge numbers of GPU cores on it.

01:02:45   Just a massive chip just filled with GPU.

01:02:48   And then it has its own 32 gigs of RAM

01:02:51   that is special, extremely high bandwidth RAM,

01:02:55   really, really close to that GPU.

01:02:57   That's just like a complete duplication of effort

01:02:59   of the stuff that's in the main SoC,

01:03:01   'cause the SoC's already got the GPU cores

01:03:03   and RAM real close to it

01:03:04   that's really high speed and special.

01:03:06   Now you've got a second copy of that

01:03:07   only without as many CPU cores,

01:03:09   although they do have CPU cores there, of course,

01:03:12   on an entire card, and they communicate over the PCI bus.

01:03:15   Right, and then you could have multiples of those cards.

01:03:16   You could have multiple GPUs on a single card.

01:03:18   It is inefficient in terms of money.

01:03:21   You have to pay for all that stuff,

01:03:23   and then power, just to power the card and everything.

01:03:26   It's a duplication of effort.

01:03:28   And yes, there is a runtime cost to that sometimes

01:03:30   in terms of transferring data back and forth,

01:03:32   but you can solve that problem

01:03:34   by just adding more resources most of the time.

01:03:36   If you have enough VRAM on your dedicated GPU,

01:03:41   most things that you do with GPUs

01:03:44   don't require constant shuttling of data back and forth

01:03:47   to the point where you're saturating the bandwidth,

01:03:49   and that is your limiting factor.

01:03:51   That's why they have so much RAM in both of these places.

01:03:54   There are some situations that are like that

01:03:55   where Apple's SOCs excel

01:03:57   because they don't have to copy the data.

01:03:58   That's what of course they're going to brag about.

01:04:00   And definitely on situations

01:04:02   where you actually have a power budget or a battery,

01:04:04   you do not want to have a complete duplication

01:04:06   of resources elsewhere.

01:04:07   But again, in looking at the line,

01:04:10   I think I say what machine should spend money, resources,

01:04:13   space, heat, power, dollars, you know, everything

01:04:17   on a duplication of resources

01:04:19   just to get that top end maximum power?

01:04:20   it would be the Mac Pro, right?

01:04:22   It's the reason they sold it with that dual,

01:04:24   whatever, Apple itself sold dual GPU cards,

01:04:28   and you could get two dual GPU cards,

01:04:30   and so you would have four high-end GPUs

01:04:32   shoved into your Mac Pro.

01:04:34   And it only seems less ridiculous

01:04:36   because we don't think of the Xeon CPU

01:04:37   as having any kind of integrated graphics.

01:04:40   I think it has none.

01:04:41   I asked that question last time.

01:04:42   I don't remember what the answer was, right?

01:04:44   But having something with a huge amount of GPU,

01:04:48   a huge amount of really fast GPU on the SoC

01:04:50   And then also third-party GPUs, you'd only do that

01:04:54   if you feel like, well, I gotta pay for this GPU on the SOC,

01:04:58   but I actually need more than that.

01:04:59   I need more GPU cores.

01:05:01   I can do my work twice as fast

01:05:03   if I have twice as many GPU cores,

01:05:06   and I'm willing to pay for them.

01:05:08   Apple should, I think, be willing to sell you that

01:05:10   and say, well, you're kinda wasting the GPU cores on here

01:05:13   unless the software is written well enough

01:05:14   to be able to use them all at once.

01:05:16   And it could be, 'cause Apple's frameworks

01:05:17   lets you sort of use the computing resources

01:05:20   wherever they are.

01:05:22   But lots of workloads are written with the expectation

01:05:25   that you're not going to have a unified memory architecture.

01:05:28   You will have to transfer stuff from regular RAM to VRAM,

01:05:31   and that's OK.

01:05:31   And the programs account for it.

01:05:34   And what some customers really need

01:05:36   is I just need as many of those GPU cores as possible.

01:05:39   I'm willing to spend 800 watts on it.

01:05:40   Just give them to me, right?

01:05:42   Apple should, I think, offer that.

01:05:45   And that's why I'm still holding out hope for a GPU card slot.

01:05:47   But I think I would actually be happier if they found some way to put enough GPUs to

01:05:57   be competitive, not the best, but competitive.

01:05:59   Because if they're competitive in their overall middle score on an SoC, they're going to crush

01:06:04   it in anything that is bottlenecked on bandwidth back and forth to the VRAM, because they don't

01:06:10   have to do that, right?

01:06:11   And so I think that solution, Apple can brag about it and say, "Oh, they'll put up one

01:06:15   on those graphs say, look how much faster we are

01:06:17   than even the top-end Nvidia card.

01:06:19   And they'd pick some particular benchmark

01:06:21   that really benefits from the SOC

01:06:24   and unified memory architecture.

01:06:25   And they would show them crushing Nvidia

01:06:26   and just ignore all the ones where Nvidia crushes them

01:06:28   or whatever, right?

01:06:29   That's fine.

01:06:30   That's the Apple way to do it.

01:06:31   Like have something, that's a unique selling proposition.

01:06:34   We make different performance trade-offs than Nvidia.

01:06:37   That's why you might consider us instead of Nvidia.

01:06:39   We're not just our own CPU plus an Nvidia card

01:06:43   'cause we don't like them or whatever.

01:06:44   But this situation where you're like,

01:06:47   we're not competitive at all,

01:06:49   and we don't offer card slots,

01:06:50   but if you need a little bit more than a MacStudio,

01:06:53   the Mac Pro is for you.

01:06:54   That's not as exciting.

01:06:57   And so I really hope, I mean, you know,

01:06:58   I hope they keep making Mac Pros forever.

01:07:01   And so they get another shot at this with the M3,

01:07:04   they can make different choices.

01:07:05   They have lots of great building blocks.

01:07:07   Their GPU cores look like they're really good,

01:07:09   like in terms of what they do for the power they're given

01:07:11   and the space and everything.

01:07:12   Their CPU cores are phenomenal, right?

01:07:15   It's just a question of how do we divide up

01:07:17   these components, what kind of ratios do we choose?

01:07:20   Maybe they'll make some interesting choices

01:07:22   in terms of arrangement of stuff for the headset.

01:07:25   Or who knows, maybe they'll just use watch CPUs,

01:07:27   I don't know, like.

01:07:28   But I think the headset might end up needing

01:07:31   a little more GPU grunt and a little bit less CPU grunt,

01:07:34   so maybe they'll make a different arrangement in that.

01:07:36   I continue to have confidence in their overall architecture

01:07:40   because I think they have all the building blocks,

01:07:42   It's just a question of how do they dull them out

01:07:44   to their individual machines.

01:07:46   And this rumor is essentially a story of them

01:07:49   changing their mind and making different trade-offs.

01:07:51   And when they ship a product, we'll see how it goes.

01:07:54   - Were you sad?

01:07:55   I mean this genuinely,

01:07:56   like I'm not trying to poke the bear here,

01:07:58   but are you sad reading this?

01:08:00   - Not really, 'cause like I said,

01:08:01   I don't think I would ever buy the Forex one

01:08:02   'cause it would just be so expensive.

01:08:05   I mean, I selfishly hope they do have GPU card slots

01:08:08   'cause that's my most efficient, money-wise,

01:08:12   path to get good GPU power.

01:08:16   Mostly at this point, what I care about is a case design

01:08:19   with a better cooling solution than the MacStudio,

01:08:22   you know what I mean?

01:08:23   'Cause I like the cooling solution on my Mac Pro.

01:08:25   It's a really big case and with really big,

01:08:27   slow-moving fans.

01:08:28   The MacStudio is not that.

01:08:30   It's not as good.

01:08:31   So if they introduce a new Mac Pro,

01:08:35   obviously whatever Mac Pro they introduce

01:08:36   is gonna be so much faster than this,

01:08:37   except maybe in GPU.

01:08:40   But assuming it is within the ballpark

01:08:41   and I eventually replace this machine,

01:08:43   I would like it to be with something

01:08:44   that I don't have to bolt to the bottom of my desk

01:08:46   because it's too noisy.

01:08:47   - We are sponsored this week by MasterClass.

01:08:51   MasterClass offers over 2,500 video lessons

01:08:55   on a wide variety of topics

01:08:57   from over 180 of today's most brilliant minds.

01:09:01   These are world-class instructors

01:09:03   at the top of their field,

01:09:04   and each class is broken into individual,

01:09:06   roughly 10 minute each video lessons.

01:09:09   Members can explore at their own pace

01:09:11   and each class has downloadable materials,

01:09:13   class guides, recipes and more.

01:09:15   Master Class is accessible anytime, anywhere,

01:09:17   on all your devices and is available also

01:09:20   via audio only modes if you wanna just listen

01:09:22   to classes on the go.

01:09:24   It is a great service, there's so much on there.

01:09:27   You know, I was looking through the night

01:09:28   and there are just so many topics

01:09:30   that I am super interested in.

01:09:31   You know, they have The Art of Negotiation by Chris Voss.

01:09:34   You have John Legend teaching songwriting.

01:09:37   Kris Jenner teaches the power of personal branding.

01:09:40   Mariah Carey teaches using the voice as an instrument.

01:09:43   Bob Iger with business strategy.

01:09:45   Gordon Ramsay teaching cooking.

01:09:46   I mean, this is just incredible talent

01:09:49   producing these classes and really great topics

01:09:52   and super interesting stuff.

01:09:54   With MasterClass, you learn from the world's best artists,

01:09:58   icons, and leaders anytime, anywhere at your own pace.

01:10:02   So I personally am really into a lot of this stuff.

01:10:05   I'm trying a lot of it out.

01:10:06   First one I signed up for were the negotiation ones

01:10:08   and some of the science ones and everything.

01:10:10   It's really cool.

01:10:11   So I highly recommend you check it out.

01:10:14   This holiday, give the perfect gift

01:10:16   of an annual MasterClass membership and get one free.

01:10:19   Go to masterclass.com/atp today.

01:10:23   That's masterclass.com/atp.

01:10:26   Terms apply.

01:10:27   Thank you so much to MasterClass for sponsoring our show.

01:10:32   Let's do some Ask ATP.

01:10:36   Arthur Dickerson writes, "Do any of you journal, handwritten or in an app?"

01:10:40   I will start and I will tell you that I am a devout Day One user.

01:10:44   I don't think they've sponsored the show.

01:10:47   They may have a long time ago, but I don't think they have.

01:10:51   Day One is a journaling app that actually, come to think of it, I think I stumbled upon

01:10:55   Day One when Marco and Tiff and Aaron and I went to Germany because I wanted to have

01:11:00   a way to take note of the things that the four of us did, but do that not on Foursquare

01:11:06   and not on Twitter, not on Instagram, but somewhere that was private for just me. And

01:11:11   I'm pretty sure it was the Germany trip that started me down this path. And day one does

01:11:15   let you say, you know, I've been, I'm here and it lets you geotag things and so on and

01:11:18   so forth. But I had been using it on and off ever since then in 2013, but I got really,

01:11:24   really serious about it around the time of the pandemic because I thought, oh, what a

01:11:29   fun chronicle this will be of the two weeks of the pandemic, right? Right? Right? Right?

01:11:36   So I was, I began the pandemic on March 13th of 2020, not that I remember, trying to make

01:11:42   sure that at least once a day I take note of something that happened that day, typically

01:11:46   with a picture if I can, but if I don't, you know, I'll write something down. And I've

01:11:50   been going, I think every single day since the 13th of March, 2020. And I really like

01:11:57   day one. Again, you know, I don't think they've sponsored in the past. They're certainly not

01:12:00   sponsoring now, but I really like day one because it's private. It's very reliable in

01:12:04   my use and it does basically everything I want it to do and not too much more. And I

01:12:12   mean that in a complimentary way. It lets you make audio recordings if that's your thing.

01:12:18   You can put videos on there. It all syncs with their own proprietary system. And it's

01:12:22   It's pretty darn reliable and it's very quick.

01:12:26   It's easy to get in and out.

01:12:27   It has a really great feature that shows you what happened on this day in years past.

01:12:31   I really, really, really like it a lot.

01:12:33   And I use it mostly for just taking note of things that we did during the day that are

01:12:39   even mildly noteworthy.

01:12:41   I try to be very, very religious with it when I'm traveling because it's so easy to forget

01:12:47   the ins and outs of a WWDC trip, as an example.

01:12:51   And my family's actually going on our first plane vacation since 2019 in just a couple

01:12:55   of weeks.

01:12:56   And I'll be using it a lot then as well.

01:12:58   And it's such a great way to say, "Oh, we ate at such and such a restaurant."

01:13:01   And while you're there, you can just create an entry and it'll geotag where you are.

01:13:05   And you could go back at the end of the day and add in notes about what you did or thoughts

01:13:10   or what have you.

01:13:12   I really, really like day one.

01:13:13   If you have any interest in this whatsoever, I can't suggest it enough.

01:13:17   You should definitely try it out.

01:13:19   Marco, do you do any sort of journaling

01:13:21   in any way, shape, or form?

01:13:23   - No.

01:13:23   - Good talk.

01:13:25   - It's one of those things, it's like playing guitar.

01:13:29   I would love to be an expert in playing guitar,

01:13:32   but I am never willing to put in the effort

01:13:35   to go from nothing to expert.

01:13:37   - You and me both, brother, seriously.

01:13:39   I really agree with you, I really do.

01:13:41   - In this case, there are some times in which

01:13:44   I look back and say, boy, I would love to have journaled

01:13:47   for that time, but I never actually do it.

01:13:51   And I love the idea of it,

01:13:55   but I never want to do it so badly that I actually do it.

01:13:59   - Jon, I have a question about your journaling first.

01:14:01   Like, the whole idea of this question

01:14:04   of calling it journaling makes me think that

01:14:07   it's like a different framing of writing,

01:14:12   'cause writing something every day,

01:14:13   or writing about your day or what happened in your life

01:14:15   can take lots of different forms.

01:14:17   And I feel like journaling comes with an expectation

01:14:20   of a sort of context, a frame of reference.

01:14:24   What I'm occurring to do is having a diary, right?

01:14:26   Which is a different way of saying the same thing.

01:14:28   I write something down every day.

01:14:29   All right, but if it's a diary, it's like,

01:14:31   dear diary, here are my deepest thoughts, right?

01:14:34   You know, my deepest feelings

01:14:35   that I would never want anyone to know,

01:14:37   and I'm writing them down, right?

01:14:38   Whereas journaling is more like,

01:14:41   I mean, you mentioned you're doing it someplace private,

01:14:42   but when I see journaling, I feel like it's,

01:14:45   I always think of it as like, how would you feel if this got into the wrong hands and

01:14:50   was published, right?

01:14:52   And I feel like a journal is written with the notion that it is slightly less bad than

01:14:57   having your diary go public.

01:14:59   Because your diary, it's like, this is it.

01:15:01   This is the base level.

01:15:02   There is nothing that I won't write here.

01:15:04   No matter how embarrassing, no matter like, I'm just going to write all my thoughts and

01:15:07   feelings and it's just for me.

01:15:09   And it's my deepest inner thoughts with no filter.

01:15:12   And journal is not that.

01:15:13   is I'm going to put my feelings and thoughts about the day here, but it's kind of like

01:15:18   something I think we're all familiar with. It's like writing not for an external audience,

01:15:23   but also thinking about if someone did see this, how dire would it be for me? And that may seem

01:15:29   like a weird thing to think about, but I have to tell you that a lot of the things I do in my life,

01:15:32   like in the internet age, I have that mindset of like, I never intend this to go elsewhere or be

01:15:40   seen outside of the context where you know where it's supposed to be but it could so let me make

01:15:46   sure if that happens it's not entirely awful like always sort of holding something back so my

01:15:52   question for casey is are your journal entries like diary entries or like a historian cataloging what

01:15:59   your family did in a day or somewhere in between like how how do you move along that spectrum

01:16:04   Um, it's a good question. I would say mine is more of a captain's log than it is a diary.

01:16:10   Like there are occasions that I will put, I mean, I can't even think of an example off the top of my

01:16:16   head. Like it's not that I'm trying to filter myself. I just, I can't even think of an example

01:16:19   of something that is really and truly private. Like I'll talk about, um, you know, oh, we had such

01:16:25   and such a medical issue that, that we needed to take care of. Like that'll get noted, but it,

01:16:30   that's not the end of the earth if that sort of thing escaped. You know, it's one of those things

01:16:34   where I wouldn't want my day one journal to be public,

01:16:36   but I don't think it would be ruinous if it was.

01:16:40   You know what I mean?

01:16:40   Like, I don't know that I would be canceled

01:16:43   on account of my day one journal.

01:16:44   I'm much more likely to get canceled

01:16:45   'cause my dumb hot takes on this show

01:16:48   or Maston or Twitter or something like that.

01:16:49   - It's not about canceling.

01:16:50   It's not like we're saying you're planning a murder

01:16:52   and you're writing a journal.

01:16:53   - For sure, for sure.

01:16:53   No, no, I'm with you.

01:16:54   - But just like writing about your feelings.

01:16:56   Like, you know what I mean?

01:16:57   The journal, as you describe it,

01:16:59   could be useful in this context,

01:17:00   but I'm saying like something you would bring

01:17:01   to your therapist.

01:17:02   Like, you could bring to your therapist and say,

01:17:03   and say, "Oh, on this day, I did this,

01:17:04   and so it helps me understand, you know,

01:17:07   what my feelings might have been,"

01:17:08   versus literally writing down your feelings,

01:17:10   bringing that to your therapist, and said,

01:17:11   "Here's how I felt on this day,

01:17:13   and here's what I wrote about it," right?

01:17:14   So it's more, that's what I feel like is,

01:17:16   the word journaling implies like a one-step remove,

01:17:19   where you're sort of like, like you said, a captain's log,

01:17:22   which has a particular context, and, you know,

01:17:25   Picard writes stuff in his captain's log about his feelings,

01:17:28   but there's still, it's written,

01:17:30   a captain's log in particular is written

01:17:31   with the expectation that it will have an audience

01:17:33   people who want to look at the captain's log to see what happened on the events of the

01:17:36   journey. And it's not like the captain doesn't write about their feelings like I fear we'll

01:17:40   never find land or whatever, but he's not writing about his, you know, secret love of

01:17:44   the first mate or whatever.

01:17:46   Yeah, no, my, my thoughts about how much I love the two of you, I keep just for me. No,

01:17:53   I mean, all kidding aside, it really is more along the lines of a captain's log. And I

01:17:58   think to back up a step, what day one does for me is fix a problem that I have. And I don't mean that

01:18:06   in a sarcastic way or to be flippant. I genuinely have a terrible memory. And I'm not proud of that.

01:18:13   I wish I had a much better memory, but my memory is garbage. And I can remember certain things.

01:18:19   And I can remember broad strokes of Declan's fifth birthday trip to Disney World in late 2019.

01:18:24   But I really enjoy, if I have the time and the presence of mind to do it, I really enjoy

01:18:31   taking note of this is where we had lunch on the first day that we were at Disney.

01:18:35   This is where we eat dinner.

01:18:37   And even to the degree of this is what I ate, maybe I liked it, maybe I didn't.

01:18:40   But I'm not going to go into how it made me emotionally feel that dinner or anything like

01:18:45   that.

01:18:46   It's really solving the problem of I have no memory.

01:18:49   This is my external outboard memory that helps me to relive events in my life.

01:18:56   Usually good ones, sometimes mundane, but just events in my life.

01:18:59   And day one is really good for me for capturing that.

01:19:02   Now you can write long form stuff in it for sure.

01:19:05   That's not a need that I typically have, and it's not something that I do particularly

01:19:09   often at all, but you could do that there if you so desire.

01:19:14   But yeah, if I were to summarize, I would say more captain's log than anything else.

01:19:18   My typical day one entry is a title,

01:19:21   often but not always a picture,

01:19:22   and typically one paragraph or less, if any.

01:19:25   Sometimes it's just the title and the picture,

01:19:27   nothing else.

01:19:28   - Yeah, and to answer my question,

01:19:29   I don't do any journaling,

01:19:31   although my wife has day one installed

01:19:33   and I hate it because it takes Command + Shift + D.

01:19:35   (laughing)

01:19:36   When I'm in the Finder, I hit Command + Shift + D

01:19:38   to open the desktop folder,

01:19:39   so I can sort by stuff or whatever,

01:19:41   so I don't have to look through her incredibly messy

01:19:42   desktop full of icons everywhere,

01:19:44   and I can just get a list view and find things

01:19:46   sorted by date or whatever.

01:19:47   - And Command + Shift + D does not open a stupid finder window

01:19:51   of the desktop, instead it opens day one.

01:19:53   - I have day one installed, and Command + Shift + D

01:19:55   definitely does the desktop.

01:19:56   - I think I'm thinking of day one.

01:19:57   It's like a teal icon, teal thing.

01:20:00   - It's like a light blue, yeah.

01:20:02   - Maybe they changed that.

01:20:03   - Looks like a book.

01:20:04   But also today I learned that Command + Shift + D

01:20:06   brings you to the desktop, so that's extremely useful.

01:20:08   Thanks, John.

01:20:09   - It's in the Go menu, it's not a secret command.

01:20:10   - I just never paid attention.

01:20:11   - You know, Command + D in a save dialog also does that.

01:20:14   (laughing)

01:20:16   - Fine, fine, fine.

01:20:17   Matt Tuohy writes, "What folders do you add

01:20:18   "to the top level of your home directory?

01:20:20   "For example, I have one for my code,

01:20:22   "another for things that might be falling

01:20:23   "off the back of a truck.

01:20:24   "Does this feel like a sacred place in your computer,

01:20:26   "saved only for special cases?

01:20:28   "P.S., I just installed Postman,

01:20:30   "and it created, you know, tilde slash Postman, yuck.

01:20:33   "I went first last time, so I'll save myself for last."

01:20:37   I think Marco, let's start with you, please.

01:20:39   - I generally don't add stuff to my home folder directly.

01:20:43   And it has a similar problem as the Dropbox folder,

01:20:48   in the sense that many apps will add their own top-level

01:20:54   folders there that I don't necessarily want there

01:20:57   and don't need to deal with myself

01:20:58   and don't want to deal with myself,

01:21:00   but they need to put their data there

01:21:01   because they're hard-coded to do it or whatever.

01:21:03   And so it ends up making the home folder not particularly

01:21:07   useful to me as a place that I want to look very often,

01:21:09   because it's cluttered up with items

01:21:12   that I don't need to be there,

01:21:13   that some app or the system decide need to be there.

01:21:17   So I don't look that often.

01:21:19   What I do have there, you know, of course,

01:21:21   you know, I'll use the subfolders,

01:21:22   like documents and photos and things like that.

01:21:25   But I do have in that top level folder

01:21:28   all of my Git checkouts.

01:21:31   So it's basically a list of projects.

01:21:33   You know, it's like overcast.

01:21:35   - In your home folder?

01:21:36   - Yeah, yeah, yeah.

01:21:37   - Wow.

01:21:38   - Home/overcast, you know, home/neutral.

01:21:42   - I find that to be madness as well, but you do you.

01:21:44   - Yeah, well, hey, that's, (laughs)

01:21:46   it makes all my paths really easy to type,

01:21:48   you know, tilde slash whatever.

01:21:50   So that's what I do.

01:21:52   So it's full of checkouts of all my projects

01:21:56   and, you know, a handful of folders

01:21:58   that I never need to look at.

01:21:59   And then of course my fake Dropbox folder,

01:22:02   which is actually the Maestro Dropbox folder,

01:22:06   but, you know, for path compatibility,

01:22:07   it's still called Dropbox.

01:22:09   - Jon?

01:22:11   I don't put things in the top level of my home folder.

01:22:14   Other apps do, and I find it incredibly rude.

01:22:17   - Yes.

01:22:18   (laughing)

01:22:19   - So Creative Cloud Files.

01:22:20   I don't like you.

01:22:21   I don't want you to be there.

01:22:23   Why are you in the top level of my home directory?

01:22:25   Dropbox does, and I have to say,

01:22:27   Dropbox is one of the only ones that if I had a choice,

01:22:29   and you do have a choice, I think,

01:22:31   I'm okay with that being top level because of what it is,

01:22:35   but I always want to have a choice.

01:22:37   There are other apps that try to put things in there.

01:22:39   Sometimes if they insist on being there,

01:22:41   I make them invisible just so I don't have to see them

01:22:43   in the finder.

01:22:44   I tend not to go to that folder,

01:22:45   but if you think about that folder,

01:22:46   there's a good history to be had here

01:22:48   going all the way back to the next days

01:22:49   of like someone somewhere back in the day

01:22:53   decided they were going to pick a reasonable top-level

01:22:57   arrangement for people's home directories.

01:22:59   And so they decided you're gonna,

01:23:01   you know, the folders that you see there

01:23:03   when you get a Mac, documents, movies, pictures,

01:23:06   not photos, pictures,

01:23:07   because some of them might not be photos, right?

01:23:09   Then there's the dreaded library folder,

01:23:11   which people don't understand,

01:23:12   and it's got the little pillars on the icon or whatever.

01:23:14   We as developers know what's in there,

01:23:16   but it's computer stuff, right?

01:23:20   Downloads, kind of makes sense as a top-level thing.

01:23:23   Public, nobody freaking knows what that is

01:23:25   because the web browser that used to ship with Macs,

01:23:27   like it used to have a GUI and then it didn't have a GUI.

01:23:29   It's still, you know, Apache I think still comes

01:23:31   with Mac OS, but anyway.

01:23:32   If you run a web server on your Mac

01:23:34   out of people's home directory stuff, it comes out of public,

01:23:36   but it's also I think the file sharing

01:23:37   public directory or whatever.

01:23:39   And don't forget, inside of public is a folder called

01:23:41   Drop Space Box, which is different from Dropbox.

01:23:44   - Yeah, that's right.

01:23:45   Now, sites is the web browser one,

01:23:46   public is like the public file sharing one,

01:23:48   and then a top level folder for music,

01:23:50   not audio, but music.

01:23:52   So you've got movies, not video, pictures, not photos,

01:23:56   music, not audio.

01:23:57   It's such a model, and of course applications.

01:24:01   I do have a applications folder in my home directory,

01:24:03   and that is a supported thing,

01:24:04   and the OS will give it an icon,

01:24:05   and I do have applications in there.

01:24:07   There are applications that only I want to use,

01:24:09   that I know no one else in the system wants to use.

01:24:11   It's sort of like a, like you don't graduate

01:24:13   to slash applications until you reach

01:24:15   a certain level of usefulness.

01:24:16   So I've got a lot of crap until the slash applications.

01:24:20   There are other folders that you can put in there

01:24:22   that the OS understands,

01:24:23   but this sort of arrangement of stuff,

01:24:26   chosen a really long time ago, it makes some sense.

01:24:28   It's not a ridiculous arrangement of things,

01:24:31   but Apple in typical fashion on Mac OS

01:24:34   has not really revisited that structure to revise it much.

01:24:39   And it's kind of strange,

01:24:40   like when they make features in the future,

01:24:42   for example, iCloud Drive, where they say,

01:24:45   what is it, desktops and documents, right?

01:24:49   The desktop folder, which is invisible,

01:24:50   but is in the top level of your home directory,

01:24:53   and documents, those are the ones that sync,

01:24:56   is desktop documents, is that it?

01:24:58   Is it downloads as well?

01:24:58   I forget.

01:24:59   - No, it's just, it's called documents and desktop,

01:25:01   and it's literally those two folders.

01:25:03   - Yeah, so those two folders sync in iCloud

01:25:07   and it's like, hmm, all right, so I kind of understand

01:25:11   why library doesn't sync, you know,

01:25:13   'cause we understand like you might have one

01:25:14   of our machine preferences or whatever,

01:25:16   but like not syncing downloads makes a little bit of sense,

01:25:20   but pictures, well, you've got iCloud photos,

01:25:22   but what if you have other pictures there

01:25:24   and should I put pictures in the pictures folder

01:25:26   and not in documents folder?

01:25:27   Well, then they're not gonna sync

01:25:28   and it's just, it's clear that current day Apple

01:25:31   doesn't consider all of these top-level folders.

01:25:33   It's not on board with the top-level arrangement.

01:25:36   Otherwise, it would sync more of them, right?

01:25:38   And like documents is so generic.

01:25:41   Isn't a movie a document?

01:25:42   Isn't a picture a document, right?

01:25:43   Aren't downloads documents?

01:25:44   What the hell is a document?

01:25:46   Documents is the thing that they sync,

01:25:48   but they're like, okay,

01:25:49   we don't wanna sync people's libraries

01:25:50   'cause we have iCloud Photo Library for that.

01:25:52   And if they actually legitimately have pictures

01:25:54   that are not photos, they should put them in documents

01:25:56   because that will sync.

01:25:57   Same thing with movies.

01:25:58   It's just, it suffers from a little bit

01:26:01   sort of someone had a had a grand plan many many years ago probably back in the next days

01:26:06   and that grand plan has the people who care about or know about their grand plan are long since gone

01:26:11   and then people just don't want to change things but they're happy to ignore stuff there was also

01:26:16   by the way back in the next days in the early mac os 10 days the idea of slash network so there was

01:26:21   slash system slash network there was slash local and then your home directory like and

01:26:26   all those things had a library directory and applications directory under them

01:26:30   That they were sort of the same arrangement of top level stuff repeated many times depending

01:26:36   on the scope.

01:26:38   Just this user stuff.

01:26:39   Just this machine stuff.

01:26:40   The quote unquote "network stuff".

01:26:42   And that has also fallen by the wayside as the internet has taken over as the dominant

01:26:45   network metaphor as opposed to file sharing across local computers and a LAN and all that

01:26:49   other stuff.

01:26:50   So lots of historical baggage here, but that's part of the reason why I just try to stay

01:26:55   the heck out of my top level home directory and do everything.

01:27:01   I don't do documents and whatever, syncing or whatever, but I try to do everything under

01:27:05   the thing they have.

01:27:06   So my movies are in movies.

01:27:08   My pictures are in pictures.

01:27:10   Everything else is in documents.

01:27:11   My downloads are in downloads.

01:27:12   I don't touch sites.

01:27:13   I don't touch public.

01:27:15   My music used to be in music, but I had this notion back in the iTunes days that I wanted

01:27:20   to share my music library between multiple users on the system because it was the biggest

01:27:23   thing on my computer.

01:27:24   before and so my music was in user shared,

01:27:27   which still exists by the way.

01:27:28   User shares music, my music is still in user.

01:27:30   - Wait, you didn't just have like the built in iTunes

01:27:33   share feature enabled?

01:27:34   - Yeah, seriously.

01:27:35   - No, no, like back in the day, if you put your music

01:27:37   in user shared music, it would literally use the same

01:27:41   iTunes library for multiple users.

01:27:43   I don't think it was ever robust enough to have both users

01:27:46   using it at the same time, but it might have been.

01:27:48   But anyway, it worked well for years.

01:27:50   Now my music is still in user shared music,

01:27:52   even though I'm the only one using it.

01:27:54   I really should move it back.

01:27:55   But anyway, my music is so small compared to my photos library

01:27:58   now, it doesn't even matter.

01:28:00   I'm not like Marco with his fish stuff.

01:28:02   But yeah, to answer the question, no.

01:28:04   No top level stuff.

01:28:05   And GitHub checkouts the top level of your home directory.

01:28:08   I say this is madness, but I have

01:28:09   to tell you that I am, in my Unix days,

01:28:13   the top level of my home directory

01:28:15   had way too much stuff in it.

01:28:17   I tried to arrange things nicely on top of my home directory,

01:28:20   and I failed miserably.

01:28:22   Like over the course of years in college,

01:28:25   my grand plans of arranging the top level

01:28:27   of my Unix home directory just were shattered.

01:28:29   And then at my jobby jobs,

01:28:31   if you looked at my home directory of my last jobby job

01:28:34   on my last day of work,

01:28:35   you would be horrified more than Ed Barker's thing.

01:28:37   I had so much crap at the top level of my home directory,

01:28:41   like my network shared home directory

01:28:42   that was mounted every time I SSH'd any machine

01:28:44   anywhere inside the company.

01:28:46   So much, I should have saved an LS output

01:28:49   just like mind boggling amount of crap in my home directory.

01:28:53   That's what happens if you give people a Discord

01:28:55   that are too big.

01:28:56   (laughing)

01:28:58   - For me, I generally don't have much in there

01:29:01   and I consider it semi sacred.

01:29:03   So I do have applications there,

01:29:05   although there's nothing in it.

01:29:07   You know, desktop, of course, downloads, of course,

01:29:08   documents, movies, music, pictures, public, all that jazz.

01:29:13   The only other stuff that I also have in there is

01:29:15   I do have, I don't know, maybe 10 or 15 shell scripts

01:29:18   that I use a whole lot.

01:29:19   I should probably find a different place for them,

01:29:21   but they're all living there.

01:29:22   - Put them in your bin directory, dude.

01:29:23   What are you doing?

01:29:24   - My bad, sorry.

01:29:25   But they're all living there.

01:29:28   - I didn't mention that, by the way,

01:29:29   but I do have top-level Unix directories

01:29:31   that I made invisible, so I have tilde/bin,

01:29:33   where all my executables and scripts go

01:29:35   that's in my shell path.

01:29:37   The Unix-y stuff is there, but I don't see it in the Finder.

01:29:40   - No, that's fair, I should probably do that.

01:29:41   But leaving aside those 10 shell scripts,

01:29:44   I also have a folder.

01:29:47   This is thought technology that I came up with.

01:29:49   I think it was an original invention,

01:29:51   like a year ago, maybe two years ago.

01:29:53   I have a folder called Detritus,

01:29:55   and that's in my top level folder.

01:29:57   And in there is like all the crap

01:30:00   that I don't care if it disappears.

01:30:03   So AudioHijack dumps all of its files there.

01:30:07   Obviously I care if those disappear,

01:30:08   but I handle all of these files immediately

01:30:10   when I'm done recording.

01:30:11   So AudioHijack dumps its stuff there.

01:30:14   Calibre has folders in there.

01:30:16   Final Cut Pro because I'm not anymore doing things.

01:30:20   - What you've got there as a temp directory.

01:30:22   - Yeah, kind of, you're right.

01:30:25   - But you wanted to be snooty and say

01:30:26   Detritus instead of TMP.

01:30:28   - Yeah, I guess, I don't know.

01:30:29   But, you know, Geotag, which is a thing I use

01:30:31   to geotag photos from my big camera,

01:30:33   and Zoom insists on having a download directory

01:30:35   for reasons unknown.

01:30:36   All of that is in Detritus,

01:30:38   and Detritus is in my home folder.

01:30:39   But other than that, in some shell scripts

01:30:41   that apparently I need to move,

01:30:42   oh, and my equivalent of Dropbox,

01:30:44   which is Synology Drive,

01:30:46   That's the only real custom additions to my home folder.

01:30:50   Finally, Kai asks, "Hey Casey, you have sometimes said

01:30:53   that you use spaces a lot as a part

01:30:54   of your window management workflow on Mac OS.

01:30:56   I use spaces a lot in the days of the 2D grid,

01:30:59   but lately with self rearranging list of variable length,

01:31:01   with intermix full screen apps,

01:31:02   I found myself getting lost

01:31:03   and feel I'm not using multiple desktops to their potential.

01:31:05   Could you elaborate on how you use spaces?"

01:31:08   Yeah, there's not a lot to say, to be honest with you.

01:31:12   Typically what I'll do is I'll dedicate a space

01:31:15   to development work so that typically,

01:31:19   if I'm on a larger screen and not just the laptop,

01:31:22   typically that means I'm running Xcode as big as I can

01:31:25   while still having a full-size simulator

01:31:29   on the right-hand side.

01:31:30   So I've got a full-size like iPhone 14 Pro simulator

01:31:33   on the right-hand side of the screen,

01:31:34   and then Xcode takes up the entire rest of the screen.

01:31:37   I'll have a space dedicated to that.

01:31:40   I'll have a space dedicated to real-time communication,

01:31:44   And so that is like 2/3 Slack and 1/3 messages.

01:31:49   So they're both full screen tiled using the semi-new MacOS

01:31:54   affordances to do so.

01:31:56   And then I typically have a space for like web browsing,

01:31:59   email, Twitter, just my junk drawer space, if you will.

01:32:03   The only reason that this works in my brain is because--

01:32:06   I have no idea where it is in system settings-- but in

01:32:08   system preferences, there was a way to say to MacOS, do not

01:32:13   rearrange my spaces.

01:32:15   It tried to get all smart about,

01:32:17   oh, the two things that you use the most,

01:32:19   let's shimmy them next to each other, which I get,

01:32:22   but because I think of this as like this whole, like space,

01:32:25   I think of it as an array of spaces,

01:32:28   it's a lateral list of screens.

01:32:31   When they get rearranged, it breaks my brain.

01:32:33   And somewhere in system preferences,

01:32:34   and I got to assume in system settings,

01:32:35   there's a place to say, don't rearrange them, leave them be.

01:32:39   I don't recall there ever being a two by two grid there,

01:32:42   I'm not saying there wasn't,

01:32:43   I don't remember that being a thing, except on Linux,

01:32:46   like, or like X.

01:32:47   - No, we've talked about this on the show before.

01:32:49   The spaces used to be two dimensional,

01:32:51   not just left and right, but also up and down.

01:32:53   - Man, I do not, I believe you,

01:32:55   I do not remember that at all.

01:32:56   But one way or another, so, you know,

01:32:59   as they're all a list, a lateral list,

01:33:01   once I turned off rearranging, it got much better.

01:33:04   And like I said, it's basically like junk drawer,

01:33:07   development work, communication, and that's about it.

01:33:10   There's not a lot of science to it.

01:33:12   Do you guys, both of you swear that spaces

01:33:15   are pretty much evil, is that correct?

01:33:16   - Yeah, I don't think they're evil,

01:33:18   I'm just super not into them.

01:33:19   I wasn't into the other spaces with the 2D either.

01:33:22   - Marco, same story.

01:33:23   - Yeah, I've never really gotten into

01:33:25   like virtual desktop kind of like, you know, sets like that.

01:33:27   I just never, it's never been my thing.

01:33:29   - I mean, that's totally fair.

01:33:30   I feel like I couldn't live without it,

01:33:32   but I know that makes me a bit of a weirdo.

01:33:34   Hey, I think that's it.

01:33:36   - Cool, thanks to our sponsors this week,

01:33:38   Memberful, Masterclass, and Nebula,

01:33:41   And thanks to our members who support us directly.

01:33:43   You can join us at ATP.fm/join.

01:33:47   And we will talk to you next week.

01:33:50   (upbeat music)

01:33:53   ♪ Now the show is over ♪

01:33:55   ♪ They didn't even mean to begin ♪

01:33:57   ♪ 'Cause it was accidental ♪

01:33:59   ♪ Accidental ♪

01:34:00   ♪ Oh, it was accidental ♪

01:34:02   ♪ Accidental ♪

01:34:03   ♪ John didn't do any research ♪

01:34:05   ♪ Marco and Casey wouldn't let him ♪

01:34:08   ♪ 'Cause it was accidental ♪

01:34:10   ♪ Accidental ♪

01:34:11   is accidental. And you can find the show notes at ATP.fm.

01:34:18   And if you're into Twitter, you can follow them at

01:34:24   C-A-S-E-Y-L-I-S-S, so that's Casey List M-A-R-C-O-A-R-M

01:34:32   ♪ Anti-Marco Armin, S-I-R-A-C ♪

01:34:37   ♪ USA, Syracuse, it's accidental ♪

01:34:41   ♪ It's accidental ♪

01:34:43   ♪ They didn't mean to ♪

01:34:45   ♪ Accidental ♪

01:34:46   ♪ Accidental ♪

01:34:48   ♪ Tech podcast so long ♪

01:34:51   - Hey, so we have potentially a problem.

01:34:57   What are we gonna do about our theme song?

01:34:59   - Many weeks, many, many weeks

01:35:01   people have been making the same joke and asking the same question, which I don't blame them for.

01:35:05   It makes perfect sense. In our theme song that you just heard, there is a section where

01:35:09   our Twitter handles are read out, without the @ signs, just with the letters so people can figure

01:35:15   it out. And everyone's going, "Oh, you know, Twitter's going downhill and you're all on

01:35:19   Mastodon and Marco recently is off of Twitter and he took the overcasting out of Twitter,

01:35:24   so you're gonna have to change the theme song, aren't you?" Here's my, was and is still my

01:35:29   opinion on this and I'm sure it's not going to be a particular popular one, but it is what it is.

01:35:33   The song lyrics are, luckily, "If you're into Twitter, you can follow them." Now,

01:35:40   the qualifier is important because, say, Twitter goes away or Twitter gets evil,

01:35:48   you can say, "Well, I'm just not into Twitter, so I'm not going to follow them." Right?

01:35:52   The other possibility is, "If you're in Twitter, you can follow them. What if Marco deletes his

01:35:56   account that would mean that even if you aren't at Twitter you can't follow him

01:36:01   because he's not there but so far we're still in the good his account still

01:36:04   exists and you could attempt to follow him there I think his account is private

01:36:09   now it is your request to follow may not be approved I don't even know where to

01:36:14   look to approve those requests I'm not approving it you work like I locked the

01:36:18   accounts in part to you know just see like you know I'm just gonna put a cap

01:36:21   on this preserve the name right yeah yeah I'm preserving the name but but

01:36:24   yeah, I'm just kind of putting a cap on my usage of this,

01:36:27   like all right, no more, you know, and look,

01:36:29   I'm not saying I'm gonna be gone forever, definitely.

01:36:31   - Yeah, Twitter could rise again.

01:36:32   Many things could happen to Twitter.

01:36:34   - Yeah, but yeah, frankly, I'm not super optimistic.

01:36:38   - But anyway, the theme song has always had that out.

01:36:41   If you're into Twitter, you can follow them.

01:36:43   Now, and the other angle is, again, say Twitter goes away.

01:36:47   It goes bankrupt, it goes out of business,

01:36:48   no one buys the assets, it gets sold for parts,

01:36:50   Twitter no longer exists.

01:36:52   Do we still have this theme song that has Twitter in?

01:36:54   I, being a nostalgic person, would say yes.

01:36:56   This theme song was written back when Twitter existed.

01:36:57   You remember those days?

01:36:58   Wasn't that fun?

01:36:59   And we had Twitter handles or whatever.

01:37:01   So I would be perfectly fine.

01:37:03   Not ever changing the song,

01:37:04   just saying the song is a historical artifact.

01:37:07   It always had a qualifier on it.

01:37:09   Even if Twitter becomes entirely evil,

01:37:10   the song is not endorsing Twitter.

01:37:12   It has always been noncommittal.

01:37:13   It's like, hey, if you're into Twitter,

01:37:15   you can follow them and then it reads our handles.

01:37:17   But the rest of the world does not agree with me.

01:37:19   The rest of the world thinks Twitter should not be mentioned

01:37:21   in the theme song if we are not actively on Twitter.

01:37:23   I'm still on Twitter, by the way,

01:37:24   but if we're not into Twitter anymore,

01:37:27   we shouldn't tell people to try to follow us there.

01:37:29   Luckily, our handles are all the same on Mastodon,

01:37:33   not the other part after the ad,

01:37:35   but if you type in C-A-S-E-Y-L-I-S-S, that's Casey List,

01:37:40   just don't type that's Casey List, you will find him.

01:37:43   If you type my last name, you'll find me.

01:37:45   If you type Mark Armand, you'll probably find him.

01:37:47   Like, we are find a bull.

01:37:49   So my second level of bargaining is, okay,

01:37:52   you demand that we change the song,

01:37:54   I still kinda like the reading of the handles.

01:37:56   I don't think you need to put the whole big thing

01:37:58   in Mastodon.social, 'cause then what about

01:38:00   when we move Mastodon servers, 'cause the whole point is

01:38:02   you can move servers really easily,

01:38:03   and it's a Fediverse, and blah, blah, blah, blah, blah.

01:38:05   So I'm like, let's not go overboard here

01:38:08   and just totally change everything.

01:38:10   Plus the rhythm of that part of the song works well,

01:38:12   reading the handles, it teaches people how to spell my name,

01:38:14   which I appreciate.

01:38:16   And it doesn't, you know, it does mention Twitter,

01:38:18   but it doesn't say anything on a particular service.

01:38:19   Let's just leave that, let's just change the

01:38:21   if you're into Twitter line to something else.

01:38:23   Maybe we'll do that.

01:38:24   So if the world says we must change the song,

01:38:28   I would hope that we will change it in a minimal way.

01:38:31   Right now we have done nothing,

01:38:33   which you will be aware of when you hear this song

01:38:36   that hasn't changed most likely on this thing,

01:38:37   but we are considering things.

01:38:39   - Yeah, I too would, I see the,

01:38:44   'cause look, here's the thing.

01:38:45   People love our song.

01:38:47   I love our song, we love our song.

01:38:49   And so I don't want to rewrite the whole theme song

01:38:54   without that line in it or whatever else,

01:38:56   or have to spell out @marcoarment@mastodon.social.

01:39:01   (laughs)

01:39:03   That's not gonna happen.

01:39:04   So I'd be inclined to take advantage

01:39:06   of the situation we're in, as Jon mentioned,

01:39:09   where our usernames are the same,

01:39:12   and just change it to something like,

01:39:14   if you're into social, or on your social network,

01:39:17   or something that has the same number of syllables

01:39:19   that is network agnostic.

01:39:21   - Yeah, we wouldn't want to put Mastad on,

01:39:23   because that would be making the same mistake again.

01:39:25   Like Twitter, we had a good run here.

01:39:26   Twitter's been around for a long time.

01:39:27   Maybe we'll Mastad on,

01:39:28   but what if we put app.net into the song?

01:39:30   Not mentioning a company that could go out of business

01:39:33   or become evil is probably a good plan.

01:39:35   - Right, and it seems like whenever

01:39:38   a new social service takes off,

01:39:41   we usually have these exact same usernames on it.

01:39:45   Or at least we can make it happen.

01:39:47   - We try to.

01:39:49   But the other thing is, and I talked to Jonathan Mann,

01:39:52   the composer and performer of our theme song,

01:39:55   to see like, hey, would you be willing

01:39:57   to do something new here or modify this?

01:40:00   And he said yes, so we just kind of,

01:40:02   basically we have to figure out what to do.

01:40:04   And the thing is though, because this song

01:40:07   is almost 10 years old, whatever we do now,

01:40:12   even if we have him record one line and try to drop it in,

01:40:16   it's gonna sound different.

01:40:18   I believe in our ability to make that work.

01:40:20   Computers, audio processing, Jonathan Mann

01:40:24   pitching up his voice and not sounding,

01:40:26   it'll be, I think we can pull it off.

01:40:27   But anyway, we'll try.

01:40:28   But we have a similar situation with,

01:40:30   'cause the theme song is so old,

01:40:31   the whole, people were making the joke when I quit my job,

01:40:33   saying, "Oh, Jon didn't do any research."

01:40:36   People don't even know what that line is about.

01:40:37   It's so long ago that that was a thing on the show.

01:40:41   But the people who do know say,

01:40:42   "Oh, now that Jon doesn't have a job,

01:40:43   "you have to change the themes."

01:40:44   No, we don't have to.

01:40:45   It's fine as a historical artifact.

01:40:47   Like when this song was written, that was a gag in the show, right?

01:40:50   So many things in the song are from like, when was it written?

01:40:53   Like pretty early in the run of the show.

01:40:55   I mean, it had to be like mid 2013 maybe?

01:40:58   Yeah, we can go back and find the first show with the theme song,

01:41:01   but like they're all like references and gags that could only have happened

01:41:05   in the very, very beginning of the show because that's when the song was written.

01:41:08   And I enjoy the fact that it exists as a historical artifact.

01:41:11   The same reason I have my wedding photo on the wall.

01:41:14   Like we don't look like that any wrong, but that's when we were married.

01:41:16   So the photo is still hanging.

01:41:18   We don't say, "Oh, we should update your wedding photo

01:41:19   "to have pictures of you being old and wrinkly."

01:41:21   Like, "No, thanks."

01:41:22   (laughing)

01:41:23   - 26th of March, 2013.

01:41:25   So that was real early in the run of the show.

01:41:28   They must, it was like--

01:41:29   - Yeah, that was what, like a month in?

01:41:30   - Maybe. - If that, even?

01:41:32   - What episode number was it?

01:41:33   - I don't know, I'm just looking at the YouTube video

01:41:35   he posted, it was March 26th, 2013.

01:41:37   - That's awesome.

01:41:38   - That's episode seven.

01:41:39   - Golly.

01:41:40   - Oh my God, that's amazing.

01:41:42   - So episode seven, I would say is pretty early.

01:41:44   So we had enough gags by episode 7 for a theme song to be written that lasted us until episode

01:41:51   whatever the heck we're on 400 and something.

01:41:53   500.

01:41:54   Whatever, I can keep track.

01:41:56   How old am I?

01:41:57   What's 100 episodes here and there?

01:41:58   500, 500.

01:41:59   We're recording 514 and this theme song is written in episode 7.

01:42:02   So that's why I'm a proponent of we'll just never change the theme song and it'll just

01:42:06   be a historical artifact but we are exploring possible changes.

01:42:09   So if you see some changes and we do have alternate versions of the theme song we do.

01:42:13   There was the bleeps and boops version that I was a fan of that we never actually put

01:42:17   into the show.

01:42:18   Go back and see.

01:42:19   It wasn't terrible.

01:42:20   It was awesome.

01:42:21   It just was incomplete.

01:42:22   I mean, it wasn't terrible.

01:42:23   It was just nowhere near as good as the regular version.

01:42:25   It was good in a different way.

01:42:27   Was it, though?

01:42:28   It was good in a bad way.

01:42:29   No, it was good in a more narrow way.

01:42:33   Marko should appreciate this.

01:42:34   Just like fish.

01:42:35   It was good in a way that is of interest to a narrower slice of the population, let's

01:42:39   say.

01:42:40   Yeah, let's go with that.

01:42:41   And yeah, let's go with that.

01:42:44   - Anyway, so we'll figure something out.

01:42:47   And yeah, and I think odds are that,

01:42:50   I would say odds are even between changing nothing

01:42:56   and changing like that one line

01:42:58   and trying to make it sound natural

01:43:01   with the rest of the song staying the same.

01:43:03   - Yeah, it relies on Jonathan having the individual tracks,

01:43:06   which given that he writes a song every single day,

01:43:08   does he keep the individual tracks?

01:43:10   I know Marco being a pack rat does keep like the,

01:43:12   he keeps the logic files from all the ATP episodes.

01:43:15   - Yeah, I keep everything.

01:43:16   - Right, but I'm not sure if Jonathan Mann

01:43:18   has the hard drive space to be keeping individual

01:43:20   track files for every song that he's ever written.

01:43:21   So I'm just crossing my fingers that he somehow,

01:43:24   you know, stashed away the individual tracks

01:43:27   for this one song.

01:43:29   - I wonder if, I wonder if we could train an AI model,

01:43:33   just feed them all of Jonathan Mann's songs.

01:43:36   - Oh my gosh.

01:43:37   And just have it, you can put it into Descript,

01:43:39   I'm sure Merlin is doing this right now.

01:43:41   You know how much he loves it?

01:43:42   Descript is a audio--

01:43:44   I am saying it like him now.

01:43:45   He says "descript."

01:43:46   The word is "descript."

01:43:47   Descript.

01:43:48   I don't-- anyway.

01:43:49   It's like description.

01:43:50   Anyway, it's an audio editor that's really cool.

01:43:52   It lets you edit text by looking at--

01:43:54   edit audio by looking at text.

01:43:56   So it does a transcription.

01:43:57   And then you just move the words around.

01:43:58   And when you move the words around in the text,

01:44:00   it moves the audio around.

01:44:01   And it has a feature where you can feed it samples of someone's

01:44:04   speech.

01:44:04   And then you can just type something you want them to say,

01:44:06   and it will say it in the voice of that person.

01:44:08   And it's amazingly good.

01:44:10   - Creepiest.

01:44:11   - But I don't think it does singing.

01:44:13   - Yeah, yeah, I doubt it.

01:44:15   But hey, I mean, look, with all the things

01:44:17   the AI can do these days, we're very close.

01:44:19   - Oh no, I'm sure this is within, someone could do this,

01:44:22   it's just that Descript, as Marilyn would say,

01:44:25   does not have this capability.

01:44:26   - Yeah, and to be clear, that's mostly a joke.

01:44:29   Like we actually are gonna have Jonathan

01:44:31   do something for us, but yeah.

01:44:32   - I'm saying this could be what Jonathan does for us.

01:44:34   We don't know, we just wanna see the finished product.

01:44:36   - Yeah, I guess it's--

01:44:37   - He can use whatever tools he wants to get it done.

01:44:38   his own voice. Yeah, I don't know. I sitting here now, I think, John, you have convinced

01:44:45   me that it doesn't need to be modified. I can't tell when people are commenting on it

01:44:49   if they're snarking for the fun of it or if they're like legitimately kind of whining

01:44:54   about it. I don't think that's the case, generally speaking. They just make it a joke. But the

01:44:58   thing is, what we have to understand is if we don't change it, how many years do we have

01:45:01   to hear them? I guess that we change the theme. I think it'll probably die down after about

01:45:06   a year.

01:45:07   I'm willing to look into changing it for the purpose of future proofing, but I'm also super

01:45:11   okay with it being the same.

01:45:13   Whenever anyone asks about it, now I'll be able to give them a timestamp linked to this

01:45:16   after show to explain why we haven't changed the theme song if we haven't changed it, or

01:45:20   why we did change it if we did.

01:45:21   [BEEPING]