PodSearch

ATP

689: The Positive Effect of Enthusiasm

 

00:00:00   So now you missed my amazing joke, which really wasn't that amazing.

00:00:04   I did, yeah.

00:00:04   Well, it's quite alright.

00:00:06   It's only uphill from here.

00:00:07   Alright, downhill?

00:00:08   Nobody ever knows.

00:00:10   It's only up from here, looking up from here, falling down from here.

00:00:13   A debate that Erin and I had, which I reached a conclusion to, but I don't think she was paying any attention because she could not possibly care less, was...

00:00:21   You lost her one sentence in.

00:00:22   Yeah, pretty much.

00:00:23   If I want the air conditioning to be colder, am I turning it up or turning it down?

00:00:29   Yeah, I feel like that, when it comes to air conditioning season, I feel like that catches everybody.

00:00:35   It's like, oh, turn it up.

00:00:37   Well, make it stronger?

00:00:39   Which direction is that?

00:00:40   Turn the temperature up, which makes it weaker?

00:00:42   Right.

00:00:43   And I feel like it's one of those situations, like Settlers of Catan trades, where you just have to add more words to clarify what you mean.

00:00:51   Anybody want to trade wood for sheep?

00:00:53   Oh, wait, no, sorry.

00:00:54   I have sheep.

00:00:56   I want wood.

00:00:57   Oh, okay.

00:00:59   God, I haven't played that game, possibly since you and I and Tiff and Aaron played it 84 years ago, but...

00:01:04   And everybody always has too much sheep, and...

00:01:07   Look, everyone always has...

00:01:08   Everyone always needs brick.

00:01:08   You know, everybody...

00:01:09   Anybody want to trade sheep?

00:01:11   No, I...

00:01:11   No one has brick, you know, Roger.

00:01:13   Like, pay attention.

00:01:14   It's too bad this isn't in the show.

00:01:21   Are we ever going to talk about Mythos?

00:01:22   Mythos?

00:01:23   Probably not.

00:01:23   What is there to talk about?

00:01:24   Wait, what are we talking about?

00:01:26   The, the, uh, the model from Anthropic that is...

00:01:30   Oh, that thing.

00:01:30   Yes, yes, yes.

00:01:31   Sorry.

00:01:31   It's so dangerous.

00:01:32   We can't let you have it.

00:01:33   All right, so I'm actually legitimately freaked out about it.

00:01:36   I gotta say, like, I'm, so I'm actually, like, I was planning a, a big overcast server upgrade,

00:01:42   um, and I'm accelerating those plans now.

00:01:46   Just, just only, not to, like, unlock any new functionality, really, just, like, you know,

00:01:51   I run a certain distribution of Linux that is a few years old now, and it's, it's, you know,

00:01:55   it's starting to leave its supported time range for certain patches of certain packages and

00:02:01   stuff like that, and I'm on, like, I'm on the other side of the big OpenSSL upgrade as well,

00:02:06   so I'm like, oh boy, there's no way to really upgrade that in place without breaking everything,

00:02:10   so I'm, I'm gonna, you know, slowly migrate to new servers, but, like, now I'm like, oh,

00:02:15   maybe I should do that a little more quickly than, than before, um, I don't know, you, are

00:02:20   you two really not, uh, you don't think it's gonna be a thing?

00:02:23   Uh, Mythos is just a marketing BS, but the idea that LLMs can find security flaws is 100% real,

00:02:29   and Mythos or not, yeah, you shouldn't upgrade your software to versions that you hope have less

00:02:35   security flaws in them, because LLMs may get a lot easier for lay people to find those,

00:02:40   flaws, so, yeah, I agree with the upgrade process, I just think the whole Mythos thing is silly

00:02:44   overblown marketing is done. Yeah, I, I think, and again, like, I think whether, whether or not

00:02:50   Mythos specifically is, okay, so just to explain, anybody who, who missed the story quickly, um,

00:02:55   Anthropics, uh, latest model Mythos, they've only released in a very limited way to, like,

00:03:01   approved people, because what they say is, it's incredibly good at finding security vulnerabilities

00:03:08   in source code, and they, it's already found, like, hundreds of vulnerabilities in old packages,

00:03:14   like, you know, like, OpenBSD and FFMPEG, like, you know, like, old software that's been around

00:03:19   forever, that's, like, run on servers and stuff, um, and so the concern is, if this gets into the

00:03:25   mass market, lots of people can find security vulnerabilities and exploit them, um, and so

00:03:30   what they are doing is having, like, a preview program where they're allowing certain software

00:03:35   vendors to have access to it so they can run it against their code, patch their code, and, you know,

00:03:41   get those patches released before, um, this is available to the public. Now, this is not a perfect

00:03:46   system, you know, Anthropic, as, as with most AI companies, has a history of overhyping, um,

00:03:52   the capabilities of their models, and kind of, you know, calling Wolf, so to speak, and, and, you

00:03:56   know, as... It's a good marketing, right? It is, and, you know, and as Ben Thompson points out, like,

00:04:00   the, the story of the Boy Who Cried Wolf does end in an actual Wolf. Yeah, although it's obvious that,

00:04:05   uh, that they didn't run, uh, Mythos on their own software, because, uh... Yes. People exploited

00:04:10   flaws, and, and, like, it got, it got, it got, a whole Discord got access to Mythos and has had it for

00:04:15   months, uh, so, yeah, you're not actually keeping this genie in the bottle, like, oh, we're only gonna

00:04:20   release it to a select group of people. That is extremely difficult to do, and they immediately

00:04:25   screwed it up by having security flaws in their own software that were exploited by people who just

00:04:28   wanted access to it. But, like, by setting this aside, like, as, as I discussed, you know, when we

00:04:32   were talking about things you can use LLMs for many, many episodes ago, you can just ask it to find bugs

00:04:37   in your, in any code, or you could ask it to find security flaws, uh, and it is a tireless searcher for

00:04:43   security flaws. Uh, you know, it will find things that have not been found by other people, because,

00:04:48   you know, it's, it's good at finding them, and it doesn't sleep. So, one thing I have found with

00:04:53   using LLMs to analyze, you know, documents, or, or code, or whatever it is, is, like, they are often

00:04:59   able to find correlations between different sections or different parts that I wouldn't have seen,

00:05:05   um, you know, just because they, they don't work like human brains. They work differently.

00:05:10   Yeah, and also, because they are tireless, like, they're finding bugs in really old code, because no one

00:05:15   has looked at, no human has looked at that code in 20 years. If a human did look at it and concentrate

00:05:20   on it, yeah, they would find the same flaws. But no one's looking at that to, like, oh, that's old,

00:05:23   it's debugged, it's fine, it's good. But a computer doesn't have the same, you know, like, oh, I won't

00:05:28   look at that code, because it's really old, and I don't need to bother looking at it. No, you just

00:05:31   pointed at the whole code base, and it just goes through all of it. So, it's not like it's even better

00:05:34   than humans at it. It is just simply, it does the things that humans don't, which is, I'll look at

00:05:39   everything forever, as long as you ask me to. Yeah, and yeah, and I'll never get tired, and I'll

00:05:43   never get bored, and I'll never zone out, and I'll never complain that it's really boring. Yeah, and

00:05:46   I'll never, like, skip over stuff, because, like, oh, obviously, that's fine. I won't do that, because

00:05:51   I don't know if it's, I just do what you tell me. Here's the code, I'll look at it. Yeah, exactly. So,

00:05:55   anyway, so, I think, regardless of whether, whether the hype about Mythos is real or valid or earned,

00:06:02   like, regardless of that, the truth is that we are in a transition point now, where, like, in the

00:06:09   long run, having LLMs look at a bunch of source code at the time will be great for security, because

00:06:14   I think fewer security vulnerabilities will get released in the first place, because the LLMs will

00:06:18   either be writing the code and will avoid them, or they will find them. I don't think they avoid them

00:06:22   when they're writing the code. I think they write plenty of themselves, because they're copying other

00:06:25   code that has security flaws. But yes, you can have a phase where you say, now, just keep looking for

00:06:29   security flaws until you think you can't find any anymore. And then the new version of the model

00:06:33   comes out, and you ask it to do the same thing over and over, right? But in the meantime, like,

00:06:36   you know, the transition between having all this sloppy human code that is full of vulnerabilities

00:06:41   nobody's ever found, and the future where LLM code is mostly what's out there, or LLM vetted human code,

00:06:48   or whatever, or LLM assisted human, you know, whatever it is. This transition period, I think it's going to be

00:06:53   just a tidal wave of patches coming out from lots of places. And so, like, what I told my family,

00:07:03   you know, when this news came out a few weeks ago, I'm like, listen, whenever you get an update on your

00:07:09   phone or your devices, just accept it, run it. Like, don't defer software updates for six months. But,

00:07:15   you know, like, this is a good time to be on the latest versions of everything that you can be on the

00:07:21   versions of. You know, certainly on our personal devices, that's a good idea. You know, I think we

00:07:25   can count on the major vendors, you know, Apple, Microsoft, you can count on them to probably

00:07:31   issue critical security patches for older OSs back to a point. And Apple has actually done that

00:07:39   recently. They've actually issued some very old OS patches recently for, like, some pretty big bugs.

00:07:44   But that's only to a point that they'll do that, and only for, like, the worst bugs, probably.

00:07:50   And so, this is a really good, like, the way I think it's wise to deal with,

00:07:55   whether it's mythos or mythos-like things in the present or future, is just be running the latest

00:08:02   versions of everything you possibly can. And when patches come out, just install them as fast as you

00:08:07   can. This is not a good time to be running old software. And unfortunately, on servers, it's a lot

00:08:11   more difficult, you know. And so, that's why, like, I'm embarking on this project to do this giant

00:08:15   rolling upgrade, basically. It's not fun, but I feel like, you know, if Overcast got hacked because

00:08:20   I didn't upgrade my Linux version from one from 2024 or whatever, like, I'm going to feel pretty bad

00:08:26   about that. Well, the tricky part of just upgrade to the latest version whenever it comes out is you

00:08:29   got to hope those latest versions don't have a bunch of vibe-coded stuff that adds a bunch of

00:08:33   horrendous security flaws in it. So, you know, it's a double-edged sword because, yes, and to be clear,

00:08:39   you don't need to use Mythos to find security flaws. Any of the existing coding models, they can

00:08:44   and will find security flaws, right? Not just Mythos, right? But also, people will use them carelessly

00:08:49   to create security flaws because, you know, just because they're good at finding them doesn't mean

00:08:53   they won't create them themselves. And because you can create so much code so much faster these days,

00:08:57   like, I just saw a stat fly by today that GitHub, for all their problems, they're having a bunch of

00:09:01   problems related to Microsoft being Microsoft or whatever and absorbing that company. But setting that

00:09:06   aside, one stat that they threw out was that the number of commits going into GitHub has increased

00:09:12   by a literal order of magnitude since this time last year. Yeah, that's wild. So, and where do you

00:09:18   think all those new commits are coming from? Did we get 10 times as many programmers in the last year?

00:09:22   We did not. So, yeah, it's a wild time out there in the software world, you know, and we can use the

00:09:30   tools to help protect us, but those same tools are also kind of screwing us over. And there were some,

00:09:35   I don't know if any of this is true, but there's a bunch of narrative that part of GitHub stability

00:09:39   problems is because the people at GitHub themselves are vide coding a bunch of things and it's destroying

00:09:42   their stability. But other people say just being bought by Microsoft destroyed their ability because

00:09:46   anyway, I'm not even weighing in on that, but just simply the stat that 10 times more commits are

00:09:52   going to, or was it 14 times more commits are happening in GitHub versus a year ago is a fairly large

00:09:59   change. It's difficult for GitHub to handle. And it is indicative that more code is being produced

00:10:04   than you would expect.

00:10:06   Yeah. But, and honestly, like, you know, I don't know what's going on with GitHub, but

00:10:09   if I ran a giant web service and it had 14 times the traffic, uh, in, in, you know, a few months,

00:10:17   I thought you did that with Tumblr.

00:10:18   Yeah. We, we grew like 20%, uh, you know, like 20% at a time, basically like this, this is very

00:10:26   different. This is like 14 X in what is it? Four months. They said they're on pace for 14 X in a

00:10:31   year. If current, if current, the current curve continues, but still like that's also, they're

00:10:35   starting from a larger scale. So like that's, they're already pretty significant. Yeah. So like,

00:10:39   I, I, I don't, I don't know what the internal politics are there, but I don't, I certainly

00:10:44   don't blame them for having some uptime issues when they're growing at that rate. All of a sudden

00:10:48   that is, that is significant.

00:10:50   Well, that's the, that's the problem that if you look at the graph of their stability,

00:10:53   the uptime issues happen coincident with the Microsoft acquisition and not coincident with

00:10:57   the rise of coding agents. So we're not entirely, it's not entirely sure where the cause is there,

00:11:02   but, uh, I'm sure that story will be told eventually.

00:11:04   Yeah. Anyway, this is a good time to run updated software. So I, I suggest everyone out there

00:11:09   keep those patches installed, like get, install them when they come out, keep current. This is a bad

00:11:14   time to hold onto an old version as much as Mac OS Tahoe sucks.

00:11:17   Yeah. Unless you're running Mac OS 15, because Apple has been patching it. So good, good on Apple.

00:11:21   Yeah. So far.

00:11:22   All right. Let's do some followup. Uh, I wanted to reiterate, even though the store is closed,

00:11:28   it happened yet again. I will save this person's name. I will

00:11:34   not share their name to save their dishonor, but an anonymous person wrote in and wrote the

00:11:40   following. I thought I would remember as I always have. I did remember on Monday, the 27th of April

00:11:45   at 1 34 AM Pacific time, the store was closed. I wanted to order the RIP Mac pro shirt to go with

00:11:51   my believe version. RIP me, I guess, pulling over next time. This is why we say to pull over the car

00:11:58   people every time, every time. Don't be that person next time. Every time.

00:12:02   There was more than one person. I saw at least one other report. It happens, you know,

00:12:06   it happens. It happens. Uh, but there is some good news, John. Can you share the good word?

00:12:10   Yeah. The ATP Neo silver shirt did in fact get enough people to order it to have the shirt printed

00:12:16   in the end, a total of 34 people bought it. So we had the first week, one person bought it the

00:12:20   second week, two people. And then the third week, everyone came storming in and they passed that,

00:12:24   you know, threshold of a dozen orders in the end, 34 people ordered it. It will be printed.

00:12:28   Congratulations. ATP Neo silver shirt owners.

00:12:32   That grew faster than GitHub. Yeah. Wow. Uh, is that the most rare piece of ATP merchandise?

00:12:38   Cause remember that we had those misprints on, I think the very first run.

00:12:41   Yeah. But I think we sold more than 34 of the misprints.

00:12:43   Yeah. That's what I'm saying. I think this might be the most rare piece of ATP merchandise.

00:12:47   Nah, we've sold fewer than 34 from other shirts. Normally what happens is the whole thing with

00:12:52   the silver is what it took so long to go over the threshold or you've had things go over the

00:12:55   threshold pretty quickly, but still only end up sending like 16 or 20. So I don't think it's that

00:12:59   rare.

00:12:59   No, either way. All right. Uh, we have a handful of very grumpy people from the university of

00:13:04   Pennsylvania who took issue with John. I'm going to say, and this is hyperbolic pooh-poohing

00:13:09   their engineering program. I, that is again, hyperbolic John, would you like to read?

00:13:13   It wasn't, yeah, it wasn't pooh-poohing their engineering program. It was, I was saying

00:13:17   that we were talking about John Ternus last week, uh, for a very obvious reason. Um, and

00:13:21   I noted that he just had an engineering degree from the university of Pennsylvania. And I was

00:13:25   saying how that's a fairly unimpressive, fairly unimpressive pedigree for someone who is the

00:13:30   CEO of one of the biggest companies in the world. Uh, just an undergraduate degree, just

00:13:34   an engineering degree, and just from university of Pennsylvania and everyone in the world wrote

00:13:38   in to tell us, to remind us that university of Pennsylvania is in fact an Ivy league

00:13:42   school. Uh, I did not remember that at all because Brown fills that slot in my mind as

00:13:47   the school that you never remember is an Ivy league school. It really is. Um, we'll put

00:13:51   a link in the show notes to Ivy league schools. If you're not in the U S and don't know what

00:13:54   they are, you can read the Wikipedia page, but it's a bunch of supposedly prestigious schools

00:13:59   for reasons that are rooted in history more than anything else. Uh, Kieran Healy writes when

00:14:04   I was a grad student, the joke amongst obnoxious Princeton undergraduates about UPenn was that

00:14:09   the students there thought the full name of the school was UPenn and other Ivy's because

00:14:14   to see the people who were there, it's like, uh, would always remind you. And you know,

00:14:18   I go university of Pennsylvania, which is of course one of the Ivies. Um, and then Kieran

00:14:22   continues. There's also a trollish bit where people deliberately confuse university of Pennsylvania

00:14:26   and Penn state, which are two entirely different schools.

00:14:29   Indeed. And one of them is Ivy league and one of them is not. Uh, and then Marina

00:14:33   Eppelman writes, arguably engineering in Penn state is superior to UPenn. That was my other point that

00:14:38   like, you know, setting aside a university of Pennsylvania being an Ivy league school,

00:14:42   I don't think it's particularly known as like an amazing, uh, engineering school, but to our credit

00:14:47   and to Casey's credit, we did, Casey did look up on last episode what it's ranked. And so it's not like

00:14:53   we, we said, Oh, this is a terrible school. And then we didn't bother looking into it further.

00:14:56   We looked up, you know, Casey looked it up in real time and said, um, on last episode

00:15:00   that UPenn is number 16, uh, ranked in engineering and for whatever us news and world's report is

00:15:05   worth. And then BU and Virginia check are 32. And what is PSU? Penn state university. Yeah. Penn

00:15:11   state is number 28. So I'm not saying it's a bad school or anything, but my, my point stands,

00:15:14   which is like, he's just got an engineering bachelor's degree. What I should have, the other

00:15:18   point that I should have made, he doesn't even have an MBA. Like if you're going to be like

00:15:21   an executive, very often you'll get like an engineering degree, but, but eventually you decide

00:15:25   like, actually I want to be like a C level executive or something. And so you get an MBA

00:15:29   on top of your, uh, undergrad or you'll get an undergrad and a master's and an MBA. Like

00:15:33   unless you're a founder of a company, you tend not to see someone going to the C level of a

00:15:39   company with, uh, without more degrees. That was my only point that, and, and, and I said,

00:15:45   you know, it shows that you don't need to worry about getting, uh, you know, a bunch of fancy

00:15:49   degrees from fancy schools to become a CEO, but obviously, uh, UPenn being, uh, a, uh, an Ivy

00:15:54   league kind of undercuts that a bit, but still just having a bachelor's degree and not being a

00:15:58   founder of a company, uh, pretty good job. And then other people argued, well, he did that a long time

00:16:02   ago and things have changed and now you actually do need the degree. And I'm not sure that's the case,

00:16:06   but, uh, all I'm saying is don't let your education, uh, dictate your destiny.

00:16:10   Also, thank God the new CEO of Apple doesn't have an MBA.

00:16:13   I mean, it's, it's rare. Like I'm obviously Apple's a special case, but look at, look at the CEOs of

00:16:19   most other companies. You will see that like, they're either like business people or have business

00:16:23   degrees or have some other degree of whatever they're interested into. And then they got an MBA

00:16:27   because that's kind of how you get into the, like, I want to be, you know, it management, uh, in,

00:16:32   in the C level. Eventually you get an MBA and you tack it onto whatever your other degree was.

00:16:37   All right. We have a kind of Mark Gurman corner. Uh, first of all, Gurman reported on

00:16:42   Tim Cook's remarks to Apple employees regarding his pending retirement or, or I don't know,

00:16:47   not retirement, but stepping down as CEO, stepping up to be executive chairman. I don't know. He's

00:16:52   stepping somewhere, stepping on the, onto the roof. Uh, so he was, he spoke to Apple employees on

00:16:57   Tuesday, the 21st of April, a couple of quotes that Gurman reports. I'm excited to continue my

00:17:02   journey at Apple as executive chairman. I am healthy. My energy is high and I plan to be in this

00:17:06   new role for a long time. That, that addresses the point we talked about last time when I asked,

00:17:10   do you think, how long do you think Tim Cook will be executive chairman at Apple? And I said,

00:17:15   I don't think he'll ever leave. And so what he's saying is, first of all, I'm healthy, which is a

00:17:20   way of his way of saying, I'm not stepping down as CEO because I'm about to die, right? Like it's not,

00:17:25   I don't, there's some health thing, you know, I'm healthy. My energy is high and guess what? I plan

00:17:29   to be here for a long time. I'm not just being executive chairman in the next year. I'm going to retire and

00:17:33   go fishing. That's not his plan currently. So all this checks out. It's going to be tough to get

00:17:39   rid of Tim Cook if he's healthy, exercising, and does not really ever feel like leaving as executive

00:17:44   chairman. Tim continues, I'll be there to support, excuse me, here to support John in any way he needs

00:17:50   and in any way I can. I'll be here to offer my knowledge and experience and be a sounding board

00:17:54   anytime I'm called upon. Apple will be my top priority. It's who I am at my core. That's a good

00:17:59   apple joke. And I can't imagine it any other way. Again, that's a, that's an apple lifer talking

00:18:04   right there. Cook finalized, uh, there can only be one CEO at a time, which is a good way of saying

00:18:11   I'm not going to be in the way. Hopefully. Yeah. I mean, certainly executive chairman of the board

00:18:14   should be a much easier, more relaxing job for Tim than being CEO, obviously. And so congratulations to

00:18:21   him to hopefully lowering his stress levels. Then with regard to Mark, Mark Ternus. Wow. Let me try

00:18:27   that again. Mark Gurman with regard to John Ternus. There we go. Uh, Gurman writes, Ternus was a

00:18:33   champion of the MacBook Neo. He urged the company to sell a cheaper laptop that could appeal to a

00:18:37   younger generation. Longtime colleagues described Ternus as someone willing to make clear calls in

00:18:43   contrast to Cook's more deliberative consensus oriented approach. Quote, Ternus will make decisions,

00:18:49   quote, when it comes to product developments. And one person who's worked closely with both

00:18:52   executives, quote, if you go into Tim with A or B, he won't pick. He'll ask a series of questions

00:18:56   instead if he has concerns. Ternus, on the other hand, will choose, said the person. It could be

00:19:01   right or wrong, but at least it's a decision. That shift could mark the end of an era in which major

00:19:06   product decisions were made collectively by a small group of top of top executives. Ternus is expected

00:19:11   to take a more centralized approach where he will be a singular decision maker. Earlier this month,

00:19:16   Ternus overhauled the hardware engineering organization around what he calls a new AI platform

00:19:20   designed to speed up product development and improve device quality. It's indicative of his plan

00:19:25   to deploy AI quickly throughout the company to improve its operations. The executive has told

00:19:29   staffers that he will remain closely involved with hardware engineering efforts aiming to shepherd

00:19:33   the next generation of technologies. So this characterization of Ternus as decisive in areas

00:19:38   where Cook wasn't sounds encouraging, but I do have to say that I have also seen reports of the exact

00:19:43   opposite, that he's not decisive and is afraid to make decisions. So it's really hard to tell what's

00:19:48   going on inside there. That's the problem I've always cited when talking about this is Ternus as he

00:19:53   existed in the company run by Tim Cook necessarily was subsumed under the way Tim Cook wanted to do

00:19:59   things. So it's very difficult to figure out how he in particular does things except for through like

00:20:04   these anonymous reports from people in Ternus as reporting chain who can can describe what he was

00:20:09   like. And that's, you know, that's presumably what this is. But what will he be like at CEO? I can tell

00:20:14   you, I would love him to be decisive, even if he makes mistakes. You know, again, he was supposedly a

00:20:19   component of the touch bar, which I don't think was a good idea. But decisively backing the touch bar

00:20:24   might have had a different result than throwing the touch bar out there and letting it languish until

00:20:29   you just give up on it and never make any changes to it. So like I'm, I'm all for decisive action,

00:20:33   especially if that decisive action can go in the opposite direction, which is we take decisive action,

00:20:38   we try it and then we decisively figure out that we screwed up and decisively stop, right? As opposed

00:20:43   to just kind of going, eh, I don't know, maybe try this, see how it goes, whatever. So I'm kind of

00:20:50   hearing what I want to hear in these stories. But until proven otherwise, I'm continued to be optimistic

00:20:55   about how things will go.

00:20:57   Yeah, this honestly sounds, you know, having Ternus come in here with a different style and

00:21:02   everything, we don't know for sure this is going to be a good thing. Time will tell. But I'm pretty

00:21:10   optimistic. I think this is what Apple needs. Obviously, any strategy can go wrong. But whenever

00:21:16   we've heard stories about Tim Cook's leadership style as CEO, there's a few themes that run through.

00:21:24   Number one is basically what Gurman says here of like, Cook does not like to be brought problems.

00:21:30   I've even heard people say like, don't bring me problems is one of his things that he has always

00:21:35   said. It seems like, you know, Cook does a really good job of seeming like a nice, calm,

00:21:42   chill guy, you know, in public or in interviews. But everything we've heard about Cook's leadership

00:21:48   style sounds a lot less nice and chill. It sounds like Cook is ruthless and leads by sheer force of

00:21:59   will in a what are you doing to ask, even asking me this question kind of way. That's the story.

00:22:04   The kind of theme of what we hear is that's more what he is, a ruthless bean counter and decision

00:22:11   avoider and very cold. Like all the descriptions of Cook in this leadership style. It's like a cold,

00:22:17   ruthless numbers guy. Nothing we've heard about Ternus sounds anything like that. And that's why I'm

00:22:26   optimistic. Like it's not that, you know, it's not that Cook is without benefit to the company.

00:22:32   Obviously, you know, he did very well in lots of ways. But like, I think the way to get great

00:22:37   products more consistently is more the kind of style we're hearing about Ternus, not the kind

00:22:45   of style we've heard about about Cook. Yeah. And like, I think one of the aspects that doesn't get

00:22:49   talked about much these days is the ability to inspire the organization behind whatever it is that

00:22:53   you think is your vision or whatever. And for all Tim Cook's, you know, good sides, I don't feel like

00:23:00   he was ever an inspiration to the people who reported to him or to parts of the organization in terms of a

00:23:05   vision of where we're going. Whereas I have again, I have no friggin idea how Ternus actually was within

00:23:11   the company under Tim Cook. But it seems like the hardware division did a good job in a way that I can

00:23:17   only imagine happening if Ternus was inspiring them to do their best work. Maybe that's not true. I don't know.

00:23:23   But like, when you have an organization, organization excelling, you always imagine that the leadership is

00:23:28   not just like making the right decisions and getting people to do things on the right track. But they're

00:23:32   also inspiring those people to do to, you know, to go the extra mile to do to be enthusiastic about

00:23:39   something cool that they're doing. And it's hard to imagine Apple's recent hardware run without there

00:23:44   being some kind of inspiration in that organization, because it's not just like, oh, we're afraid of the

00:23:48   boss and we'll just try to do a really good job. And it's not just like, oh, we're really smart,

00:23:51   because there's tons of really smart people at Apple. And there's tons of people want to do a

00:23:54   good job. But having leadership that inspires, and that's one thing Steve Jobs did, he had

00:23:58   enthusiasm for certain things. And if the thing he was enthusiastic about, he would inspire people do

00:24:04   it. He was obviously Steve Jobs was enthusiastic about what he considered to be good user interface

00:24:09   design. Very often, he had weird ideas. But it's so clear that the people who were also inspired by

00:24:14   were on the same way make the same like, yeah, I do want to make the best looking button that you want to

00:24:18   lick. Was that a good idea? But I can tell you, he sure as hell inspired that team to make Aqua

00:24:24   to his taste. And that team was enthusiastic to do it because they could see the vision like, yes,

00:24:28   we're going to do this. Again, you can quibble with the ideas, but it was so clearly not just like,

00:24:32   this is what we're going to do and everyone get in line.

00:24:35   How could you not help but be inspired by Steve Jobs' obvious enthusiasm for what would become the

00:24:42   Aqua interface? And you know, and so that's what I hope Ternus will provide that Cook didn't,

00:24:46   which is whatever Ternus is excited about, whether it's the touch bar or cheap laptops or whatever,

00:24:51   whatever things that he's excited about, I want him to inspire the Apple organization to get behind

00:24:57   his vision because they want them to see that he's super jazzed and into it. And like, what the hell

00:25:02   was Tim Cook super jazzed and into, you know, product wise, I guess, maybe the car and maybe also like

00:25:08   the iPad for a while, but it was so hard to tell.

00:25:10   I don't know if Cook was super jazzed about anything. Like, that's just not his style.

00:25:13   He's super jazzed about earnings calls, I guess.

00:25:15   Even, yeah, I mean, he felt like, but that's, I think also, you know, kind of another aspect of

00:25:20   these styles is, you know, one of the things we've always heard about Cook is, again, like,

00:25:25   they don't bring me problems. What that meant is that it kind of pushed problems down in the

00:25:31   organization. The idea, I believe, that I think the goal of a, of a, of a site of a, you know,

00:25:36   philosophy like that is to make the people under you figure things out and deal with things

00:25:43   themselves and deal with conflict, resolve conflicts themselves. Don't bring it to you.

00:25:46   That's kind of the, the idea. But one of the downsides of that approach is that many problems

00:25:54   in the organization can result, or many problems in the products can result from like weird personality

00:26:00   problems, like in like some mid-level of, of leadership. And so if like some VP doesn't

00:26:06   get along with some other VP or there's friction between two different divisions or whatever,

00:26:10   that can like hold back a product or that can cause certain dysfunctions or that can, you know,

00:26:15   create the wrong incentives. Ideally, the leader through both inspiration, as John was just saying,

00:26:22   and I think through being a little more involved in, in, in product development in a company like

00:26:27   Apple should ideally bring those, bring those conflicts better to resolutions that are directed

00:26:34   from the philosophy at the top, rather than Cook saying, figure it out yourself. Don't come to me with

00:26:40   your BS. That just pushes down. That's like, you know, that's like, you know, deal with, deal with

00:26:46   your problems by suppressing them. Like that doesn't really work. Um, and, and we've heard

00:26:52   over the years, we've heard so many stories of like, why is this thing this way? Oh, it's because

00:26:56   this one leader of this one group like refuses to do it any other way. And there's some conflict

00:27:01   between them and some other group, but like we've heard these stories enough of them that it's probably,

00:27:06   there's probably some, some truth to this general dynamic. So the turn of style, if it's more like

00:27:12   the job style, you know, jobs would get in there and like, you know, command things to happen and

00:27:18   they would happen. It's a harder process to scale. Like the jobs, the job style is harder to scale as

00:27:24   a company gets really big. Uh, but that does result in certain frictions being removed. Now it's also

00:27:31   bottleneck certain things. If, if more things are going through the CEO, you, you do have to, you know,

00:27:36   deal with bottlenecks here and there, but like, like the Apple under the jobs to our, you could kind of

00:27:41   tell what they were concentrating on and other things suffered during that time because of that

00:27:45   bottleneck. Right. Um, and so, you know, there are, you know, the, the approaches are not without their,

00:27:50   their downsides, but the, the jobs approach and what it sounds like, hopefully the turn of this

00:27:55   approach is closer to, I think that's better for Apple's product quality overall. Um, you know,

00:28:02   they, you can, we can argue about finances and operations and everything. And there's, you know,

00:28:07   there's ways to delegate those things. You know, you don't have to, the CEO is not the one going to

00:28:11   every factory or whatever, you know, you, you can delegate, but the general leadership style of

00:28:16   jobs and allegedly turn us sounds better for Apple and for Apple's products.

00:28:23   Yeah, I think you're right. And I'm really hopeful that once things get settled, we're going to see

00:28:29   kind of a new normal. That's maybe better, which would be really nice.

00:28:32   Also, can I, this is just pure speculation here. Do you think it's worth noting that we have a chief

00:28:39   hardware officer now, but Craig Federighi did not get elevated to chief software officer?

00:28:44   I mean, he wasn't, he wasn't the one who was threatening to leave and he's not as indispensable

00:28:49   as Surugi. So that's what that boils down to. I mean, look, I don't, I don't want to start any

00:28:53   crap or anything, but like that to me kind of stood out as like, maybe Federighi is not going to be with

00:28:59   us that much longer in this role. I mean, that's up to him, right? Like that's another thing that I

00:29:04   haven't seen mentioned as much. It got mentioned a couple of times in the CEO transition, but like,

00:29:07   and we've been talking about it for months leading up to this, uh, like, Oh, you know, when, when

00:29:11   there's a CEO transition, obviously there's a lot of turnover. Cause like now's the time to re re-examine

00:29:15   what you're doing and maybe you're old as well, blah, blah, blah. But that, that turnover does not

00:29:20   necessarily end when the CEO transition is announced. So don't be surprised if you see more executive

00:29:25   turnover surrounding the CEO transition, not just only before it. So I agree with you. Like,

00:29:31   I don't think Federighi is going anywhere, but like, don't be shocked if you see some more people

00:29:34   leave and it's like, well, why didn't they leave before the transition? Surely they knew it was

00:29:37   coming. It's like, sometimes it happens before, sometimes it happens after. But I mean, just the

00:29:41   Federighi situation, him not becoming chief software officer or whatever, um, from an outside perspective,

00:29:47   because I have no idea what's going on in there. Apple's hardware is amazing. Apple's chips are

00:29:52   amazing. Apple software is less amazing. I mean, it's as simple as that. The hardware guy becomes

00:29:57   CEO, the chip guy becomes chief hardware officer and the software guy stay where you are. You got

00:30:01   some work to do. Well, and I think like, you know, if we, maybe, you know, whenever Federighi does

00:30:07   actually retire or leave, I'm sure we'll do a whole thing about like what we feel about his era. But I do

00:30:11   think the, the Federighi era has been defined by some things being pretty good and some things being a

00:30:19   rough. And also I have to say, I mean, this is one of the things that I would say in the defense

00:30:23   of Federighi is at various times, including currently, he hasn't actually been in charge of

00:30:30   all software to the detriment of like, I mean, for example, Siri has been passed around like a hot

00:30:34   potato, not spending too much time under Federighi. How can he be expected to pass around more like

00:30:41   nuclear waste? Right. But like, and I believe Siri was under Federighi for some period of time,

00:30:45   but for a thing that for like long suffering weaknesses in Apple software platform,

00:30:50   Federighi, if you had given that to Federighi 10 years ago, and it had always been his, I think it

00:30:54   would have done better than if it had been passed around the org and you would hire the Google guy

00:30:57   from the outside and let him run it and then given it to Rockwell. And like, it's been all over the

00:31:01   place. And like, I just feel like it's unfair to say, well, Federighi, you're, you're the big,

00:31:06   you're the highest level software guy. So anything involving software with Apple is your fault.

00:31:10   Some things are, but not everything. And in particular, not Siri, which has been one of

00:31:14   the most glaring weaknesses that has only become more glaring with the advent of LLM. So I give

00:31:19   Federighi some cover for like, I mean, Suruji has been in charge of Apple Silicon, right? There's no,

00:31:24   it hasn't been passed around. It hasn't been divided up between him and some outside hires and stuff.

00:31:27   It's just him. So it lizardized based on him, but Federighi has been gaining responsibility,

00:31:32   but at various times has not been given responsibility over all the things. So I don't know. Yeah,

00:31:38   we will talk about that eventually. I don't say like, like the, the Federighi era,

00:31:42   like some of its, I think some of its shortcomings are around like, like product or design decisions

00:31:50   about the software that he probably was not the one making those decisions, or at least not exclusively.

00:31:55   Yeah. Like he's, he's involved in that process, but he is not the head of design.

00:31:58   If you look at like the, the challenges of the Federighi software, even let's leave out Siri. Cause

00:32:03   yeah, we know that's been nuclear waste, but like the, I think the main challenges of the,

00:32:07   of the Federighi era have, have been software design, which isn't really him software quality,

00:32:12   which is him. And you know, that's we, that we can talk about. Um, and I would say the features

00:32:20   that are developed with Federighi, there's like, there's immense engineering sophistication

00:32:27   and an immense care for privacy. But then you have the user facing features are like really kind of

00:32:37   like finicky and complicated configuration stuff, like focus modes and like the, like the home screen

00:32:44   configuration stuff. And like, there's, there's a lot of stuff there that's like, I see where this

00:32:50   came from, but it is delivered as a very kind of complicated story that I think could have used

00:32:58   better product direction. So like, and again, I don't know the way things work between software

00:33:02   design and marketing and like how the, how the, how the product decisions are made. I don't know how

00:33:07   much of that was Federighi's call, but the software quality is under him for sure. And, and I do think

00:33:13   like regardless of whatever, whatever, um, other inputs there are on design and product decisions,

00:33:21   software quality itself is a hundred percent on Federighi's leadership and his organization.

00:33:28   Like that, that is definitely him. And I do think that story with software has been spotty,

00:33:36   you know, and, and, and I, and I like Federighi like as, as a leader, as a presenter, as a character.

00:33:41   And, and I like, I like that he, you know, I've, I've had the chance to meet him a few times

00:33:46   and he is very clearly a very good engineer. The technical sophistication of the things that he

00:33:54   for, you know, both understands and, and has implemented in the past and has led the implementation

00:34:00   of more recently. Like he really knows his stuff at a technical engineering level and everything they

00:34:07   say about like privacy and everything. He embodies that he values that he enforces that. So I give him

00:34:15   a hundred percent credit for being a real nerds, nerd, a really good engineer and having, having the right

00:34:21   principles about things like quality, uh, or, and, and privacy. But, but the quality part is had the

00:34:27   execution on the quality part has been spotty. And, but I do wonder like with this new elevation

00:34:34   of Ternus, I, we don't know how Ternus and Federighi get along. Probably I'm sure that's fine. But like

00:34:38   with that elevation, Federighi, I don't think Federighi wanted to be CEO. So I don't think there's

00:34:44   that kind of weird friction, but like that's gotta maybe like put an end date in his mind of when he

00:34:51   wants to leave. Um, and I, I, it would not surprise me if the Federighi era ends sooner than, than people

00:34:59   think. I, I would be, I would be shocked if he's still there in say certainly five years, maybe even

00:35:05   three years. Um, and then finally my, my one really crazy wild card, which I know is, this is ridiculous,

00:35:11   but what if they brought Scott Forstall back?

00:35:15   They won't. I mean, this is not a serious, I don't, I don't think he wants to come back.

00:35:21   Honestly, I don't, I don't want him to come back, but, but you know, like that's the type of thing

00:35:28   that a new CEO can make happen. Cause honestly, as much as Tim Cook was responsible for, uh, for

00:35:32   Forstall being booted out or leaving or whatever, Tim Cook wouldn't stop Ternus from bringing Forstall

00:35:38   back. If you wanted, I, it doesn't seem like a thing that is going to happen. And I think I don't

00:35:42   agree with a lot of things that Forstall did, but to give an example, Forstall definitely inspired the

00:35:46   people under him. And the same way I feel like Federighi and his, Federighi's obvious enthusiasm

00:35:50   for the technical aspects of his work inspires the people who work under him, who are also enthusiastic

00:35:55   about those things. So there's a positive effect of the positive effect of enthusiasm is what I'm

00:36:01   endorsing in this episode that like, if you're excited about a job related thing, you can't be excited

00:36:05   about every job related thing, but Federighi, when we see him speak, he's clearly animated

00:36:09   enthusiastic about specific aspects of his work. And it's infectious, right? If you are also on that

00:36:16   same wavelength, like, yes, my boss's boss's boss is also super into this thing that I also think is

00:36:21   cool. And we're going to do some awesome stuff together. And they do. And then there's the parts

00:36:24   that he seems less enthusiastic about or has bad taste with, right? So, you know, nobody's perfect,

00:36:28   but yeah. Forstall coming back. Ternus could do that if he felt like it was the right thing to do.

00:36:33   And Forstall was willing to do it, which it seems like Forstall's happy out there doing whatever he

00:36:37   wants. But, you know, stranger things have happened. Yeah. To be clear, like, I don't, I don't think

00:36:41   that that's a realistic outcome for lots of reasons, but it is, it is an interesting thing I think to

00:36:47   think about as fans, just like a wild fantasy to think about. Especially if Federighi leaves and

00:36:52   there's a gap and it's like, who can we get to fill this gap? It's like, well, we have one guy who kind

00:36:55   of has already been in that seat. And if he suddenly has a change of heart and is like,

00:37:00   you know what? Yeah. You know, there's a gap, there's a power vacuum and you're going to bring

00:37:04   me back. Hey, it worked for jobs.

00:37:07   We are sponsored this week by Aura Frames. Aura Frames are electronic picture frames. You know

00:37:13   these and you probably hate them, to be honest, because for a long time they were all trash. They

00:37:19   were terrible. But I'm happy to tell you that in the intervening years between that time you bought

00:37:24   your terrible photo frame for your mother and now they've actually gotten really good and Aura Frames

00:37:29   have in particular. So what makes Aura Frames so great? Well, not only do you have a really well done

00:37:34   iOS app that you can use to search through your photo library and upload to one or more frames.

00:37:38   But beyond that, you can even use, say, the share sheet. So if you're in the photos app,

00:37:42   you can hit the share sheet and then share the photo in question to the Aura app. And then it will go

00:37:49   ahead and beam it to whatever photo frame or frames you so desire. Additionally, one of the other great

00:37:54   things about Aura Frames is that you get free unlimited storage. You can add as many photos and videos as you

00:37:59   want. Furthermore, if you are getting one as a gift, which honestly, you might want one for

00:38:04   yourself. We have one and they're great. We've gifted them. And that also is usually a absolute

00:38:08   home run. But let's say you're getting a gift. You can personalize it in advance. And so you can

00:38:13   add a message that will appear when it's powered on for the very first time. You can do this without

00:38:18   taking it out of the box. You can even get the Wi-Fi credentials of the person's house that it will

00:38:23   live in programmed into the Aura Frame without removing it from the box. It's really good stuff.

00:38:27   So you should really, really check it out. Aura Frames were named number one by Wirecutter.

00:38:32   You can save on the gifts moms love by visiting AuraFrames.com. For a limited time, listeners can

00:38:38   get $25 off their best-selling Carver Matte frame with code ATP. That's A-U-R-A-F-R-A-M-E-S.com,

00:38:44   promo code ATP. Support the show by mentioning us at checkout. Terms and conditions apply.

00:38:49   Thank you so much to Aura Frames for sponsoring the show.

00:38:52   All right. I think about an hour ago we were in follow-up. So let's continue. The ability to be

00:39:02   hotter. What was the context for this, please, John?

00:39:05   This is Apple making computers that can cool CPUs that have more transistors and that dissipate more

00:39:12   energy in exchange for doing more computation. They're, you know, as of now, their biggest computer is the

00:39:18   Mac Studio. It can only dissipate, you know, under 300 watts of power, or it is only dissipating under

00:39:24   300 watts of power with the current most power-hungry chip. What if you could dissipate twice that heat?

00:39:28   Could you put a chip that had twice as many transistors and was maybe 1.5 times as fast? I think you could,

00:39:33   but Apple hasn't done it. So on that topic of the ability for Macs to be hotter in exchange for doing

00:39:38   more computation.

00:39:40   Yeah. So, uh, Rachel Neuer in Scientific American writes, as reported in science, a metallic material

00:39:45   called theta phase tantalum nitrate, that too, tantalum nitrate, uh, achieved a thermal conductivity

00:39:53   of about 1110 watts per meter Kelvin, about three times higher than copper's 400 watts per meter Kelvin.

00:40:01   And it works in a way scientists have never seen before. Quote, our result breaks the historic

00:40:06   ceiling for heat transport metallic materials, says senior author Yongji Hu, uh, physicist and

00:40:11   engineer at the University of California, Los Angeles. Given this conductor's superior performance,

00:40:16   it has the potential to complement or even replace copper.

00:40:19   So Apple, even if you want to keep the Mac studio the same size, boy, do I have a way you can get

00:40:24   more heat out of it? Maybe three times as much heat. Uh, I don't know how much this, this alloy costs

00:40:28   or how difficult it is to manufacture, but, and it's probably actually much more relevant for things

00:40:32   like phones where a tiny amount of this exotic material could dissipate heat even better, but

00:40:37   I'm excited about material advances. So Apple just make the bigger Mac case to dissipate more heat,

00:40:42   but on the phones, put one of these suckers in there. Cause it sounds cool.

00:40:45   Yeah. For it literally sounds cool, John. Yeah. Wow. All right. Let's do some topics. And we've had,

00:40:51   this is one of those topics that we've had on the docket for a couple of weeks, but we had just

00:40:55   haven't had a chance to talk about it. Um, we're going to talk about Backblaze and how they're no

00:41:01   longer backing up everything that we thought they were. Uh, before I go any further, I wanted to

00:41:05   remind everyone, uh, Backblaze is a former sponsor. They are not currently a future sponsor. In fact,

00:41:10   they last sponsored us, I believe the summer of 2023, as far as my records indicated, but

00:41:15   I just want to make that plain for everyone. Uh, with that in mind, uh, Robert Anton, Anton

00:41:20   Reese wrote, uh, sometimes semi-recently, despite claiming to back up all of your data, Backblaze

00:41:26   quietly stopped backing up one drive and Dropbox folders, along with potentially many other things.

00:41:30   And later in the same piece, which we'll link in the show notes, of course, uh, he said the

00:41:35   following, which I thought was very, uh, at first I thought it was a bit dramatic. And then the more I

00:41:39   thought about it, I was like, no, actually I think he's onto something here. Anyway, Robert Anton,

00:41:43   Anton Reese wrote, by deciding not to back up everything, Backblaze has made it as if they

00:41:48   are backing up nothing. His point here is that if it always used to be that you could just assume

00:41:56   that Backblaze was backing up everything. And over the years, they seem to be chipping away at that

00:42:02   in bits and pieces and fits and spurts. And so now there's some amount of question. Are they backing

00:42:10   up X? Are they backing up Y? Are they backing up Z? And at that point it's like, well, what am I even

00:42:15   doing here in the first place? Now, to be fair, I am a Backblaze person. I still like it. I still use it.

00:42:21   I am unaware of a better solution for my needs. We're going to talk about other options toward the end of

00:42:27   this topic. But that's where I stand. And maybe once you guys start chiming in, you can tell me

00:42:32   your situation. But one way or another, John, it looks like you did some spelunking and discovered

00:42:37   in Mac version 9.2.2.878, which was released roundabouts of the end of August 2025. There's the

00:42:46   following in the release notes. Recent macOS versions can mount cloud storage, for example,

00:42:51   Google Drive, OneDrive, and Dropbox, in local paths, causing the backup client to mistakenly back them up.

00:42:57   These cloud-mounted folders are now detected and excluded to avoid performance issues,

00:43:01   excessive data usage, and restore complications. Meanwhile, in version 9.2.2.877, and this is for

00:43:13   Windows now, the backup client now excludes popular cloud storage providers from backup, including both

00:43:18   mount points and cache directories. This prevents performance issues, excessive data usage, and

00:43:22   unintended uploads from services like OneDrive, Google Drive, Dropbox, Box, iDrive, and others. This change

00:43:29   aligns with Backblaze's policy to backup only local and directly connected storage.

00:43:33   People were not happy about this, and people got very unhappy about this in a lot of places,

00:43:38   including on Backblaze's Reddit, or subreddit, excuse me.

00:43:41   And by the way, just to reiterate the date for those versions that I just pulled out here,

00:43:45   and I was guessing based on like archive.org pages and some random web page I found that listed the

00:43:50   date as being around end of August, early September 2025, because Backblaze does not put dates in their own

00:43:55   release notes on their website. That's why I had to do the spelunking to figure it out.

00:43:59   Anyway, that's summer 2025. When this when these releases came out that started excluding these

00:44:05   things. So it's taken people to Casey's earlier point, it's taken people like almost a year to go,

00:44:10   hey, wait a second, you mean it's not backing up my stuff in my Dropbox? And that's one of the worst

00:44:15   things about this is like, you just assume it's doing a thing and backing up your stuff. And if you

00:44:21   don't have occasion to check, you know, maybe you won't know that suddenly a thing that was being

00:44:25   previously backed up is not being backed up. So yes, people have discovered it almost a year later

00:44:29   and are making lots of noise about it, which is why it's in the news now.

00:44:31   Right. So on their subreddit, Natasha, who is an employee at Backblaze, wrote the following,

00:44:36   this decision was driven by a consistent set of technical issues we were seeing at scale,

00:44:41   most of them driven by updates created by third party sync tools, including unreliable backups and

00:44:45   incomplete restores when backing up files managed by third party sync providers. These cloud storage

00:44:50   providers now rely heavily on OS level frameworks to manage sync state on Windows, for example, files are

00:44:55   often represented as reparse points, which is an NTFS thing in the via the cloud files API. When they can

00:45:02   appear local, they are still system managed placeholders, which makes it difficult to reliably back them up

00:45:06   as standard on disk files. Additionally, Jim and other Backblaze employee chimed in once a reparse

00:45:13   point is added Backblaze is unable to back up the file, even if it physically resides on the computer.

00:45:19   Not great, Bob. Continuing from Natasha, we built our product in a way to not back up reparse points

00:45:25   for two reasons. One, we wanted the backup line to be light on the system and only backup needed user

00:45:30   generated files. Two, we wanted the service to be unlimited. So following reparse points would lead

00:45:35   us to backing up tons of data in the cloud. We've made targeted investments where we can, for example,

00:45:41   adding support for iCloud Drive by working within Apple's model and supporting Google Drive. But

00:45:45   extending that same level of support to third party providers like Dropbox or OneDrive is more

00:45:49   complex and not included in the current version. We are currently exploring an add-on that either

00:45:54   follows reparse points or backs up the tag data in another way. We also hear you clearly about the

00:46:00   communication gap. Both the sync providers and Backblaze should have been more proactive in

00:46:04   notifying customers about a change with this level of impact. Please feel free to reach out to me

00:46:08   directly if you have any questions. So here's the thing. I've read into this some, and I think two things

00:46:14   are simultaneously true. It seems to me that by way of the OS level affordances for this stuff becoming

00:46:23   more of a thing and arguably somewhat more opaque for anything that isn't like Dropbox itself or,

00:46:30   you know, whatever your file provider is itself, it does legitimately make it harder for Backblaze to

00:46:35   the point that it seemed so unreliable, if you are to believe Backblaze, so unreliable that it's almost worse

00:46:42   for them to try to back this stuff up because then some parts will get there, some files won't get there,

00:46:47   some files will be in a weird state, and it's just inconsistent. And that's arguably worse. It's better off to

00:46:52   just say, screw it, we're not going to back any of this stuff up. I genuinely have no reason to believe,

00:46:57   maybe John will correct me here in a minute, but I have no reason to believe that's untrue.

00:47:00   However, to just kind of quietly slide this into release notes and not say anything about it,

00:47:05   and also to say that it's on the sync providers, let me reread what Natasha said. We also hear you

00:47:12   clearly on the communication gap. Both the sync providers and Backblaze should have been more

00:47:15   proactive in notifying customers. I don't think this is on Dropbox or whatever or whoever else to

00:47:20   notify Backblaze customers that this is the case. This is a Backblaze problem.

00:47:23   Well, I mean, I see what they're saying. You remember when Dropbox rolled out their file

00:47:29   provider version, the one that works with Apple's framework for doing cloud files as opposed to

00:47:33   using the custom system that Dropbox did? When they rolled that out, Dropbox, first of all,

00:47:39   I think it was optional or opt-in, and within the little Dropbox app on macOS, there was a thing

00:47:46   you could click, and it would be like, what is this? What are you talking about, file provider? Which would

00:47:49   lead you to a web page on dropbox.com, which would say, hey, here's the deal with the file

00:47:53   provider version. We're doing this because Apple's making us, they basically said, which is true,

00:47:56   because Apple's trying to make them use their framework. And if you go to this version,

00:48:01   here are the limitations. And those limitations were like, other programs that were previously

00:48:07   accessing files in your Dropbox might have problems because of the way file provider works. And so I

00:48:12   feel like Dropbox actually did do a better job of saying, if you choose to go to the file provider

00:48:16   version of Dropbox, it won't work the same as the old one in ways that affect other apps.

00:48:22   Like we are changing the way our stuff in Dropbox looks to the rest of the system, and that will have

00:48:28   an effect on those other apps. And I think that's what Backblaze is getting at in terms of like,

00:48:31   you know, they would like, because like they didn't, Backblaze didn't decide to make Dropbox files

00:48:36   harder to back up. Dropbox decided to do that. And Dropbox, I feel like, did communicate that to the best of

00:48:42   their ability much better than Backblaze communicated. That's for sure. Because

00:48:45   I know that when this happened to Dropbox and Dropbox made it up. And in fact,

00:48:49   I still have not opted into file provider on Dropbox. I'm still right now running the non-file

00:48:54   provider version of Dropbox because it's better because it just, they're just regular files and

00:49:00   they are backed up by Backblaze and all that. Backblaze is probably excluding it by path or

00:49:04   whatever, but like they could in theory be, be excluded, be backed up by Backblaze. So they're just

00:49:08   plain files. It's not using file provider at all. Um, so that's the other

00:49:12   half of that, but I get what you're saying. And it's like, well, in the end, you're the backup

00:49:15   provider, regardless of the reason it is that you can no longer back up these files that's on you to

00:49:20   communicate. Like, so Dropbox says, Hey, we might affect other apps because we're changing the way

00:49:24   we're doing things and Apple's making us. And then the other people who are affected have to say, Hey,

00:49:29   Dropbox just released a new version. If you're using the file provider version,

00:49:32   our backup app can no longer back up your files. So seek an alternative method. And they just didn't say

00:49:37   anything. Like they put it in their release notes, you know, in a door in a basement behind a sign that

00:49:42   says beware of tiger. And then, and they did it a year ago in the summer and then didn't say anything.

00:49:47   And by the way, I would like to return to the, what they said in their release notes. Um, it says

00:49:53   like, uh, recent versions of Mac OS can mount cloud storage was the weird way of describing it in local

00:49:58   past, causing the backup client to mistakenly back them up. These cloud mounted folders are now detected

00:50:03   and excluded to avoid performance issues, excessive data usage, and restore complications. Restore

00:50:09   complications is the only one that is later referenced. No one references performance issues

00:50:13   or excessive data, excessive data uses. Like that's what we're paying you for to take our excessive

00:50:17   data. Like the whole using backblaze is excessive data usage. Exactly. It's like, Oh, if we backed up

00:50:23   those files, we would have more files in the backup. Yeah, that's the whole thing. We watch. So,

00:50:28   so their release notes characterize it as like, Oh, this is a bug. We shouldn't have ever been backing up

00:50:33   those files anyway. And we're doing it because there are performance issues and blah, blah, blah.

00:50:37   restore complications is the only legit one because it's like, if we can't back them up and you go to

00:50:41   restore them and we have like half of the files, then you're going to be mad about that. So we're

00:50:45   just excluding the whole thing. So they're, they're release notes, which are, you know, release notes

00:50:49   are not the official communication mechanism for a nuanced discussion of this thing. It's like a

00:50:53   paragraph, but the release notes compare the release notes to what Natasha says on Reddit.

00:50:58   Totally different reasoning. Now, all of a sudden it's like, Oh, well, here's why we're doing it.

00:51:02   And we can't back them up and restore things and blah, blah, blah. It's, it's like they weren't

00:51:07   speaking with one voice in this and they were definitely not trying to advertise the fact that

00:51:11   they were doing it, which is like, that's the communication gap. And that is the, the biggest,

00:51:15   easiest slam dunk against backblades is like you failed to communicate this to your customers.

00:51:20   And it's an important thing that your customers needed to know. And so now they're doing damage

00:51:24   control. Um, but for the technical aspects of it, having, having dealt with cloud files in hyperspace,

00:51:29   I have some experience with this. Um, I get where they're coming from because like in hyperspace,

00:51:36   I've got it easy. And even though I had it easy recall that version 1.0 wouldn't touch cloud files

00:51:41   at all because I knew there were going to be complications. And I'm like, it's a 1.0,

00:51:44   let's not overcomplicated. But when I did tackle cloud files in whatever version that was 1.3 or

00:51:49   whatever, I have it easy because given the APIs in Mac OS, I can tell when the file is cloud synced.

00:51:56   And I can tell when like, all right, you know, there's APIs where I can say, Hey, is this file

00:52:03   cloud synced or not? And the OS will tell me yes or no. And most of the time it's right. Although

00:52:06   sometimes it's not. Um, and then I can also say, all right, so if it is cloud synced, is the file

00:52:12   actually on this disc or is this just like, you know, a stub for the file? And if, and if I were

00:52:17   to read it, I would get the file contents, but it's not actually here. And if it tells me, Oh,

00:52:21   that's just like a stub that the file isn't there. I don't have to touch it because my app saves disc

00:52:25   space. And if the file's not there, I'm done. Hey, it's not even there. Right. Because like,

00:52:32   what the hell am I, you know, I'm, I'm not going to force the file to be downloaded so I can see if I

00:52:36   can see if I can dedupe it. It's, it saves them the most space by not being on the disc. So that's,

00:52:42   that's, I have it easy. Uh, and even doing that, there are lots of nuances and blah, blah, blah.

00:52:46   Right. But a backup client can't do that. A backup client can ask those questions. And if the answer

00:52:51   is, yeah, that file's not even on the disc, the backup client now has a difficult choice. It's like,

00:52:55   well, I could skip it and be like, well, how am I going to back up a file that's not on the disc

00:53:01   or I could cause it to be downloaded, which takes disc space. And also, by the way, I can tell you

00:53:08   that asking Mac OS's file provider thing to cause a non-downloaded file to be downloaded. You might

00:53:14   as well put a letter in the mail. Like it's, it's like on iOS where you make a suggestion about a

00:53:20   thing that might happen in the future. And then you go away for a month and come back and see if it's

00:53:23   happened. It's not as if like, you can sit there with your arms crossed and being like, so is the file

00:53:28   going to be downloaded? Is that going to, is it going to happen? It's like, well, there's a whole

00:53:32   other demon that does that. And it's busy right now. And it's a low priority thread and you're like

00:53:35   tapping your fingers. And it's like, okay. And how many of those can you wait on at the same time?

00:53:39   And then how long do your backups take? And then what if you just, what if hours pass and that one

00:53:43   file doesn't download? Is there a big button you can push to make that one file downloaded? Hell no.

00:53:48   Otherwise there'd be that button in photos too, where you could say, just download the photos now.

00:53:53   Think now, now you don't do low priority threads, do everything. There's no do it now. Like you're at

00:53:58   the mercy of the system. And so if you're a backup app, you're like, I can't tell you that I will

00:54:03   successfully ever reliably get all these files to the point where they download. And by the way,

00:54:08   what you'd want to do is get them to download, back them with the backblades, but then evict them

00:54:12   again, because you don't want to take up the disk base. The whole point is like the cloud storage.

00:54:16   It's like, oh, they're not even all on your disk. Maybe they don't even all fit on your disk.

00:54:19   So you can't ask them all to be downloaded because you'll fill the disk. So you got to ask a small

00:54:23   number to be downloaded, then wait for them to be downloaded, then upload them, then re-evict them.

00:54:27   Oh, by the way, what if they were editing the file during that hour long process?

00:54:30   It, I understand why a backup client is going to be like, we just can't,

00:54:35   like, even if we tried to do this, we will not be able to give you a coherent, internally consistent

00:54:42   backup of this because they can't even use things like, like APFS snapshots, right? Because

00:54:47   you snapshot a directory filled with a bunch of stub files. That doesn't, like, you have a point

00:54:52   in time image of a directory with half the files that aren't there, right? If the files were all

00:54:57   there, you could snapshot it and then read the snapshot and backup like time machine. But if

00:55:00   they're stub files that are in cloud storage, what the hell are you going to do? You can't

00:55:03   cause them to be downloaded into the snapshot. So I am totally on board with Backblaze saying,

00:55:09   we can't actually back up these files in a reliable fashion.

00:55:12   I know everyone in all the threads is like, but what if they're all downloaded? Backblaze can't

00:55:17   know that they're all downloaded. Do you want it to iterate over the whole thing, confirm that they're

00:55:21   all downloaded, then start backing them up? Guess what? In between the time that it confirmed they're

00:55:24   all downloaded and started backing them up, some of them got evicted. Now what do you do? Like,

00:55:27   there's no, there's no 100% solution to this. And it's just the nature of the, like, the pros of

00:55:33   the system is, oh, so you can have a Dropbox that's, that's four terabytes, and you can mount that on

00:55:39   your one terabyte Mac. How does that work? Well, half the files aren't there. They're just stubs.

00:55:42   And it's all transparent to you. And if you just double click them, they all work as long as you're

00:55:45   internet connected and blah, blah, blah. Like, that's the benefit. The downside is you can't

00:55:50   back it up in a reasonable manner. It's just not possible. So I think this is a case where

00:55:56   Backblaze did a really poor job communicating a real technical issue. And they continued, by the way,

00:56:03   I think, to do a really poor job of communicating both the technical issue and what they're going to do

00:56:06   about it. But I do think they're not just like, oh, they're just being lazy and trying

00:56:10   to save money. Like, I don't actually think there's an easy way for them to do this successfully. And as

00:56:16   they noted, like, we're looking into this, that and the other thing. And we've worked with Apple

00:56:19   directly to do iCloud Drive backups. Like, how are they doing iCloud Drive backups? Well,

00:56:23   if you can work directly with Apple, they can figure something out. Like, I'm sure this has taken a long

00:56:27   time where, you know, they're talking to Apple, Apple does actually control the, you know,

00:56:32   whatever it is, the, I forget what it's called, Cloud D or the ubiquitous storage back. Like, there's demons

00:56:36   on macOS that do this. Apple writes those, right? And so

00:56:40   they can work something out with Backblaze. Here's how you can actually back up iCloud files

00:56:44   in a somewhat safe and consistent manner. But I bet even there, there are situations

00:56:48   in which a particular Backblaze run will fail to back up a file because it

00:56:52   couldn't get the stupid cloud demon to download the thing locally. But

00:56:55   whatever, they're doing that. And the same thing with Google Drive,

00:56:58   but Dropbox is a little bit more exotic in the same one drive. And then on Windows,

00:57:01   I have no freaking idea what a reparse point is. And I don't even want to know.

00:57:04   But apparently there are additional complications that I do not understand there. So I can't,

00:57:09   I can't chime in on what that situation is. But

00:57:11   yeah, I'm kind of the same situation as Casey. I again, again, Backblaze has been a past sponsor.

00:57:16   I would gladly have them as a future sponsor, because I think Backblaze is the best cloud based

00:57:22   backup product for the Mac. It's just that in this world we live in, the number of limitations on that

00:57:27   product are just increasing with the use of the file provider and various cloud storage things.

00:57:32   There are other alternatives that we'll talk about in a little bit, but I still use and pay for

00:57:36   Backblaze. And I think it is the best software that does what it does. But I think they did a terrible

00:57:41   job communicating this limitation, communicating this limitation. And you know, like any backup

00:57:45   strategy, you got to have multiple ones, right? So even if Backblaze is totally ignoring my Dropbox,

00:57:50   I have other things that are like my super duper clone that are ignoring Dropbox because super duper

00:57:54   doesn't care about Dropbox. And because I don't use the file provider version of Dropbox,

00:57:59   all my Dropbox files are plain normal files on my disk and super duper backs them up fine. And by the

00:58:04   way, so does Time Machine. So, you know, backing up is difficult and the price of good backups is

00:58:11   eternal vigilance. And in this case, our vigilance was someone thwarted. Our vigilance was thwarted by

00:58:16   Backblaze doing a really bad job of communicating this change.

00:58:19   We'll put a couple other links in the show notes. There's a blog post from Backblaze that's kind of

00:58:24   we're victims too, which again, like I think there's truth to that, but I just don't love the way in which

00:58:30   it was communicated. There's also some other reports and roundup from MJ Psy as always.

00:58:35   And then there's some alternatives here that we can at least have you look at. I don't know if we

00:58:43   necessarily recommend them, but we can point them and point them or point you in their directions.

00:58:47   Yeah, this first one is not a piece of software, but it is a strategy that I saw mentioned many,

00:58:51   many times. And I do want to address it because it it can work, but only in certain circumstances.

00:58:57   So here's the suggestion. Number one, if you're a Dropbox user, don't use the file provider version

00:59:03   in Dropbox that there is an option for that buried in the app. You can even opt out of it once

00:59:07   you've opted back in. It's the same piece of software for a while. I thought it was like a different

00:59:10   version, but like right now it's like the same. It's just the Dropbox client for the Mac and the

00:59:14   same piece of software can run in two different modes. Apparently if you can find your way to the

00:59:18   option, you can disable it. I don't know what the limitations on that on. I maybe, maybe some people

00:59:22   were opted in permanently, but anyway, I have the option turned on and off. And I know this because

00:59:26   to test hyperspace with both the non-file provider, non-file provider version, I have the non-file

00:59:31   provider version on this Mac and on the other Mac, I have the file provider version so I can, you know,

00:59:35   see the different behaviors. And it's the same account and it's the same client software.

00:59:39   So the idea is don't use the file provider version. So then every file in your Dropbox is on your

00:59:45   disk taking up space. There's no weird file stubs. There's no ubiquitous files. They're just plain

00:59:50   boring files that are just a hundred percent there once they're synced. Okay. But Backblaze will probably

00:59:57   still skip them because they're in a folder called Dropbox and, you know, a well-known location.

01:00:01   They'll just say, Oh, well, I'm not going to check whether these are cloud backed. I'm just going

01:00:04   to say anything in a folder called Dropbox, anything under there where you won't back up or whatever

01:00:08   heuristics they use to figure out, is this a Dropbox folder? Maybe there's even a way they can just look

01:00:12   it up in Dropbox prefs where your Dropbox folder is. They might still be skipped. Right? So, and people

01:00:17   complain about that. Like, look, Backblaze, don't skip it based on the path. I'm telling you,

01:00:21   I'm not using the file provider version. Just back them up. It's pissing me off. So the solution,

01:00:25   the suggestion very much in the hyperspace style is how about you do an APFS clone of the entire

01:00:31   Dropbox tree, which will take up essentially zero extra space. Not really zero because it's a tiny

01:00:36   directory entry for every single file, right? So, but effectively zero space in the grand scheme of

01:00:41   things. Do a complete, complete clone of your Dropbox tree to a location that does not have Dropbox

01:00:46   in the path. And then Backblaze will back that up because Backblaze has no idea where those files came

01:00:50   from, but it also won't take up any additional space on your disk. The problem is doing a giant

01:00:55   recursive clone of your Dropbox tree. A, it's the painting the Golden Gate Bridge thing where when

01:01:00   you start that recursive clone, things could have changed by the time it finishes. So it's not really

01:01:03   a point in time snapshot, although you could do it. You could do it from a time machine snapshot if you

01:01:07   wanted. But B, that's time consuming. Like just creating that directory trees and all the directory

01:01:12   entries and entering all the metadata and all that stuff. That is time consuming. And if you have like

01:01:18   a terabyte of Dropbox stuff, just doing that operation on a periodic basis, like, and what tool you,

01:01:23   if you use our sync, our sync won't, won't copy all the metadata. If you use our clone, you might get

01:01:27   more data. If you use Apple's like ditto tool, you'll get even more. Like you're starting to become

01:01:31   like, you're starting to deal with like file system APIs and stuff that you don't want to deal with

01:01:36   because like, uh, faithfully reproducing every aspect of files is difficult. And setting aside

01:01:41   that Backblaze doesn't even back up all that metadata. Anyway, metadata on Mac OS is

01:01:46   terrifyingly complicated these days, as I've had to find out with, uh, with hyperspace. And so doing

01:01:52   something as simple as, Oh, I'm just going to make a copy of the directory. I'm sure, you know, it will

01:01:56   take up effectively zero space, but will the metadata and every one of those files be identical to the

01:02:01   source files? No, it absolutely won't. Uh, it'll differ in ways that probably don't matter, but it will

01:02:06   differ. And so there's now you have that in the mix. So it's another complication, but I just want to

01:02:10   throw that out there. If you do this, you have to have the non-file provider version and you have to find

01:02:14   some way to carefully and efficiently periodically make a full APFS clone. And by the way, don't just make a

01:02:19   plain copy. You got to make sure those copies are clones, which they probably will be, but depending

01:02:22   on the tool you use, they might not be. And anyway, back point, point Blackblaze is that, you know,

01:02:27   Blackblaze will back up that directory because it's not excluded because it has no idea this came from

01:02:30   Dropbox. So that is the zero cost. As long as your time has no value solution to this problem.

01:02:38   There are also some not free solutions to the problem. Uh, this is software. First is arc A R Q. Uh,

01:02:45   is it Marco that's used this? I feel like one of you sees this. Yeah, I use it a long time. Years ago,

01:02:49   I used it. Oh, do you have anything you'd like to add about that or just going to? No, I mean,

01:02:53   it's been a long time. So anything I say would be very outdated. This is, this is, I think just an

01:02:58   independent developer making a piece of software that backs up files. And the person obviously has a

01:03:04   passion for it and has developed this program over years and it is tries to be very careful and

01:03:08   reliable. Um, it is not big corporate bloatware, uh, but it is also very clearly an idiosyncratic

01:03:16   indie style app that does one very specific thing. And a lot of people like it and have success with it

01:03:20   and it doesn't have the same limitations. Now, how does it handle the, the difficulties that I just

01:03:26   described with Dropbox? I don't know. Uh, one thing I can imagine is that if you have the non-file

01:03:31   provider version that it will, it will just back them up because it's like, well, they're just

01:03:34   plain files. I don't, I don't, I'm not going to exclude them based on the directory name containing

01:03:37   Dropbox. It's not going to do any of that. So thumbs up for ARC in that respect. But if you use the file

01:03:42   provider version of OneDrive or, you know, the Dropbox or any of these other things that Backblaze is

01:03:47   skipping, I'm not sure how ARC will handle that. You could ask the ARC developer, but that is a third

01:03:51   party alternative to a piece of software, like an app that you run that will periodically back

01:03:56   stuff up that is entirely divorced from the Backblaze code base and technique.

01:04:00   Yeah. And to be clear, my limited, very limited understanding of this is that it's mostly a

01:04:07   bring your own backup destination kind of thing where you use ARC to manage sending data somewhere

01:04:13   else, but the somewhere else is also your problem. Uh, I also just realized as I'm loading

01:04:19   arcbackup.com again, A-R-Q, uh, there's some quotes on here, quote, I honestly love this app,

01:04:24   quote, Marco Arment, which is funny. Uh, but anyways, I did say that, uh, but the, the, I think

01:04:31   they do, it appears they do dabble in providing the, the destination as well. But the, I think the

01:04:37   bread and butter of ARC from my limited understanding is you have a place like a Synology or, you know,

01:04:42   is one of like Backblaze B2 or, or, um, the AWS equivalent. What is that that I'm thinking

01:04:47   of the drawing the blank? S3. S3. S3. Thank you. Something along those lines that you are using

01:04:52   ARC to manage backing up and restoring from there. Yeah. It's very funny. Like you look on their page,

01:04:56   like here's, here's the storage locations, Amazon drive, AWS, B2, Dropbox, file-based, Google cloud,

01:05:01   Google drive, one drive, poly cloud, SharePoint, SFTP, Wasabi, attach disc or network share S3

01:05:07   compatible server. Like it is backend, not backend agnostic, but it's pluggable backend. So if you just

01:05:13   want to copy things from one hard drive to another, ARC will do it. If you want to copy things from a

01:05:16   hard drive, S3, ARC will do it. If you want to copy things from a hard drive to, you know, Wasabi,

01:05:20   which is another like S3 type server, like it, that's, that's what this app does. And so people

01:05:25   like the flexibility of like, I just want to copy things from A to B and let me pick what A is and

01:05:30   let me pick what B is. And let me use an app that is, you know, uh, seemingly, uh, been developed over

01:05:36   many, many years by a development team or an individual that are really enthusiastic about this

01:05:41   problem space. And so there you have it. ARC is that comes up a lot on these discussions.

01:05:45   And then there's a, I don't remember if this is a newcomer, which is what I was about to say,

01:05:50   or if it's just been, um, purchased by another group, but the same folks that do kaleidoscope,

01:05:55   which is my, uh, diffing tool of choice, they now own or run or whatever a parachute backup,

01:06:01   which is again, just software. But what this is about is specifically with regard to iCloud. So iCloud

01:06:06   photos and iCloud drive, this is a way to slurp up the stuff that's in iCloud and put it somewhere

01:06:11   else as a backup. So if you're, if you're limited, if your use is limited to iCloud stuff, you might

01:06:17   want to check out parachute as well. We'll put a link in the show notes. Yeah. This is another one

01:06:20   of the problems. So I talk about this a lot with my photos backup thing, you know, um, Apple's photos

01:06:25   app on the Mac has options in its settings. It says, Hey, do you want me to download what they call

01:06:31   like download originals to this Mac, which is, uh, I'm going to get the original full resolution

01:06:35   version of all your images and I will download them on this Mac. Or do you want me to quote

01:06:39   optimize storage, which basically means some of these photos aren't going to be on your Mac.

01:06:43   We'll put the thumbnails on your Mac or whatever, so you can scroll through them. But like when you,

01:06:46   when you double click on one to see the big version, we might not have downloaded that.

01:06:50   So we'll download it. And that's to save disk space because if your photo library is terabyte

01:06:53   and you have a 500 gig laptop, how can you ever have your whole photo library in there? Well,

01:06:58   if you do optimize storage, it will make it look like your whole photo library is on there,

01:07:02   but obviously there's a demon behind the scenes that is downloading and evicting,

01:07:05   the full resolution photos periodically. So what if you're trying to back up your Mac? You're like,

01:07:10   Oh, I'm backing up my Mac. So I have my Mac has my whole photo drive library on it and I'm backing

01:07:14   up my Mac. Therefore I have a backup of my photo library. It's like, well, if you have optimized

01:07:17   storage on and your iCloud photo library and you make a backup, you're not backing up your whole photo

01:07:22   library. You're backing up whatever portion of your photo library is on your Mac at that moment.

01:07:26   Um, and so I always advocate having at least one Mac that you say download originals on,

01:07:31   right? But what if you don't have any Mac that has enough space to download all your originals

01:07:36   on because your photo library is huge or your Mac SSDs are small parachute. What it will do is

01:07:42   it will essentially go through your photo library and say, is this photo on the Mac? If it is,

01:07:48   they'll back it up. If it's not on the Mac, it will say, okay, photo thing, download it,

01:07:52   wait for it to download, back it up and then evict it again. That's my impression of what

01:07:57   parachute is doing for you. It will laboriously piecemeal download your whole photo library,

01:08:04   a few photos at a time. So it can back them up to wherever you're backing them up to. And then it

01:08:08   will chuck them back off again. And obviously if, if your whole photo library is on your Mac,

01:08:12   you could use arc to back it up or parachute backup will have less work to do. It was like,

01:08:15   Hey, look, all the files were there. I backed them all up. Um, but that's, it's trying to capture

01:08:21   a niche, which is like, well, lots of things in the Apple ecosystem won't even bother backing up

01:08:26   your, your photos library is like, Hey, it's in the cloud. And the cloud is the backup. And that's

01:08:31   something that we should mention here. Like people think, uh, all my files are in Google drive.

01:08:34   So they're backed up. All my files in Dropbox, so they're backed up. A backup is not the same as

01:08:39   having your files on a server somewhere in the cloud. Having your files on a server somewhere

01:08:43   in the cloud is great. If your laptop falls into the ocean, but if you accidentally delete all your

01:08:47   files or accidentally delete a file and don't notice until it's 30 days later and it's been removed from

01:08:51   the trash, that's where backups come in handy because backup is not the main location where your file is.

01:08:57   It's a second location that is not affected by changes to the origin. So if you have a backup of a file

01:09:05   and you delete that file, you can get it from the backup. If you have your file in Google drive and you delete

01:09:11   it from Google drive, you can't get it anymore because you deleted it. Again, setting aside, it'll be in the

01:09:16   trash of 30 days, yada, yada. The whole point is like, if it's really gone from where it was, it should still be in

01:09:23   your backup. But if something is really gone from your photo library, you don't have any backups, it's really gone.

01:09:28   So that's why apps like parachute help for people who, uh, you know, are paranoid about this type of

01:09:34   thing. I mean, I, I'm paranoid about photos to the point where I have my photos library upgrade uploading

01:09:40   to Google photos. So my originals are also in Google photos and Google photos. There's no connection to

01:09:47   Apple photos. Like if Apple burns down and my whole iPhone photo library dies and I lose all my time

01:09:52   machine backups and I lose all my M disc optical discs and you're like this, like my, my fifth level

01:09:56   backup. It's like in the end, Google photos also has all the originals from my photo library. And it

01:10:02   doesn't know they came from my photo library. It just knows it has a bunch of images in it. And so

01:10:06   yeah, the more, uh, more diverse your backup environment can be, the more protected you are.

01:10:12   And that's what apps like parachute and arc are. Therefore they're not made by Apple. They're not

01:10:16   made by backwards. They're not made by Google. They're made by another company. And so you're trying

01:10:20   to like Marco's point earlier about security bugs and stuff like that. Try to spread the risk

01:10:25   around across multiple companies, multiple things. Some of them are, uh, you know, in your house.

01:10:30   Some of them are offsite. Some of them are clouds. Some of them are big companies. Some of them are

01:10:33   small companies. I know it seems like it seems to me often that the main job of me and my computer

01:10:38   is to back up my computer. Like it seems like what does your computer do other than back itself up.

01:10:43   But if you want, if you have digital data that you care about to the degree that I care about my photos,

01:10:48   for example, it is a considerable amount of effort to make sure that that data is protected in some

01:10:55   way. Or you, otherwise you're just rolling the dice and hoping a disaster never strikes. And,

01:11:00   and even I am not fully protected. I am absolutely sure there are many failure modes in which I will

01:11:04   lose all my data, but I'm just trying to minimize them to the best of my ability. And diversity in

01:11:10   your backup strategies is a great way to do that.

01:11:12   We are sponsored this episode by Delete Me. Delete Me makes it easy, quick, and safe to

01:11:18   remove your personal data from hundreds of data brokers online at a time when surveillance and

01:11:22   data breaches are common enough to make everyone vulnerable. It is so scarily easy to find personal

01:11:28   info about people on the web. You just do a search for someone's name and you get all these sites

01:11:32   advertising, like find their home address, find their phone number, like all this crazy stuff. And

01:11:36   what that is, is data brokers. These are sites that collect as much info as possible about as many

01:11:41   people as possible and list it in search engines and try to get people to come by and give them a

01:11:45   few bucks for it. And they all have some kind of like opt out or deletion procedure that you can get

01:11:50   your data removed, but they're all different. And there's hundreds of these sites and there's new

01:11:54   ones cropping up all the time. So what Delete Me does, this is an online subscription service and they

01:12:00   automatically go out there, find data brokers out there. They find your info on data brokers and they

01:12:06   mass opt you out and file the takedowns and deletions for as many data brokers as they can

01:12:11   find. They're constantly adding more. They give you reports about what they were able to find,

01:12:15   where they found it, how they were able to delete it. It's a great service. There's lots of services

01:12:20   out there that do this. Delete Me was rated the number one pick for data removal services by the

01:12:24   New York Times Wirecutter. I even personally, like when back before they were a sponsor, I was looking at

01:12:29   these kinds of services and I found Delete Me and I used it myself and I like it a lot. So take control of your

01:12:35   data. Keep your private life private by signing up for Delete Me. Now at a special discount for our

01:12:40   listeners. Get 20% off your Delete Me plan when you go to joindeleteme.com slash ATP and use promo code

01:12:46   ATP at checkout. The only way to get 20% off is to go to joindeleteme.com slash ATP and enter code ATP

01:12:54   at checkout. That's joindeleteme.com slash ATP, code ATP. Thank you so much to Delete Me for sponsoring our show.

01:13:05   All right, let's do some Ask ATP. And Brian Ash writes,

01:13:10   Hey, John, what year or generation is your Honda? Any plans on when you'll replace it? If you don't

01:13:16   have plans, but if you had to replace it, what would you get? Brian, John does not believe in

01:13:20   hypotheticals. I am right there with you with this question, Brian Ash.

01:13:24   We did a whole member special on this hypothetical. How are you saying? I don't believe in it.

01:13:27   Uh-huh. Uh-huh. Well, you're going to poke a hole in it. I guarantee it. Anyway,

01:13:29   I looked at Honda's website, writes Brian Ash, and unless I missed it, it looks like they don't

01:13:34   offer anything with a manual anymore. Are you driving less now that you're independent and

01:13:37   or because of what your kids are doing now? All right. So let's take this, uh, one piece

01:13:41   of time here. What year generation is my Honda? It's 2014. So I hope it's, uh, over a decade old

01:13:46   now. It's hanging in there. Uh, any plans on when you replace it? My current plan is to replace

01:13:51   it when it dies. What does dies mean? Well, you know, it's a certain point. Cars get so old that

01:13:56   repairing them is no longer worth money. I will say that I'm probably willing to put, uh, a surprising

01:14:02   amount of money into this to keep it on the road. But at a certain point, you know, it becomes

01:14:06   untenable. The car gets sold or whatever. So I'll tell you what that point is. But right now I have

01:14:09   no plans to replace it because I like it and it fulfills my needs and it's paid for. Uh, so yeah,

01:14:16   it's hanging in there. It's low mileage. It's very low mileage. About how many miles are on it

01:14:20   just off the top of your head? 45, less than 45, I think 43. So 2014 car with 43,000 miles on it.

01:14:26   So this is a low mileage, low mileage example, as they would say. My, uh, my, what is it? A 2018.

01:14:32   I'm pretty sure I had that right. My 2018 Golf R, uh, has like 32, 33,000 miles on it. So similar

01:14:39   situation, you know, when you don't drive, if you don't have a commute and you're not driving long

01:14:43   distance to offset that commute on a regular basis, you know, for whatever reason, uh, you end up putting

01:14:48   not too many miles. And for, for probably, I don't know, almost a decade when it like my last job that

01:14:54   I had for like 12 or 13 years, I commuted to it, but the commute was not long. Like having a short

01:14:59   commute, like my commute was like three miles or so or something like that. I mean, it took forever

01:15:02   because traffic, but you know, it was distance wise. It wasn't long time wise. It was very often long.

01:15:07   And so even when I was commuting, I wasn't racking up the miles. I know some people commute, you know,

01:15:11   10, 15, 20, 50, a hundred miles to their job and that will really rack up the mileage. So that's another

01:15:17   reason this is low mileage. And then obviously, uh, quitting my job and working from home is just reduce

01:15:21   it further. Um, if I did have to replace it, what would I get it? Like with, like with big TVs, I'm always,

01:15:27   always looking for what would I get? That's, that's my main activity. Right. And in the car space,

01:15:32   um, like I, can you, did you find the number special that we did about that? No, I'm looking

01:15:37   as we speak. I haven't gotten there yet. Yeah. And the, in the car space, it's tricky because I want my next

01:15:41   car to be electric, but I don't like any of the electric cars that are available. Right. Like I, I like

01:15:46   them, but like, you know, when I ranked, I don't like any of them better than my current car. You know

01:15:49   what I mean? Like they all have something that annoys me about them. Makes me think I'll, I'll just wait

01:15:52   until they figure that out and fix it. And they just keep always messing up one, even my beloved

01:15:55   lucid air. Like I hate that interior. I don't like what they did with the controls. And I'm like, I'll, they'll fix it

01:16:01   with the next revision of the air. It's like, no, we're never revising that car. All right. So I,

01:16:06   I don't know. There are other cars that I like. If I, I think the member special hypothetical was more

01:16:11   like if you had to, like if your car broke, you know, you got totaled or whatever, and you had to

01:16:15   get a new car, what would you get this right now? The answer would be, I would get a stick shift Honda

01:16:19   civic without a sunroof on it. Um, because my wife has that car with the sunroof. Um, and I really like

01:16:25   her car, except the stupid sunroof hits my head. So she insisted on the one with the sunroof is I think

01:16:30   we talked about before, cause she wants the heated seats and all the blah, blah, blah. And you can't

01:16:33   get that car without a sunroof with the trim oval she wants. And it's her car and she's way shorter

01:16:36   than me. So it's fine for her. And she likes it. She likes her car. I like her car too. I just don't

01:16:41   like the sunroof. I would get a stick shift Honda civic without a sunroof. Uh, I like the current

01:16:46   generation. I like the current generation. It's a really nice car. I recommend it to anybody.

01:16:49   It's excellent. Uh, even the hybrid is excellent, which I wouldn't buy because it doesn't come

01:16:53   with a stick shift, but the hybrid is also excellent in the current model year, uh, current generation of

01:16:58   Honda civic. Um, and then are you driving less now that you're independent? Um, yes. Cause I don't

01:17:03   have to commute to work. Um, so even though it was only three miles, it was three miles there and

01:17:08   three miles back every single day. Whereas now, you know, run errands a couple of times a week,

01:17:12   but it's not every single day. Um, and because of what my kids are doing, it's probably more because

01:17:17   of what my kids are doing. Cause my kids are doing is going to schools that are far away and I'm

01:17:19   driving back and forth to either pick them up or drop them off or bring them stuff or whatever.

01:17:23   So my, my son was like 45 minutes to an hour away, depending on traffic. And my daughter is

01:17:27   like an hour and a half to two hours away. And I've gone back and forth to the, both of those

01:17:31   schools a lot. Uh, so what my kids are doing is actually driving up my mileage.

01:17:35   Fair enough. Uh, ATP neutral car shopping from the 14th of May, 2024. We will put a link in the show

01:17:43   notes. Jason writes, Mac hardware represents just under 8% of Apple's revenue. It will probably be

01:17:50   overtaken by AirPods soon. Woof. Uh, my question is this, which Apple CEO will remove the Mac from

01:17:56   their lineup? Will it be John turn? I don't think I want to keep reading this. Even I don't want to

01:18:00   keep reading reading. I think this is the answer to this one. Will it be John turnus? He could be CEO

01:18:04   until 2040. Do we think the Mac will exist in hardware form then? Will it be the next CEO after

01:18:10   turnus? If, if the best selling Mac of all time is basically an iPhone and a bigger chassis, why not

01:18:15   sell Mac OS as an app for iPhone users and save on manufacturing costs? Everybody wants everything to

01:18:21   be in their iPhone. My whole computer could be in my iPhone. Then I would just connect my phone to a

01:18:26   keyboard and a mouse and a big monitor and it would become a Mac. The files are in the computer. You

01:18:31   just described a Mac mini with a screen on it. Yeah, that's just a Mac. Yeah, exactly. So the easy answer

01:18:36   to this and the one that everybody says, because it is the easy answer is as long as the Mac is the

01:18:42   platform you use to develop for all of Apple's other platforms. You can't get rid of the Mac

01:18:46   because then there would be no apps for the phone. There would be no apps for the iPad.

01:18:50   There would be no apps for Apple TV. There would be no apps for Vision Pro. Well, anyway,

01:18:53   like it and that's app that's on Apple. Apple did make, you know, a way to develop iPad apps on the

01:19:00   iPad, but it's extremely limited. But like they never went whole hog and said, here's Xcode for the

01:19:05   iPad. Here's Xcode for Vision Pro. Here's Xcode for the phone for crying out a lot. Full fledged,

01:19:09   100% can do everything. Xcode like an Apple could have done that. There's nothing stopping them

01:19:15   from porting Xcode to the iPad and getting rid of the Mac. And, you know, like like like that's but

01:19:19   until they do that, they can they can't they literally can't get rid of the Mac. Now, setting

01:19:25   that aside, say they aside to that. Oh, the new CEO comes in and says, we want to get rid of the Mac.

01:19:28   So, of course, we have to port Xcode to the iPad and then we'll be done.

01:19:32   I think especially in the current age of excitement over AILM stuff, where is all that action happening

01:19:40   on Apple's platforms? It's not happening on the phone. It's not happening on the iPad. It's not

01:19:43   happening to the Vision Pro. It's happening on the Mac because the Mac is the more technically capable

01:19:49   and more open platform. And whether Apple recognizes it or not, I would encourage them to recognize

01:19:55   the openness and power of the Mac is what is allowing Apple to participate in this stuff at all.

01:20:01   Because if Apple didn't have the Mac, nobody would be doing trying to do this stuff on the phone or on

01:20:08   the iPad because they're just too closed and make it too difficult. Whereas you can let command line

01:20:12   stuff run wild on your Mac and screw everything up. And that's what people are currently excited about

01:20:16   and doing. And so I think Apple should not get rid of the Mac. I think Apple should not close down

01:20:22   the Mac any further. And I think they won't. I don't think Ternus is going to do that. He seems

01:20:27   like he's a big Mac guy and he's a nerdy person and those people tend to like it. So don't look at

01:20:32   the percentage revenue. Look at how important this platform is to the company. It is immensely important

01:20:37   and like necessary for the company to continue to exist now. And if Apple removes that necessity by

01:20:42   allowing you to develop software for their other platforms, not on the Mac, it would still be dumb

01:20:48   because the openness of the Mac is super important to Apple being even in the conversation for future

01:20:54   innovations like AI stuff.

01:20:55   Yeah, I would I would go a little further and like, first of all, like what is a Mac? It's a

01:21:01   computer running Mac OS. And that that has shifted a little bit over time. And you know, all the different

01:21:07   form factors and technical implementation details of what exactly that means. But for the most part,

01:21:12   it's, you know, in broad strokes, a Mac is a computer that runs Mac OS. If you have an iPhone

01:21:19   and you can plug it into other hardware and it becomes a keyboard pointing device and a screen

01:21:26   running Mac OS, that's a Mac. And, you know, maybe it's an iPhone when you remove it from all that

01:21:33   stuff. But like the combination of all those things, that's a Mac.

01:21:35   And as we've discussed in the past, putting two OSs on the SSD on the iPhone and having to deal with

01:21:40   this, but like it's a probably a bad compromised Mac and a bad compromised phone. But if it is truly

01:21:47   a Mac, it will be running Mac OS and you don't want your phone to be running Mac OS. So it's also

01:21:51   got to be running iOS. And now you've got a device that's running two OSs and virtualization and blah,

01:21:54   blah, blah. And there's so many compromises that have to do with that, which is why Apple has never

01:21:58   actually shipped that product. Could you do it? Absolutely. A hundred percent. You could do it.

01:22:01   But I don't think anyone would want that product. It's just easier to make phones and Mac

01:22:04   separately. And so Apple does. Yeah. And and as as for like, you know, the the software

01:22:09   differences between them. I don't think it's possible to open up enough pathways in the iOS

01:22:20   software architecture for apps to make it able to do the kinds of apps people love the Mac for

01:22:27   people love the Mac. You know, software development is is a big one. But what is it about software

01:22:33   software development that makes people want to and be able to do it on a Mac and either not want to or

01:22:39   literally not be able to do it on iOS? It's a bunch of characteristics of what tools software

01:22:44   developers use. Part of it is, yeah, command line stuff is often involved or scripting. A lot of it

01:22:50   is multiple applications operating on the same data, different tools integrating with other tools and

01:22:56   lots of different things, fast switching, multitasking, like a lot of that stuff either is impossible on

01:23:03   iOS or is just such a pain in the butt that nobody wants to do it when a Mac is available.

01:23:08   And to actually change that about iOS, you're just making a Mac at that. Like if if you have all

01:23:17   those. Yeah, like they wouldn't the way they would do that on iOS, like it's totally technically possible.

01:23:21   You know how you run a virtual machine that essentially has Mac OS inside it and everything

01:23:25   that runs in that environment is running in a big shared Mac inside your phone. But like

01:23:28   you can't break the sandboxing and security of the iPhone because like, why would you? You destroy the

01:23:34   security model of the iPhone. But hey, I want to do all this stuff where everybody can see everybody

01:23:37   else's data. I want to run full flight X code. No problem. Or run you in a little VM inside the phone.

01:23:41   Now you've got a little tiny world that's kind of like a Mac inside there where everything can see each

01:23:46   other and you can run command line stuff. It's like it's like the Linux virtual environment and

01:23:49   Windows or whatever. Totally technically possible by essentially creating a new separate little world

01:23:55   inside your phone where that stuff is allowed to happen. So even though you have technically

01:23:59   modified iOS to do all the stuff that X code needs to do, you've done it by basically boring a little

01:24:06   hole in iOS and saying, here's this little playpen where X code gets to live. And it's essentially a

01:24:10   environment that's more like the Mac inside the iPhone environment.

01:24:14   Yeah. And it isn't just software development that works this way. There's lots of tools and

01:24:19   applications and needs that people have on computers that just don't fit well with the iOS like super

01:24:28   isolated software model. And I don't think iOS will ever get there because that's not what makes iOS

01:24:35   good. Like if anything, what we've seen, you know, in recent years with trying to make the iPad more

01:24:41   Mac like in some of its windowing and multitasking features, what we've seen is that it often is kind

01:24:46   of a sidestep. And a lot of iPad users actually don't like it and want to disable it or switch it

01:24:50   back because it's a little clunkier or it's a little bit different. And it's kind of a mixed back. It's

01:24:54   hard. It's very difficult to advance iOS, especially like once you're at the iPad scale, you're trying to

01:25:00   make it closer to Mac. It's very hard to advance iOS towards the Mac in a way that doesn't make it worse

01:25:09   at being what people loved about iOS before or what has worked very well for iOS before. So I don't

01:25:16   think I think the whole idea of iPhones and iPads replacing Macs, that idea is probably not going to

01:25:24   happen. You know, what happens in computing usually when we have a new platform come out, usually or a

01:25:31   new form factor, usually it doesn't replace anything. It just adds people. People might choose to have

01:25:37   multiple ones or, you know, like when the watch came out, the watch didn't replace the phone. Some people

01:25:41   said it would. It didn't and it won't. And I don't think it's an inevitability that our phones or iPads

01:25:49   are going to replace Macs or PCs. What we've seen in the computing industry is that there are certain

01:25:56   form factors and ways of interacting and ways that the devices work that just hit a real sweet spot for

01:26:03   people. And they tend to last, even though all the people who love talking about tech in the future

01:26:08   are saying, oh, this thing is going to die. Like certain things never die. I think the most obvious

01:26:13   thing in what we know today that probably has the longest lifespan is the smartphone. The phone is

01:26:19   just so ideal for people in so many ways, in so many like like basic physical ways. The phone is great

01:26:28   and very compelling and such a great balance of lots of things. I think the phone is going to be with us

01:26:32   for the rest of our lives and probably significantly after that and not just existing. I think being like

01:26:38   the central computing device for most people, I think it's the phone for a long time. But also

01:26:43   the computer, the way we know it as like the Mac and Windows PC today, like that style of computer,

01:26:50   I think is another one of those sweet spots. People have tried over the years to replace it and to make

01:26:58   it obsolete and or to phase it out. And it doesn't work because people love it and it keeps working

01:27:03   really well. And lots of other things like, and look, you can part of what you can say against the

01:27:08   Tim Cook era at Apple is that Tim Cook clearly had the idea for whatever reason that the iPad was going

01:27:15   to replace the Mac and that, you know, the iPad was the modern platform and the Mac was the legacy

01:27:20   platform. And eventually he seems to have changed his mind on that. But for the first, I don't know,

01:27:24   five to five to ten years maybe of his leadership at Apple, that seemed like that was the plan.

01:27:30   And customers didn't go for it. The iPad was not the future of all computing. It was the future of

01:27:36   the iPad. And it has a market. It has uses. But it didn't replace the Mac or the PC for that many

01:27:43   people. And Apple sure did try. They tried really hard to make to push it that direction. And it just

01:27:49   didn't go that way. I think people have shown over and over again, they love their phones and they love

01:27:55   their laptops and they don't necessarily want those things to merge. And no one's asking for those

01:28:00   things to merge. And finally, one other thing to consider is that the the computing industry for a

01:28:07   long time has kind of been it keeps pushing this this this command and this assumption and this goal

01:28:14   of like computers are too hard to use. We need to make computers easier for everyone, everyone else,

01:28:21   all the non nerds easier to use. And there's a lot of truth in that. There's also a lot of like

01:28:29   kind of, you know, false like advocating for for groups that don't exist anymore or that exist a lot

01:28:34   smaller than you think or speaking for people or making decisions for like theoretical people that

01:28:39   you're not really talking to real people. And I think what's also happening is all the people who

01:28:44   don't know how to use computers are aging out. Most people today are fine operating a computer to the

01:28:53   degree they need to operate it. And so the the whole push of like, let's get rid of the Mac. It's too

01:28:59   complicated. We're gonna make everything simple. Not only is has that failed over the years, and not

01:29:05   only is ios getting more and more complicated over the years to basically match the complexity of of a

01:29:11   Mac or PC in many ways. But also, the demand for that the need for that is just a lot lower than it

01:29:18   used to be. Because way more people are when more people have basic proficiency at using computers now.

01:29:26   I don't think that's why the demand has gotten lower. I think the demand has gotten lower because

01:29:29   Macs have actually gotten less intimidating, like the Apple has done a lot of work on the Mac over

01:29:36   many, many years to take those things that were the most difficult and confusing and make them less.

01:29:42   So there's still a gap. Obviously, a Mac is easier to screw up than a phone by a million miles or whatever.

01:29:46   But compare how difficult it is to wrangle a Mac today, like compare like a MacBook Neo to like a

01:29:53   a powerbook running classic macOS night and day level of expertise required. So I think people

01:29:58   haven't changed. I think Apple has successfully over many decades made the Mac easier to use,

01:30:05   not as easy as an iPad, not as easy as an iPhone. But like to your point, I think they're reaching a

01:30:10   point where they're they're getting to a threshold where it's like you can no longer argue for the

01:30:15   elimination of this platform due for it due to its intractable difficulty, because now we've made it

01:30:20   it reliable and easy enough that to the average person, it is not impossibly difficult. And that

01:30:26   wasn't true. In the early days of the PC, the PC was essentially impossibly difficult to wait for way

01:30:31   too many people. If you weren't a computer nerd today, you don't have to be a computer nerd. You can throw

01:30:35   a MacBook Neo at anybody. It's more complicated than a phone. It's more complicated than an iPad. They can

01:30:41   screw it up more easily, but they will be successful in using it to a degree that you could not say for

01:30:48   a Mac or any PC 20 or 30 years ago. Yeah, but but I think, you know, both things have happened like the Mac

01:30:54   is easier to use today, but also people like way more of the population is proficient at using Mac and PC

01:31:01   style computers today as they used to be. But and also no, they they get they can I think you're

01:31:06   right. When it was your case, you said they can do what they need to do with them. But it's the it's the

01:31:11   joke about Gen X being the only generation that knows how to use computers because we have to we had to help

01:31:15   our parents with it when we were kids. And now I have to help my kids with it when I'm an old person.

01:31:19   Yeah, but but it's because proficiency does not mean power user. It means you you're able to use the

01:31:23   computer to do what you need to do with it. I know. But even that like still, it's like I can't print. I mean,

01:31:28   the eternal one. I can't why can't I printing is not a power user feature. It's a thing that you have to do every once

01:31:33   in a while. Why can't I print? My kids can't figure out why they can't print. My parents can't figure out why they

01:31:38   can't print. I can figure out why we can't print. And my son has it now has a computer science degree. So like, what's his

01:31:43   excuse? Right. Well, but also and keep in mind also like having the smartphone take over the world has also affected

01:31:51   this greatly. So like, first of all, way more people are able to use smartphones because they are simpler and in a lot of

01:31:58   ways. And so many people's needs are perfectly satisfied with their smartphones. They many

01:32:02   people. Yeah, they don't they don't even have a Mac or PC because they don't need one. Right. Like many

01:32:06   people who don't who don't want a computer or can't use a computer get their needs solved with a phone

01:32:12   just fine. Because having it and for that point, having a personal computer or a Mac is not a requirement

01:32:18   to live in today's society. Having a smartphone essentially is. And that also like, I feel like,

01:32:25   you know, many people who never learned how to use Mac and PC computers, they didn't need to like

01:32:32   they weren't really ever forced to whereas today, phones are pretty important for lots of society. And

01:32:38   also, because people tend to like them more than they used to like old computers. People are people who

01:32:44   are who are not into computers are are more motivated to use their phones more often and therefore actually

01:32:49   become more proficient just by experience with them. Whereas that motivation like wasn't there for

01:32:54   them for computer necessarily. So the need to wedge all the all this simplicity into Mac OS and ruin what

01:33:03   the Mac is or get rid of the Mac over time. That need is gone. Meanwhile, the MacBook Neo just came out.

01:33:11   It's an it's a Mac for the price of an iPad. And everyone is loving this people are like, people are so much

01:33:18   more excited about the MacBook Neo than I have heard about the about any iPad model released in probably the

01:33:26   last eight years. Like I can't I can't imagine I can't remember the last time anybody was this excited

01:33:31   about an iPad. The MacBook Neo is a hit. Why? Because people who want an iPad mostly that's like, you know,

01:33:38   more basic needs like watching TV and stuff like that. And and that needs been solved for a long time by the iPad.

01:33:42   That's nothing new. But people love having the Mac in a lower price point. And it doesn't need to be a super

01:33:52   powerful MacBook Pro. It like it wasn't about computing power. It's a and that's how we like you take an iPhone chip

01:33:58   and you put it in a Mac and you have, you know, a pretty limited Mac in terms of like sheer hardware capability.

01:34:05   But people love working that way. They love using a computer like that. The great reception of the

01:34:13   MacBook Neo, I think, proves this point. What people have wanted all this time that Apple was trying to

01:34:18   make the iPad the future of computing. What most people want for their computing is a computer. And

01:34:24   what most people like about the Mac is not replaced by an iPad with a keyboard that ends up costing

01:34:30   $1,300, you know, for the combo or whatever. Like people like a Mac and making just a cheaper Mac with

01:34:37   lower hardware specs. People love that. You know, meanwhile, the idea of making the iPad the future

01:34:43   computing, it seems like that is increasingly not the case. The iPad is another platform that many

01:34:48   people use and love, but it is not going to replace the Mac.

01:34:53   All right. And then finally, Paul Franz writes, a recent Mac OS bug report suggests that Mac OS will

01:35:00   crash after reaching 49 days of uptime. John Gruber in his article about it says that 91 days of uptime

01:35:05   is remarkable. Both of these numbers seem really small to me. I run Linux at work. Some new desktop

01:35:09   software has been causing me problems and our desktop support team pointed out my uptime was over a month

01:35:14   and could use a reboot. Is my experience with Linux servers giving me unreasonable expectations for

01:35:18   desktops. Unless I'm out of the loop, I don't hear this brought up with other devices running a Linux or

01:35:22   Mac OS kernel like phones, watches, TV, or cars. To be fair, I looked at my Mac OS uptime going back to

01:35:28   November on my laptop and the longest uptime was 40 days. My Linux server's longest uptime was seven

01:35:32   months. It's W temp goes back to 2022. I believe in all these cases, there were OS updates or needing to

01:35:40   physically move my server. Rebooting for updates seems less annoying than things breaking and requiring an

01:35:45   immediate restart. I don't think the expectations are out of whack. But as was kind of alluded to here, if you don't

01:35:52   shut down your Mac every day, like Jason Snell, because you're a weirdo, if you have a Mac, and you're

01:35:58   conscientious, as Marco suggested at the beginning of the show, your uptime is going to be determined by Apple's OS

01:36:04   release cycle, because you have to reboot when they update when they do a point release. So 15.7.7 comes out, you do the

01:36:10   update, you reboot. That is for me, for someone who leaves my Mac on all the time, and you know, I just put it to sleep when

01:36:16   I'm not using it. And then I wake it up when I'm using it, but I never ever ever shut down. My uptime is 100% determined by when

01:36:23   Apple releases software updates. I tend not to do a software update right before I record a podcast. So maybe it'll be delayed by a couple

01:36:29   days. But that's it. But if I didn't do that, how long would my uptime be? I mean, essentially forever. Like I, you know, my I have had

01:36:37   kernel panics on this Mac at various times, especially with all my video card swapping and stuff. But

01:36:41   it doesn't happen very frequently. So I I'm sure I could have had like multiple years where I just had

01:36:47   an uptime of multiple years. But like, well, here's the problem is the other problem. You reboot for all

01:36:52   sorts of reasons. If when I go on a long vacation, I like to shut down and unplug my computer when I

01:36:57   reboot into Windows, there goes my uptime. But like, theoretically, if you have a stable system running

01:37:04   Mac OS 10 or Mac OS or whatever, there's no reason why you should ever need to reboot a

01:37:08   short of software bugs. And software bugs requiring a reboot are very rare. Even bugs that require like

01:37:14   everything is frozen and dead. If you just kill login window for SSH and kill login window, your uptime

01:37:20   will be maintained, even though you just you just booted yourself back to the login screen. But that

01:37:24   doesn't affect your uptime because, hey, the machine is still up. So now obviously Linux servers that are

01:37:29   just doing one thing over and over again and don't have any software updates that are running on them,

01:37:33   although I think that is also rare and it should be rare because you should be patching your you

01:37:37   should be updating. So again, to Marco's point, you should update everything, even your servers every

01:37:41   once in a while. But in theory, people have these Linux servers. They never update them. They run the

01:37:45   same software year. Like the only thing that's going to take them down is a hardware problem or

01:37:49   whatever. And I think that is also true of any modern device like an Apple TV or a phone or a Mac or

01:37:55   whatever. But the world we live in is a world where there are software updates and you should apply them a lot of the

01:38:02   times because a lot of times they are security related. And also software does have bugs.

01:38:06   So, yeah, I don't think it's that if someone just says, hey, your Mac has been up for a long time and

01:38:11   you should reboot. That's maybe not great advice. But the the one of the things that I tell my kids

01:38:18   every time they tell me their computer is doing something flaky, like, you know, it's the whole turn

01:38:23   off and turn on again. Rebooting is one of the first best easiest diagnostic steps. Did you have

01:38:29   to reboot? It's like, no, but like it takes two seconds. Max boot really fast these days. And if

01:38:33   it solves the problem, like then try that first before you come to me, like X isn't working,

01:38:39   Y isn't working. And very often it's like some app isn't working. It's like, I don't know, log out and

01:38:43   back in, which is the software should never reboot, which is like maybe that app has a bug and it's

01:38:48   screwed up in some way. And if you log out and back in, it will fix it. But if nothing fixes it,

01:38:52   you can reboot. But if there's nothing wrong with your computer and you're like, oh, you have up

01:38:56   time of a month, you should just reboot just to be safe. No, if it's working fine, you don't need to

01:39:01   reboot. I remember when I was first on my own, I ran some flavor, if I want to say Ubuntu, on like a

01:39:10   ThinkPad or something like that. This was 20 years ago, so I can barely remember now. But I remember

01:39:15   being very impressed by the uptime of that machine. And I remember being so impressed by

01:39:19   that when I at one point needed to either shut it down or reboot it or whatever the case may be,

01:39:23   I actually captured a PDF of a little like diagnostic page that I'd written and the record server uptime,

01:39:30   which was recorded actually a month after me being married to Aaron. So this was 27, it's 2007,

01:39:37   excuse me, server uptime, 299 days, six hours, 28 minutes. And I was very proud of that at the time.

01:39:42   Yeah, if you search the internet, you'll find people with like this Linux server has been running

01:39:46   for 37 years, not joking 37 years, like literally, I don't know if that fits with Linux, but there are

01:39:52   Unix servers, they're like, I'm sure there are some of those like, you know, very fancy mainframes with

01:39:57   redundant CPU and power supplies that are doing like visa transactions for all of our lives. Like,

01:40:01   that's a little bit fudge where it's like, well, it has the same thing been up because they swapped

01:40:05   out the CPU a few times. But anyway, yeah, there are lots of long running things out there. And

01:40:10   Linux, Linux is one of the first ones you start to see people brag about it, because like,

01:40:14   regular tech nerds would just like set up a Linux server in their house somewhere. And then they'd

01:40:20   wake up 25 years later, and they realized they'd never rebooted it. And they'd post a cool screenshot

01:40:24   or whatever. But yeah, as with all things, again, I always find it hilarious when people post

01:40:31   screenshots as proof of anything. It's like, you know, you can make any pixel in any arrangement and

01:40:35   a bitmap. You know, you can just like, like, I can turn on and off any pixel, and I can make any pixel,

01:40:42   any color screenshots are not proof of anything. But yeah, you can just edit WTEP directly. And then,

01:40:50   and then type uptime and be like, wow, uptime of five millennia. But now I'm assuming people are being

01:40:56   honest, because yeah, computer systems can be made reliable in the right circumstances.

01:41:01   Who knew?

01:41:02   All right, thanks to our sponsors this episode, Aura Frames and Delete Me. And thanks to our members

01:41:08   who support us directly. You can join us at atp.fm slash join. One of the many perks of membership

01:41:13   is ATP Overtime. This is our weekly bonus topic. Every single episode, there is more content that

01:41:20   non-members, sorry, you're missing out. Members, you get it. It's great. This week on Overtime,

01:41:24   we're going to be talking about, basically, should Apple make Neo versions of more products,

01:41:29   more affordable iPhones, Macs, etc. We're going to talk about that in Overtime this week. You can

01:41:34   join to listen at atp.fm slash join. Thanks, everyone. And we'll talk to you next week.

01:41:43   And you can find the show notes at atp.fm. And if you're into Mastodon, you can follow them.

01:42:13   At C-A-S-E-Y-L-I-S-S. So that's Casey Liss. M-A-R-C-O-A-R-M-N-G. Marco Armin. S-I-R-A-C-U-S-A-C-R-A-C-U-S-A-C-R-A-C-U-S-A-C-U-S-A.

01:42:29   It's accidental. Accidental. They didn't mean to. Accidental. Accidental. Tech podcast so long.

01:42:42   I did want to call out the update to Pedometer Plus Plus. This is our friend David Smith,

01:42:47   underscore David Smith's app, Pedometer Plus Plus. I've mentioned a little bit here and there as I

01:42:52   prepare for my giant walk around Manhattan, which is actually this weekend, my big 32-mile walk.

01:42:58   It's in a couple of days. As I've been testing out different watch options for that, the app I've

01:43:05   been using on the Apple Watch Ultra has been the beta of Pedometer Plus Plus 8. And that just came out,

01:43:12   I think, yesterday or the day before. It's brand new. And underscore has a couple of great blog posts

01:43:17   about it. Our friend Stephen Hackett over at 512 Pixels also did a great post about it.

01:43:21   It is such an amazing app because what I love about Pedometer Plus Plus is, and I made a little

01:43:30   Mastodon post about it earlier, but I'll repeat myself here.

01:43:34   In the app store, you know, in the kind of Apple developer ecosystem, sometimes you get an app that

01:43:41   somebody has just poured so much love into, even if there might not necessarily be like an ROI that

01:43:49   justifies that level of investment. This is something that like, this is what people love about the Apple

01:43:58   indie developer scene. This is what made a lot of us fall in love with the Mac and certainly the

01:44:04   software scene around it. That, you know, just people who put in a huge amount of care and whether

01:44:12   that's design, engineering, polish, you know, some combination of all three, like a huge amount of

01:44:18   care going into a fairly specialized app. More care than that, than that app necessarily like quote

01:44:24   needs in terms of like what would the market pay for or support or whatever.

01:44:28   But mostly because that person or company just cares so much about making something great in that

01:44:34   area. Usually it's like, it's something they personally love and they just want it to be great.

01:44:38   You want to hear the great phrase from the gaming industry for the phenomenon you're describing?

01:44:42   It's called over delivery.

01:44:44   Hmm. I like that.

01:44:46   It's said in a pejorative way in various contexts, but yeah, over delivery, you are delivering too much.

01:44:52   Right. And, and we hear, you know, I think another term for this is a labor of love, which is also

01:44:56   sometimes used pejoratively. I think there, there is a lot of truth to that. Like, like there are certain

01:45:00   apps that you, you look at and you're like, oh, this was a labor of love. And I, and I personally

01:45:05   take that as a compliment when somebody calls something that I've made, you know, something like

01:45:08   that. The way to use it pejoratively is like, wow, this, this idiot wasted their time doing this thing

01:45:13   that's not worth it. But the way I see it is, is much more generous. It's more like, wow, like this,

01:45:19   this person put so much love into a thing, not because they needed to, but because they wanted

01:45:24   to. And that's, you know, as long as, you know, the bills are getting paid, it's fine. Right. And so

01:45:29   pedometer plus plus, you know, underscore makes widgets myth. And by all accounts, widgets myth is

01:45:35   so ridiculously popular that numbers wise, he probably shouldn't spend any time ever working on anything

01:45:43   else. Like if you just calculate, like, you know, what, where does it make sense to devote your time?

01:45:48   Like you have this one massive mega hit product, like it doesn't make a lot of sense. You know,

01:45:53   like Tim Cook wasn't having like a side gig, you know, making, making little apps for himself. Like

01:45:57   he had enough to deal with, you know. And, you know, so underscore doesn't need to make pedometer plus

01:46:03   plus. He makes it because he loves it and he makes it for himself. And this update to this app,

01:46:07   like you think it's just a step counter and it is a pretty good step counter, but then behind the step

01:46:13   counter is this like world-class hiking and walking tracker, especially in the Apple watch.

01:46:21   Like if you are walking or hiking and it actually supports other types too, but those, I think those

01:46:25   are the big ones, walking, hiking, anything where you, where you could use a map on your wrist or

01:46:31   distance measurements on your wrist or step counts or whatever. Um, it actually probably would be good

01:46:35   for cycling too. Um, but anyway, um, if you are doing like, you know, walking, hiking, running, cycling,

01:46:42   things where you are outside following trails or maps or whatever, the experience of using pedometer

01:46:48   plus plus on the watch, it was always good, but now it is even better. He brought in a designer to

01:46:54   help him make these amazing, he brought in a cartographer to help him make the maps look better.

01:47:00   How many people do you know who, who have hired a cartographer for anything? Um, but that's,

01:47:05   he brought in a cartographer, he brought in a designer. It looks incredible. It works great.

01:47:10   The engineering behind it is great. It destroys the like, you know, in terms of like how nice it looks,

01:47:17   it destroys all the other fitness watches and Apple's built-in apps and any other app I've ever seen,

01:47:23   uh, third party or first party, um, in terms of niceness of these, of this experience. Uh, and

01:47:28   my, one of my favorite things too, is that it has this, um, this expedition mode where to extend the

01:47:35   Apple watch battery life, there's a few tricks you can do, but he, if you want your Apple watch to last

01:47:41   like multiple days of a hike, you can turn an expedition mode and that's a special workout mode

01:47:46   that doesn't use the heart rate at all. Like if, you know, if you're just walking, if you're hiking,

01:47:50   the heart rate might not, might not be that relevant to you. Um, so you can turn that off the heart rate

01:47:55   and get 40% more battery life, which is massive. So all this is to say like, this is an amount of

01:48:03   coolness and niceness and, and, and effort into an app that this one guy has, I think embarrassed the

01:48:12   rest of the industry with how good his app is for these purposes compared to the built-in Apple

01:48:19   workouts app, the, the Suunto, the Garmin's, like all those watches of the world. Like this app is

01:48:24   awesome. So if you go outside and walk or run or bike or hike, which is just walking, I think, but

01:48:30   in the woods or something like if you do those things, check out Pedometer plus plus, especially

01:48:35   on the watch where it really shines. It's a great phone app as well, but the watch and like

01:48:40   no one else is doing Apple watch development worth a damn except underscore. Like he, he blows

01:48:46   everyone else away in the term, in terms of like how much he, how much effort he puts into the Apple

01:48:50   watch as a platform because it, the platform fights you at every turn. It's very developer hostile for

01:48:56   lots of, lots of good reasons and some bad reasons. Um, and he gets through it and he does great work on

01:49:02   the watch. And so it has made me like re fall in love with the Apple watch as a fitness device for that

01:49:08   particular purpose for the walks. And even though I also like the Suunto for other reasons, um, man's

01:49:14   Pedometer plus plus is so good. So check it out.

01:49:16   Yeah. When we, uh, went on spring break a couple of weeks back or a month back, uh, we did some hikes,

01:49:22   including one that was, I think I described it on the show, two miles out, two miles back.

01:49:26   And what I did was I found on like all trails or something like that. I found a GPX, which is,

01:49:30   you know, a series of waypoints. Uh, I think it's XML, but it doesn't really matter. It's really ugly.

01:49:34   Yeah, there you go. Um, I found that and sent it to Pedometer plus plus, and then it could,

01:49:40   it would show me, am I on or off the track that was provided via the website, the all trails website,

01:49:45   which was really, really slick. It's really good stuff. Um, and that was the old version.

01:49:49   Uh, that was version seven or whatever of Pedometer plus plus. This is, this is really

01:49:53   good stuff. Uh, underscores put in a, an astonishing amount of engineering work. And like Marco said,

01:49:59   I mean, who says I hired a cartographer? How do you even do that?

01:50:03   Where do you find, yeah, do you go on LinkedIn? Like what?

01:50:05   Right. Like how do you find your friendly neighborhood cartographer? I mean, just ridiculous.

01:50:09   Next, he's next to the cobbler.

01:50:11   Right.

01:50:11   And the blockbuster.

01:50:12   I've said many times throughout the course of the show. The reason, one of the main reasons I

01:50:17   think that we're, we're successful, the three of us at doing this show is, is really two actually.

01:50:22   Number one, you know, I think the three of us have a pretty good chemistry. And even though

01:50:26   sometimes we want to murder each other, by and large, we're really dear close friends. And I

01:50:29   think that comes through and shows, but number two, and what's relevant here is that we really

01:50:34   and truly give a crap. We don't always succeed and execute perfectly, but we really and truly

01:50:39   give a crap. And that is something that is extremely true of underscore broadly and of Pedometer

01:50:45   plus plus specifically. And like Marco said, I mean, I don't know anything about underscores,

01:50:51   you know, uh, business finances or anything like that, but I have to assume that just an

01:50:57   astonishing, overwhelming share of the money that his business earns is from widget Smith,

01:51:02   which he also gives a crap about. But nevertheless, because this is a labor of love complimentary,

01:51:09   uh, because this is something that is scratching an itch that, that underscore very much has,

01:51:14   uh, he still spends a lot of time on Pedometer plus plus and it shows, and I haven't had yet had the

01:51:19   chance to read. Um, he also put up a blog post about how he spent, this is literally the title of his

01:51:26   blog post, six years, perfecting maps on watchOS. And that's both a testament to underscores,

01:51:33   tenacity and, and perseverance, and also how shitty watchOS development is. But anyway, you get the

01:51:40   point. So, uh, yeah, you should check it out. It's, it's good stuff.

01:51:43   I have one, uh, more unrelated tidbit on this. Um, back in the classic Mac OS days, uh, when it was the

01:51:50   wild west and there was no memory protection and there was cool little system extensions to change

01:51:54   the way everything looked in the operating system. Um, there were various engines that could do cool

01:51:59   things like, uh, make it look like your windows are transparent. It's, uh, you know, something that

01:52:05   is trivial now, everything has it, but before the age of compositing window managers, when your app

01:52:11   drew a window, it didn't know what was like, like the, the system when you're dragging a window around

01:52:18   didn't know what was behind it. It was just like telling the things to draw the new dirty areas of

01:52:21   the screen. There was no compositing window manager, which meant that like in a compositing window

01:52:25   manager, some part of the operating system says, okay, give me all your windows. You want a white

01:52:30   window here. You want a yellow window here. You want this window here. And by the way, the windows

01:52:33   can also have transparency values. Like I got all the windows. I know where they all are. You've told

01:52:37   me what's in the bitmaps for them. And you've also told me for every single pixel, how transparent it

01:52:41   should be. Now I will combine all those windows into a final image that you will see on the screen by

01:52:47   compositing them together. I'll take the back thing and then I'll put the next layer over it. Then I'll put the

01:52:51   next layer. Then I'll put the next layer. And as I'm layering them on, Hey, if I'm layering a window

01:52:55   on top of some other stuff and that window has some pixels that have some transparency, I will blend

01:52:59   them with the stuff that I know is behind them. That's what a compositing window manager does. That

01:53:03   was introduced with Mac OS 10 before that, that was not how it worked before that application is drew

01:53:08   directly into like a backing buffer for the entire screen. And they just drew opaque pixels. Um, but if you

01:53:14   were clever, you could say, Hey, I'm about to draw my window onto the screen and there is stuff already

01:53:19   on the screen. So before I draw what's on the screen, I'll look to see what's in the area that

01:53:24   I'm about to draw the window. And I will blend those pixels with my pixels in a transparent kind of way.

01:53:30   And so what they were doing is drawing an opaque window that contains contents that look like it's

01:53:36   showing what's behind it. Now, the tricky part of that is if you then move something, if something

01:53:40   behind it changed, your app had no idea that it happened. You know, it's not like Mac OS 10 with the

01:53:45   operating system. I'd be like, I'm constant, every frame I'm compositing, compositing, compositing.

01:53:49   This would be like, well, when I drew this window, like if you, if you have like a clock behind you,

01:53:54   when I drew this window, here's what was behind me. But then a second passes and the clock ticks over

01:53:57   and your window has no idea that that happened because it has no awareness of what's going on

01:54:01   behind it. Right? So what I'm saying is you could get fake transparency, which only became apparent

01:54:06   when like you either move something in the background or something changed in the background.

01:54:09   And you're like, oh, I see. You're not really compositing. And like, you could fudge it by

01:54:14   saying, okay, every, every few seconds, you know, look at what's behind you and then redraw it

01:54:19   yourself. And every window could do that. Like that fake transparency. Here's why this is relevant to

01:54:23   your underscore discussion. Uh, I feel like Casey with this long, long way to, to, all right.

01:54:28   And speaking to like, you know, underscore cares about this app and everything. And Casey said,

01:54:33   and Widdersmith too, he cares about, I recall seeing a thread go by, I think on Mastodon where

01:54:38   underscore was describing what he wanted to do was make essentially transparent widgets,

01:54:43   but there was no good way to do that. So he's like, well, if I just know what's behind them

01:54:47   and I draw them exactly how they would be appear, if they were transparent, they'll look transparent,

01:54:54   but Oh, to do that, I have to know down to the, not just down to the pixel, but down to the,

01:55:02   like the, you know, but yeah, down, down to the, the screen pixel, not just the point,

01:55:06   but down to the screen pixel where everything is on the screen, because I have to draw my widgets

01:55:11   as if you can see through them. So I need to know where every single other element is on the screen.

01:55:17   So the illusion is perfect, but Oh, there are so many different sizes of iPhones. And also, by the way,

01:55:24   people can have that magnifying mode on the iPhone. So he wrote this program. I forget it was vibe

01:55:29   coded or not that would automate the process of figuring out every single Apple device,

01:55:34   every possible set of display settings, every possible widget arrangement on every possible screen.

01:55:38   So he would know, so he could exactly fake with widget Smith widgets that appeared in a way that

01:55:44   seemed impossible because he's faking it down to the pixel for any iPhone and any display resolution

01:55:50   and any zoom factor. And like, he cares a lot about everything he does, not just phenomenal plus plus.

01:55:56   And that's for a feature, which I think, by the way, is stupid, but he didn't think it was stupid.

01:56:01   He thought it was cool. And so he went like, talk about going the extra mile. It's like anybody would have

01:56:06   looked at that and said, yeah, you can do it. But I mean, it would be ridiculous. And underscore is like,

01:56:10   but you can do it, right? And he did it. He, I believe this is shipping. So all this is to say

01:56:17   that underscore doesn't just care about a widget Smith. He cares about, uh, or doesn't just care

01:56:22   about a pen hour plus plus. He cares about widget Smith and all his other apps as well. So check him out.