685: The Ability to Be Hotter

22:55Chromebook, that's more of like an os platform thing because you just sign into the chromebook with your google id and all your stuff is there because all your stuff is never there it's in the cloud that pulling that off like i don't think if you forgot your your macbook neo went to the library even if they handed you a new one could you have it up and running to the point where you could use it for something in time for your class given how long it would take for the mac to do its whatever network boot setup home directory configuration blah blah blah you know what i mean like having all your stuff there.

23:24I'm not sure how people configure MacBooks in school systems, but if it's a one-to-one type thing, I have to imagine there's more local stuff, or even if there's not any local

23:33Marco Arment:stuff. I don't see, I don't know. Like, I think a lot of, I mean, my kid only had a MacBook Air in a couple of, in the Fire Island grades, like a couple of elementary school grades, but now we're in a bigger district. It's all Chromebook. He's in eighth grade, but even districts that use MacBook Airs, I think they still just use a whole bunch of web-based tools, much of the time, like for most or all of the work. But you got to log in, right? Do they all log in? It's like a school account. Yeah. He has like a school Google account and logs into all these weird websites. Right. But is it individual school Google?

24:05I'm saying you got to type something in the login prompt. You have to log in as a user. Oh, I see what you mean. Yeah. I don't know. I guess I don't know how they do that. You know what I mean? And it's like, okay, well, if I just give you a blank one, what's the user account? And if there is no user account, how long does it take? Even if there's like a network thing that just sets it up, like you just it's single sign on, you enter in your school email address, like the time taken to just set up the home directory or whatever. I'm not saying it's going to take forever, but I just feel like Chromebooks are have an advantage here in that they can probably let you do a Google ID, you know, log in with Google type thing to a Chromebook faster than you can do the equivalent on the Mac.

24:38I'm guessing people can write in and tell me. The second thing is the repair. and that's where i think a lot of the teardowns in the neo come in here they're saying like oh well we have students replacing the screens on the chromebooks because it's easy enough to do that a student can do it and we have the parts and so on and here i think apple has come a long way so first of all as of several years ago apple will supply you with parts and instructions on how to do it for lots of its products including the neo second as we've covered the other teardowns the neo is way easier to take apart than all the other mac laptops there's this stuff glued in there you don't have to like you know take a heat gun to any part of it and break any seals and do all the other stuff a student with the appropriate weird you know i fix it pentalobe whatever things can in fact take apart a macbook neo and replace maybe not the screen because i think all the tear downs were like well we just we don't know how to get the screen off of the top lid but if you've got the screen and the top lid i think a student could replace it so although chromebook still have a lead here, I think the Neo is closing the gap by making it feasible to have a bunch of students whose job it is to grab from the bin of, you know, spare lids to the Neos and take a broken screen and put it in the recycle thing and put the new one on.

25:55So, you know, I agree with Henry that it's still not quite the same, but I do think the Neo is actually making some progress in the direction of the Chromebook advantages here.

26:04Marco Arment:Yeah, it'll help a lot. And I think whether, whether you know this whether you side with henry or ari on this one i think depends a lot on the supporting infrastructure around the the laptops themselves you know and your budget yeah budget because even then like you know even with the macbook neo being about 500 bucks for education that's still about twice as much as most of those chromebooks cost those schools so you're still looking at a huge price difference and of course then you know like you know at my kid school again it's all chromebooks um and they have a whole program there's like an insurance program that we buy for damage uh you know and you know there's he whenever something happens you know some the corner of the screen cracks or whatever um keys fall off like the weird stuff happens these are middle schoolers um but but like you know there he brings it into some place in the school and he gets a loaner and then he gets his original one back maybe like a month later you know there's all this support around it there's service depots there's in school service there's the insurance thing that you can buy to you know make get a damage waiver like there's all that stuff around and then even setting aside the service and the damage and the price then there's as you mentioned john the software and application and login management all of that chromebooks are popular in schools for lots of reasons it isn't just cost although the cost is a big one.

27:29Marco Arment:But all of those management things, like if you are a school that is all invested in that ecosystem, that all of those management sides of it, they make that very easy on the IT departments. And when you're running fleets of hundreds or thousands of these for all for a whole bunch of kids and a whole bunch of teachers, like it's a huge job just to manage those. And so if you have the infrastructure in place to manage Chromebooks and it works well for you, Apple's not really going to be changing a lot with the Neo. But if, as John said earlier, if you already have Apple administration, if you're already using fleets of MacBook Airs, then the Neo is a better situation for you because it costs less and is easier to service.

28:14Marco Arment:But if you don't already have all that in place, this might not move the needle that much. We are sponsored this episode by Masterclass. There's always something that you might want to get better at. So it can be things like how you communicate, how you negotiate, how you cook, how you think. There's so many wonderful options to help you with whatever you want to get better at at Masterclass. And their classes are taught by the people who are the best in the world. So unlike other learning platforms, Masterclass puts you in the room with the people who defined their fields, not just experts, but the best in the world.

28:51Marco Arment:And this goes across so many different categories, business, writing, cooking, creativity, wellness, and so much more. And Masterclass classes can fit into any schedule because not only do they have the video versions, but they also have lessons you can download for offline access, including audio only mode. So you can have your commute or your workout turn into a classroom. You can learn on your terms, whether you just have like 10 minutes sometime or whether you want to sit down and really like buckle down and see a whole bunch at once. You can do that with Masterclass. It's totally flexible.

29:24Marco Arment:Three in four members feel inspired every time they watch. 83 % have applied something they've learned to their real lives. So it also really works. And you can do all this with no risk because every new Masterclass membership comes with a 30-day money-back guarantee. Masterclass keeps adding new classes. There's never been a better time to get in. Right now, as a listener of our show, you can get at least 15 % off any annual membership at masterclass.com slash ATP. That's 15 % off at masterclass.com slash ATP. Head to masterclass.com slash ATP to see the latest offer. Thank you so much to Masterclass for sponsoring our show.

30:08John Siracusa:All right, let's talk about the M5 Pro and Macs and their, I guess, medium cores. We have a 9to5Mac write-up that's about an article on the Mac and I website, which is in German. So Johannes Schuster wrote and is quoting Anand Schimpie, who is on platform architecture at Apple. Apple's new performance core is a completely custom-designed microarchitecture, So it differs significantly from both the super core and the efficiency core. Furthermore, it manages to surpass the efficiency of the efficiency core. That's straight from Apple. They're making a straight up claim for what we're calling the medium core, just because it makes more sense.

30:46They're saying it's actually more efficient than the old efficiency core, which kind of leans towards maybe the M6 will use these cores instead of the old efficiency cores. But we'll see. Remains to be seen if they're even going to use the chiplet architecture for the plain no suffix M6. But that's a bold claim. don't worry about the medium cores not being as efficient as the efficiency cores they're

31:05John Siracusa:actually more efficient claims apple all right and then doug brooks who is mac product marketing explained that while the chips do support pcie 5 they are actually custom designed controllers yeah usually apple doesn't talk about these things like what you know what standards does it use or whatever pcie 5 i think is a standard from like 2019 or something so there you know there are later standards that are still in progress but in case you're wondering if apple is keeping up with the times. I'm not sure. Maybe the M4 and M3 also use PCI-E5, but you just so rarely hear Apple talk about this. And unless you find someone who's really dug into the chips, it doesn't come up, especially since it's not like there's any PCI slots or anything.

31:43But yeah, PCI-E5 is in there.

31:46John Siracusa:And then from the write-up on 9to5Mac, we were interested in whether two chips with the Fusion architecture could be combined using UltraFusion to create a single system on a chip consisting of four dies. That would then be the M5 Ultra, but Apple traditionally doesn't talk about upcoming products. Quote, at the moment, we have only announced the M5 Pro and M5 Max, was all Shimpy would say. That's such Apple. This is from the Mac and I website, by the way. It's not from the 9.5 Mac thing. So as I said, like this is translated by Google. So forgive us Germans if this is not good. Like literally, I'm just using Google Translate to translate a webpage.

32:20It's the best we got. But the, and again, the quotes are translated too. Like even though I don't think Anon speaks German, these quotes are in German and then translated back to English. But this is as close as you'll ever get anyone from Apple to confirm the M5 Ultra or some other M5 chip. At the moment, we have only announced the M5 Pro and M5 Max. You sure have. Yeah, so I'm looking forward to something that's not an M5 Pro and M5 Max in the Max Studio for WWDC.

32:46John Siracusa:Really, John? Are you now? I am. We'll get to it. All Pro and Max chips differ in the number of GPU cores, memory bandwidth, and maximum addressable RAM. We got the impression that Apple wanted to gain experience in experimenting with the core count. Quote, we're not trying to experiment, figure things out that way, explained Shimpy. Ultimately, we're not commercial semiconductor manufacturers. We develop chips specifically for a particular product generation. So many people have said this, both officially and off the record, this sentiment that Apple builds chips for products. Apple doesn't build chips and then figure out what to do with them.

33:16That's part of the whole deal with Apple Silicon is they're not just building. Oh, we should just have a, you know, a good line of chips and figure out what to do with them. it's like no when a chip is designed it's for specific products now sometimes it's easy it's like oh we're making the chip for the iphone you think we're gonna have an iphone we're totally gonna have an iphone but even there it's like what features will this iphone have what are the requirements of this iphone what things do you want us to make better than the previous one um and sometimes it goes awry as we'll discuss later um but even like we've heard things about um chips and entire computers they were originally designed and built and planned for a product that never shipped or shipped in a different form and just ended up being repurposed so you know things happen but this is gets back to what we're talking about with the uh the soic mh chiplet architecture thing for the m5 pro and m5 max where there's separate dies for the cpu and gpu it's like well now they've got this chiplet architecture they can mix and match all sorts of stuff and they're like no that's not how we do it yes we will mix and match things but only in the specific ratios the specific combinations that we know we need for our specific products that we're planning this for So, again, don't expect a pop-up menu on the Apple configurator where you get to choose the CPU cores and choose the GPU cores beyond the typical, like, binning and sliced up into pieces options that we get today.

34:34John Siracusa:All right, let's talk ubiquity. First of all, half of the internet wrote to ask, hey, Marco, why not use site-to-site VPN for your swarm in the data center and your house? And I haven't spoken to Marco about this, and I don't know what your answer will be, but I wanted to mention that former and I think future sponsor Tailscale, which I am freaking in love with, one of the great things about Tailscale is that it basically removes the idea of what network are you on when it comes to connecting to your other devices. So, yes, a site-to-site VPN would fix Marco's problems perhaps when he's at home.

35:10John Siracusa:But what if he leaves the home? And yes, I'm aware that there's like Unify's Teleport, which, by the way, I have not had a particularly great experience with. But one way or another, what's great about Tailscale is I can be with my phone. I could be out. I could be watching Declan, you know, play football and check on something real quick. I could be in an airplane. I could be on another continent. I could be on Wi-Fi, on cellular. it doesn't matter because you're always connected to your other devices, no matter what the networks are between them. So I'm reckoning that maybe, Marco, that's why you haven't tried this option, or maybe you're about to tell me you've given up on tail scale and that site-to-site is the right answer for you.

35:46Marco Arment:So I actually tried this with the restaurant back about a year ago. I just couldn't figure out a way to do it because the way Ubiquiti does those site-to-site bridges, by default it basically like it lets you you know transfer it lets you tunnel your traffic through the other place to go to the internet but it like totally blocks access to all of your network

36:07John Siracusa:devices casey you had this issue too right well no um i have briefly tried this with the unify travel router um which i am really a big fan of but my problem with teleport is that right i've never tried site-to-site vpn but with teleport it's extremely fickle like even with their own device sometimes it works and sometimes it doesn't and it's considerably less reliable

36:31Marco Arment:than tail scale is which is really weird yeah anyway so i i tried i tried teleport and i tried like various bridging options about a year ago i did not have any luck getting them to work to be able to see devices on the local network uh on the remote local network so i don't i can try it again sometime i would love to not need tail scale on every machine just to be able to access it remotely um but there are other options too like for instance i could like tunnel i could like set up tail scale on one machine there and then like tunnel to that one to access the other ones like i have other options um but right now this was just easy to do and it's fine um and i and i set the tail scale keys on those machines not to expire which is an important step i learned i I learned that the hard way for the first few machines I set up.

37:22Marco Arment:Yeah. So anyway, so far, the tail scale option is working fine for me. In the future, if it doesn't anymore, then I'll look at other options. Or if there is something that's really, really easy and that is reliable and flexible, then that's fine, too. But for the moment, it works fine.

37:38John Siracusa:Yep. All right. We talked, I don't know, maybe a month or so ago, maybe two, about MLO. What is that? Multilink, something like that? Multilink operation, I think. There you go. Thank you. and how basically it doesn't work anywhere as per some article that we had talked about several episodes ago. Well, Ubiquity in the last couple of weeks has debuted a new product called Airwire. And their release video, their trailer, if you will, and they do this for both software and hardware, their trailer basically said nobody could use MLO, particularly because the clients were always trash. And what we're going to do is we're going to have our own Wi-Fi client that you then plug into via Ethernet, if I'm not mistaken.

38:19John Siracusa:So the literal quote from the video is, the infrastructure was ready, now the clients are. And so there's also a blog post about this, reading from the blog post, nearly every Wi-Fi 7 client today advertises MLO compatibility, excuse me, capability, yet they only use one link at a time. Instead of combining bands through simultaneous multi-radio operation, they simply switch between them. Airwire, this new product, is a plug-and-play USB-C Wi-Fi 7, I'm sorry, it's not Ethernet, it's USB-C, Wi-Fi 7 client engineered for true simultaneous multi-radio performance with STR MLO. It operates independently on 5 gigahertz and 6 gigahertz at the same time, aggregating spectrum instead of alternating between bands.

38:56John Siracusa:The result is a real multi-gigabit throughput, ultra-low latency, and improved connection resilience. So, pretty cool. To describe this product, it's kind of like a, I don't know how you would describe it. It's bigger than a phone. It's kind of like a, you know, maybe 4-inch, three and a half inch by three and a half inch by one inch usbc rectangle that you plug in and it's got like a lid that kind of opens like a laptop lid and that's the antenna and it's like it's basically saying your laptop's wi-fi can't do this your max wi-fi can't do this whatever you have can't do this because nobody seems to ship uh simultaneous mlo wi-fi seven clients but we have a box that will do it plug this box into your wireless device stop using the wireless radio that's in your device and instead have your device through this USB-C connection, use our box and we'll do the Wi-Fi for you.

39:48And it's one of the clunkiest things I've seen since the track balls that would clip onto the side of a laptop really compromises the portability of your portable device. But if you absolutely positively need to have simultaneous MLO because it's the only thing that can provide you that extra bit of bandwidth and lower latency that you desire, apparently Ubiquiti will sell you this thing.

40:11Marco Arment:Yeah, Ubiquiti sells a lot of very specialized products. Like if you have very specific needs for this kind of thing, now here's a product for you. Almost no one should buy this because almost no one needs this. But if you do, here it is.

40:29John Siracusa:That's exactly right. Yeah, I would say this thing looks, and it's hard to get a good idea of scale, but it looks to me like the Mac minis that we had up until the most recent release with like a flap on the top, like John described. Oh, it's like an Apple TV, actually. No, yeah, that's a good, that's a better analogy. Yeah, I would say that. Anyways, it does look cool, but I agree with Marco, probably not for you or me. Now, a lot of you are probably already firing off emails to us because we said in the beginning of the show that we stand with Ukraine, and yet we are also talking positively about Ubiquity.

40:59John Siracusa:We have gotten this feedback a million and seven times, and so let's talk about it. on January 27th, what is the name of this place? Hunterbrook Media reported, official Ubiquiti distributors appear to have continued supplying Russia after the invasion of Ukraine, sometimes rerouting shipments through intermediaries in high-diversion-risk countries like Turkey or Kazakhstan, trade records show. Some appear to have used intermediates later sanctioned by the U.S. for export control evasions. Ubiquiti openly admits, quote, we do not have any visibility, quote, over purchases from its distributors, but legal experts told Hunterbrook that that's not a viable defense.

41:33John Siracusa:U.S. export controls and sanctions operate on a strict liability basis, meaning even unwitting violations are still violations. Ignorance is not really a practical excuse or rather a legal excuse, a former senior State Department sanctions official told Hunterbrook. A sanctions compliance lawyer added, you doing very little effort were able to determine that it's available for purchase by the Russian armed forces. The company's compliance team should be taking additional steps to prevent that. This sounds really, really crappy, and it might be really, really crappy, but there's a catch, John. Do you want to tell us about this?

42:05Yeah, there's a couple other things from the article as well, talking about how Ubiquiti did actually cut off direct sales to Russia when they invaded and all this other stuff or whatever. But here's the thing. One of the reasons this has been in the notes forever and hasn't even made it into overtime is because as soon as the story came out, I added it to the document and I'm like, oh, I'll just let that, it didn't make it in time for a show, but like by next week, I'm sure we'll have more stories and I can add more links to it. And there just weren't any. and every time someone brought it up i would say where did you hear about this and they would tell me and every single source was always pointing back to this hunter brook is it hunter brook media

42:38Marco Arment:i gotta look up hunter brook capital hunter brook media i guess is their media arm but it's it's a capital fund yeah anyway and like everything pointed back to hunter brook i'm like there was no corroboration from no other sources from literally every place you saw there's a youtube video about it all but it's the person for hunter brook on the video there's a blog post about it but they're just linking back to the Hunterbrook. Like other, is there any other source for this? And eventually it's just like, there's never, apparently there's never going to be a second source for this. It's just going to be Hunterbrook.

43:05So I figure we should talk about it just in case people want to know about it. Cause we do talk about Ubiquity a lot, but the disclaimer on the actual article that we will link from Hunterbrook says, uh, based on Hunterbrook media is reporting Hunterbrook capital is short on the stock symbol for Ubiquity and long a basket of comparable securities at the time of publication. So I forget, I couldn't look this up by Googling or, you know, I didn't couldn't find this in our history, but I believe we have talked in the past about a similar situation. Maybe it was also Hunterbrook where, where it's like some company shorts a stock.

43:36And if you short a stock, you make money if that stock price goes down and you lose potentially unlimited money if the stock price goes up. So some company will short a stock and then post a negative story about that company in the hopes to make the stock go down and disclaiming it and saying, here's a big negative story about ubiquity and by the way just so you know we make money if the stock price of this company goes down which is fine maybe they're shorting it because ubiquity is doing a terrible thing like you know it's not it doesn't mean that they're lying to try to make the stock go down maybe they're saying hey we're shorting it because we know they're doing a bad thing and we think people will agree with us and that will make the stock go down and then we'll make money but the fact that i cannot find a single other source that doesn't eventually lead back to this 100 brook media story makes me question this a little bit now i fully believe that first of all it's a fact that russia is using ubiquity stuff okay the question is what is ubiquity doing to stop that are they doing enough and again in the article you'll see that they have cut off direct sales to them and they they've banned all russian ips and they've done a bunch of other stuff but still the fact remains russia is getting these products through intermediaries and resellers and ubiquity is claiming ignorance or whatever so i'm not sure what to make of it because i'm not an investigative reporter on the Russia beat.

44:49But obviously, as we noted before, we do stand with Ukraine. Russia is terrible. They shouldn't be getting Ubiquiti stuff. The fact that they are something that Ubiquiti should stop. But I don't know how much credence to give this report that Ubiquiti is knowingly acting badly for its own self-interest. It's plausible. Companies do it all the time. But, you know, do with this information what you will. The link will be in the show notes. You can read it yourself.

45:13Marco Arment:Let me add a bit more information here. That's, I think, extremely pertinent to this. So this is a investment firm that has taken a short position. And again, so that means they make money if the stock goes down. Now, if you do this to something like Apple or Google, it's going to be hard to make a dent. But Ubiquiti is different. If you take a look at Ubiquiti's stock, it's pretty unusual in a few ways. One of the biggest ways is that the founder and CEO still owns almost all of the company. There's very little public stock actually issued, and there's not that much liquidity, and there's not that much trading volume because there's not that many shares.

45:56Marco Arment:So if you look at Ubiquiti's stock price over the last few years, it goes all over the place. It's easy to manipulate. So if you were trying to make a bunch of money on a short position, if you could put a story out there and the stock dives 20 or 30%, you're going to make a lot more money than if you try to hit a bigger stock that has a huge amount more trading volume and a huge amount more shares, and maybe you only make it go down 1%. So if you wanted to manipulate a stock in the tech business, Ubiquiti would be a really interesting choice for that because of the very unusual dynamics of its stock relative to other well-known tech companies.

46:40Marco Arment:So that's why you see, I think, a lot of these kind of weird stories trying to move the Ubiquiti stock because it works, and it works pretty effectively compared to most other targets. So I think stories like this, you have to take with a large grain of salt. If the Russian military has equipment from Ubiquiti, well, everyone has equipment from Ubiquiti. it's a very popular company that sells a huge amount of tech and networking gear all over the world if there's a huge market of demand for it and again this is not saying ubiquity sold it to russia this is saying that ubiquity is not adequately policing its third-party sellers to make sure that third-party sellers who buy it from ubiquity don't resell it to russia that's kind of a lot to put on the company honestly like i mean that's the law the whole point is that that's the law that you don't get to say well we just give it to distributors what they do with it after is not our problem is the question is are they doing enough that's what it comes down to right and it kind of sounds like they're doing a reasonably okay job of like doing their diligence and like not like knowingly setting it that way but like if they sell to somebody and then that person goes and sells it like you're never going to get 100 compliance And a product like products like Ubiquiti's products that have they sell in such high volumes and they sell all over everywhere.

48:10Marco Arment:That's going to be really hard to get 100 % compliance on. So given like the problem space that this is and then given that this – only this particular media reporting seems to be about this issue and that the company doing the media reporting stands to make a lot of money by a stock going down that is very volatile and swings in huge amounts based on media reporting, I have a hard time taking this story super seriously. i mean i i take it seriously but i just think like you don't know where the cause and effect is if you're trying to make money shorting a stock and you find a stock that is like as you described margo has high volatility um you would look at that and say is there any way we can make the stock go down and one of the ways you can make it go down is revealing true information about a bad thing the company is doing it doesn't mean the thing isn't true your motivation for doing it is for you to make money not because you're just magnanimous and want the world to know about this bad thing but you know it's like doing negative uh you know a research on your political opponent or whatever you're doing it because you you want to win win the race or whatever but the information can also be true so i don't know that's why i said i don't know what to make of this like the motivations are clearly we're doing this to make money and we want the stock to go down and and the reasons we would pick ubiquity is the reasons you say but all that doesn't mean that the story is not true um and so that's that's why you would i would hope that i would see some other reporting about it if this is actually an issue if ubiquity is uniquely badly policing its distributor network or one of the worst in policing its distributor network or something like that but just no one else picked up this story no one else wanted to i've found no other reporting it doesn't mean doesn't exist it's just every single person who reported this to me i always said what's your source for this and they all pointed back to the hunter brook so that's why i say do with it what you will decide on your own uh whether you this changes your opinion of Yubikuti or not, but the story is out there and you can read it and it's from this particular source.

50:04Marco Arment:We are sponsored this episode by Factor. Now with Factor, these are meals made for busy people. I love Factor meals. This is a service that I have had. They sent me some and I kept it going because it's a great service. Here's how this works. Factor sends you meals, pre-made meals. You heat them up for like two minutes in the microwave or you put them in the oven and they're done. So it's not like a typical meal kit where you have to like be chopping stuff for what always seems like an hour somehow, even though it always says 20 minutes, I'm always, maybe I'm slow. Factor meals are great for all of that because when you're busy, you don't necessarily have time for that, but you still want like healthy, good food.

50:42Marco Arment:That's what Factor gives you. All their meals are ready in two minutes. They shop, they prep, they cook, and they deliver it all straight to your door. So you have more time for everything you love this spring and all the time. Factor has many meal options built around your goal. whether that's weight loss, overall nutrition, high protein, GLP-1 support, strength and workout recovery. They have all sorts of great chef-crafted meals. They're all functional ingredients, lean proteins, colorful veggies, whole foods, healthy fats. They don't use over 175 ingredients. They don't use artificial colors, sweeteners, high-fructose corn syrup, refined seed oils, just good, nutrient-dense food.

51:19Marco Arment:It's always fresh, never frozen, and they have over 100 rotating weekly meals, so there's always something new to look forward to. I strongly recommend you check out Factor. Head to factormeals.com slash ATP50off and use code ATP50off to get 50 % off and free daily greens per box with new subscriptions only while supplies last until 9-27-2026. See website for more details. Once again, factormeals.com slash ATP50off. Use code ATP50off. Thank you so much to Factor for sponsoring our show.

51:57John Siracusa:Without further ado, John, it's a sad day. Reading from 9to5Mac, Apple has confirmed to 9to5Mac that the Mac Pro is being discontinued. It is being removed from Apple's website as of Thursday, March 26th. Apple has also confirmed to 9to5Mac that it has no plans to offer future Mac Pro hardware. I'm sorry for your loss. I mean, it's nice that we just recorded an episode before this came out where I basically said, I can't believe they haven't killed it yet. What are they waiting for? Well, they were just waiting for us to drop the episode, apparently. I mean, a lot of people will write again. And when people write in and say, you know, I wonder what John thinks of this.

52:32I do wonder if they're like, you know, only a sporadic listener of ATP because, God, we've talked about this topic so much. And if you had listened to every episode and had not tuned out like Casey every time I mentioned the Mac Pro, I feel like you wouldn't. I feel like you would know how I feel about this and how I feel about it is what I said on the last episode. Like the machine has been effectively dead for a long time. like especially with the you know the the past few milestones that happened when it was like we were waiting for you know news and rumors or whatever and i think it was an episode a while back where i was like notice how there's no rumors at all about anything having to do with the mac pro and sure enough whenever the events came along there was in fact nothing to do with the mac pro and it's like well that's it it's over like they upgraded the studio to the m3 ultra which itself was weird did they upgrade the mac pro to it no they did not well maybe they'll do it the next event no they didn't it's like well that's it it's dead it's done like this is not what you do to a machine that's a going concern and then it became ridiculously embarrassing i made the snarky toot about uh the the neo being faster in single core than the mac pro because it's just so old and expensive and of course the price has never changed and you know it just needed to be put out of its misery apple you know apple was clearly no longer making this product and i would much rather see it put out of its misery and discontinued immediately rather than uh waiting for the mac studio to replace it and saying this new m5 ultra mac studio replaces the mac pro they're not even waiting until then why wait there's no point uh so it's gone and buried um if you want to hear me talk about the mac pro in the more positive sense not the m2 ultra one because it's hard to say anything positive about that one but uh the 2019 mac pro the one that i'm actually using i'll put a link in the notes to uh episode one two three of the last detail on relay with dan and tom and i was their guest and uh you're supposed to talk about like an item that means something to you design wise and i chose the 2019 mac pro and talked about it at length if you're wondering why i like this machine obviously it's from a design perspective but uh take a listen to that episode if you want to hear me uh think on uh better days when the mac pro 2019 mac pro was a really amazing machine when it was released that was in 2019 we're in 2026 now and the m2 ultra one was really never a very good machine um and people ask what are my plans obviously again if you listen to the show you know my plans my plans are i'm going to get uh an m something mac studio most likely assuming i can stomach the ram and storage costs which that's not the mac studio's fault that's ai's fault but you know it is what it is it's my fault for waiting too long presumably although i'm i'm glad i waited because not that i don't want an m4 max but i wanted to wait until the you know until all the possibilities had foreclosed and now they have everything everything has for this there's going to be no mac pro apple says there's not going to be one and by the way don't wait for a new one because that's not happening either they're really making sure you know that this is an x product uh unlike for example like i think gruber gave this example like the weird thing where they stop selling the big home pod and it's like well that's it the big home pod is dead and then a little bit later they introduced a new big home pod it's like what i thought this was a dead products it was just dead for a little while but apple wants you to know this is not a home pod situation this is a dead dead situation as in we have no plans for another one it's not like there's going to be a new mac pro next year's wwdc or something like that just forget it um yeah and

55:54Marco Arment:apple doesn't often tell you they're definitely not going to make another thing like i think the only one i can think of is when they a couple years ago gave a statement to some press saying that they were not going to make any more bigger iMacs remember that yeah yeah that was like don't don't wait for a 27 inch one when this 24 inch one is out like we're not going to happen um and and you know these statements are always like we have no plans um which means they currently have no plans in five years there could be plans but you know all they're saying is it's not going to be a home pod situation we're not discontinuing this because we ran out of stock before wwc that's not the situation um yeah and and again it's shouldn't be a surprise for anyone who's watching Apple listening to this program because the product was so incredibly dead.

56:36I mean, obviously, as soon as they didn't put the M3 Ultra in it to even continue the farce of it being a Mac Studio in a bigger case, they didn't even continue that. That was just, you know, so clearly this has been dead for a long time. The only thing I have to say retrospectively about it is, you know, again, a repeat of things I've said in the past, but people don't listen to every episode. The Mac Pro is essentially the only failure of Apple Silicon because the job of Apple Silicon was to replace you know intel processors with apple processors and all of apple's products and it did do that eventually to great effect in every single product to amazing effect in every product except one and it's the mac pro and it's not because apple had a difference of opinion and said well we don't need a big computer like the mac pro they planned to make a chip for the mac pro but they could not do it economically there was at least the m1 plan for like the m1 like big giant mega yeah the quad stuck together chip um there was diagrams of it there was things leaked from apple or whatever i'm not sure how far extended past that german says in 2022 they can their final plan for like a quad processor or whatever but it's a thing that they wanted to do but they could not do it and there's a million reasons why like you're not going to sell a lot of them it's very expensive yada yada like it makes sense but it is the one failure and lots of people are saying okay but that was a different era nobody needs expandable ram or internal storage or slots all those things are you know we don't need those anymore uh setting aside whether there's really zero use for slots i mean part of that is apple's decision to not support third party gpus but setting that aside especially in the ai age i would argue that potentially having a giant case stocked with a bunch of nvidia gpus for doing ai stuff would actually be a useful thing but apple doesn't get along with nvidia anyway um set the slots aside entirely just within the realm of is there any room in apple's product line for something that is not something more powerful than mac studio i've always said and i continue to say absolutely yes and here's why right now we have great examples of applications that run on macs that we wish would go faster that we wish had more capability.

58:50Local models, local AI models. Yeah, you can gang together a bunch of Mac studios, but basically local models will use every ounce of computation and memory you could possibly throw at them. The Mac Studio is great, but it's a plus-size Mac Mini. It's not that big. With the new SOIC MH chiplet architecture, where Apple can do separate dies and put them into things, they're now able to make chips that are bigger than they used to when they would take the M-whatever Macs, which was close to the reticle limit, close to the biggest chip that they could possibly make. and they would take two of those and combine them and call it an ultra and it's like well there's no point if the ultra fits in the studio there's no point in ever having anything other than the studio because that's the biggest chip we can make but now with the cpu and gpu at least being separate the cpu can be as big as m4 max or whatever the gpu can be as m4 max and you can combine two of those and now you've got a monster chip but guess what you cannot put that monster chip inside a Mac Studio because it can't dissipate the heat.

1:00:04So why does Apple still need something, whether it's a Mac Pro or whatever, something bigger than the Mac Studio? Because they're leaving computation on the table. They can make, it is possible with the current technology we've seen in the M5 Pro and Macs, to make a chip that is too hot to fit in the Mac Studio. Now, obviously, if they ship a chip in the Mac Studio, it's not going to be too hot to fit in the Mac Studio. But they're also not going to plan and design and build the chip that would be too hot, but they could you could make it bigger faster hotter why would you make it bigger and hotter because it will do more computation like it will be faster it will hot you could put more memory on it like you need a bigger case that can dissipate more heat just for the cpu gpu stuff in their current plans no slots no third-party gpus no expandable memory no none of that just basically something bigger than the mac studio and that is my argument maybe i'll write a post about it someday but probably not because i'm old and tired um that you know the case for a true mac pro successor they did they made it the 2019 mac pro was amazing it fulfilled everything that it needed to fulfill but that was a different era in the current era there is still a place for a mac that is more powerful than the mac studio because the mac studio it's got its own little thermal corner and it has amazing capacity and it can cool basically the biggest chip apple could make with like the m3 ultra but now i believe apple has clearly the capability to make a chip that is more powerful but too hot to fit in the mac studio and where are they going to put that that's why i think i will still be i will still be lobbying for and dreaming and hoping for that apple should make a mac that has better more cooling capacity than the mac studio so they can put a bigger chip in it that wouldn't fit in the mac studio i will also separately believe that the role of slots is still potentially good or whatever but that's that's nebulous but the the heat capacity thing is not nebulous that's and that's that's why the old m2 ultra mac pro was so ridiculous because we know that chip fits in the studio you can buy it in the studio and yet they're putting in this cavernous case with massive cooling it was just a complete waste so fingers crossed for let's say seven years from now we'll we'll report back here and see if apple has made a bigger Mac than the Mac Studio.

1:02:19The rumor we had was that there was no plans for a larger chip than the Ultra until at least the M7 generation, but that was a vague rumor and it was a long time ago, so I'm not holding my breath for that. But yeah, RIP Mac Pro, it was already dead in all but name. Now it is dead, dead. It is clarifying for purchasers. It is clarifying for me. And yet that still remains the one glaring failure. Not maybe glaring, glaring to me. The one glaring to me failure of Apple Silicon as they could not figure out how to make a replacement for the biggest, fastest, hottest chip that would require a big case to cool.

1:02:52They tried. They failed. It's not a big deal to anyone except for me, but I still think they should shoot for that star.

1:02:58John Siracusa:What is it that you think, you mentioned AI processing and stuff like that. Is there anything else that you can think of that would require just that much memory, that much compute? I mean, obviously gaming. I know people don't care about that, and Apple's gaming performance is amazing but gaming will also use all the transistors you can throw at it like you know well the frame rate is good but could it be better yes it can always you can always just especially gpus they're so embarrassingly parallel and everything that would be another applicant not that apple's going to make an eight thousand dollar computer to play games on or whatever but that's that's an easy one but but no the ai thing is just so such a no-brainer now because we are in such an early part of that curve where like the models that everyone wants to use literally can't run on anything that even on a desktop they can't run they only run on sixty thousand dollar nvidia rack mount things and probably multiple ones of those given how many tokens they use and that's that's a a place where we haven't been for so long where there's a thing we want to do in computing that cannot fit on your desk and ai is currently that it cannot fit on your desk i mean i guess it could if you you know spent a few hundred grand and put a rack there but like it doesn't fit in a normal desktop computer of any reasonable size and dimension and it's going to be a while before we catch up with that and why do we ever need to do that why wouldn't we just always run it on the server because local has advantages obviously like you know setting aside all the privacy and other stuff like that being able to do stuff locally is advantageous for multiple reasons which is why we never fully converted to thin clients where everything is in the data center so um yeah gaming and i guess ai and saying ai is like oh that's just one use case it's like a million use cases like yeah this is not a a narrow uh use case this is we don't even know what the full possibility space is of this technology

1:04:45Marco Arment:until it starts to level off a bit yeah that's like saying servers well servers can do a lot actually there's a lot of different yeah anyway yeah i i think the the mac pro you know for apple killing it now it's kind of like when they killed the ipod like not that many years ago like when by the time they killed the ipod it had really effectively been dead for a pretty long time And yet it was still a useful product for DJs, and you can't even say that about the Mac Pro. Yeah, well, but, you know, the Mac Pro, I'm sure the Mac Pro had, like, five or six customers. You know, I'm sure there were people who filled it up with the sound processing cards, like, you know, which was about the only reason to have it.

1:05:24Marco Arment:But if you look at the arc of the Mac Pro, I assert that the last good Mac Pro was 16 years ago, the 2010 Mac Pro. Oh, 2019 was amazing. What are you talking about? Here's why it wasn't. Okay, so let me go back a minute. This product from 2008, the first Mac Pro, which was basically the Intel version of the Power Mac G5, that was a great line. The Power Mac G5, it had some flaws, but it was a great line of computers. when the Intel transition happened the 2006, 2008, 2009, 2010 those were the four Mac Pro models that were unique there was the update in 2012 which is how we got our first favicon with the new badge over the Mac Pro because it wasn't really an update it was just like they just raised the minimum spec and added one more top end configuration option for the family that they already had been using for two years because Intel released one more little tiny model up top but it was not really a new Mac Pro, not really even a speed bump because the components were all the same, just a different selection of them.

1:06:32Marco Arment:But anyway, so you had four unique Mac Pros in that old cheese grater case. And it was every couple of years, Intel would release a new series of the Westmere Zeons. I guess that was one of the other. The new series of the workstation grade Zeons about about every 18 months. And that's when we get Mac pros. And they were great. They were extremely capable. They were very competitive with other options at the time. Like you could if you bought like an 06 or 08 Mac pro, like my first one was the 08 one, that 08 Mac pro not only outperformed everything else on the market for the for the Mac user in 08 in every way, include in single thread, it would at least match it and multi thread, it would destroy it and you know you'd have gpu advantages you would have huge amounts of potential ram and storage the ram was ecc and everything was more reliable like huge reasons to have the mac pro in those years and it was reasonably price competitive like it wasn't cheap a decent one might cost you like 2500 to 3500 back in those times and that was reasonably competitive with high-end desktops and high-end laptops of the time.

1:07:44Marco Arment:And then that would be relevant. That would be like a well-performing computer for most people, five years maybe? For John, 25 years? A little bit over 10. I used my 20. I think my 2008 Mac Pro was, I used for a little bit over 10 years. Sure. So those computers, at the time they existed, they were competitive. They were reasonably priced. They were extremely capable for a huge variety of what Mac users needed and wanted to do, and they remained competitive for years after they were out. Look at what happened after that. So 2012 wasn't a real update. 2013, the trash can. The trash can was almost none of those things.

1:08:31Marco Arment:Apple redesigned it to take away almost all of its advantages. It was no longer reasonably priced. It was no longer as capable because almost immediately after it was released, it lost single-threaded performance, or maybe even it already was. Multithreaded performance was good for a little while, except until it died from thermal problems with the GPUs. GPU performance was, again, okay for a little while, but it didn't remain competitive for very long. It was way less expandable and way more expensive. So they shrunk them by their own choices. The market for professional workstations was still a healthy market back then.

1:09:12Marco Arment:But by Apple's own choices with the trash can redesign, they shrunk its market. And then they also pushed the price way up. Then what happened? Then nothing happened for a while. The iMac Pro was good. That was kind of a one-off. We're not talking about that right now. The next Mac Pro was the 2019 one. Yes, six years later. So they neglected this market for six years. So during that time, there were no Macs with expansion slots. Like, go ahead, try to kill the market as best you can. Then in 2019, they released John's current Mac Pro, the Intel 2019 Mac Pro, the last Intel one. it started at what was it six thousand dollars john something like that yeah so they designed a computer that starts at six thousand dollars so you've already this was something that you know the the previous tower version started at like twenty seven hundred or three thousand dollars and you've now doubled the entry price skyrocketing this market way up like you are losing so many people from that all of the gpus you could buy were also pretty expensive so you know you're really pushing price here so you've now you're limiting it really to extreme high-end buyers only no hobbyists are going to buy it no power users are going to buy it like no one like john who mostly wants to use it regular ways but wants to play games with good gpus like none of those people are going to buy except john but like so they shrunk the market so far down less than a year before the Apple Silicon transition.

1:10:52Marco Arment:And of course, they would have known that when they released it. So what they did was through their own choices and their own, in some ways, negligence, in some ways, just mistakes, they kept shrinking the market for this product massively by their own choice. Now, what was also happening in that time was the market for professional desktops was itself shrinking, not just from Apple, but from everybody as more people were moving their work to laptops. And as Apple was losing a lot of that high-end computing market, anything that was using NVIDIA's CUDA, that all moved off of Apple platforms. A lot of high-end video editing was moving off of Apple platforms.

1:11:32Marco Arment:Apple was losing a lot of the high-end pro markets already for other reasons. Then Apple Silicon happens. And what Apple did was design this amazing architecture that serves pretty much everything below that need amazingly well. We have amazing phones all the way up to amazing workstation class laptops and the Mac Studio, which is like a pretty great overall, you know, asterisk here and there, but like pretty great overall computer. Except look at the performance of the Ultra chips versus the Max chips. It doesn't scale up that well. It scales up basically only in some of the GPU compute benchmarks, like the Metal and OpenCL.

1:12:17Marco Arment:Those scale up. Nothing else really does. And part of the Apple Silicon architecture also is you don't get upgradable RAM. So you have to buy it all up front. And of course, then that means you have to pay Apple's prices. So the RAM, super expensive, cannot be upgraded ever down the road. Neither can the SSDs in any of these products. so you have total lack of upgradability everything being expensive up front and limitations on the capacities of these things compared to where they were before when they were socketed so what apple did with the entire apple silicon transition was lop off the high end of the market in either performance price or limitations so and so that you have you know john's 2019 mac pro then less than a year later, the Apple Silicon transition begins.

1:13:08Marco Arment:Then the Mac Pro M2 Ultra generation, which, as John was saying, was just not useful. It was like, what is the market for this? If you cut off GPU expansion and it's still insanely expensive compared to everything else, no longer GPU expansion, no longer RAM expansion. They just kept cutting off markets, reducing the selling proposition, increasing the pricing on everything, on everybody trying to buy it for whatever they did need it for, just cutting off markets one by one. And now AI is happening and Apple is just nowhere in that business. And maybe they'll get there, but they are currently nowhere.

1:13:50Marco Arment:So by Apple's own failings and decisions and directions, they are doing great in every single other part of the Mac hardware lineup. The Mac hardware lineup is awesome, but the Mac Pro, they just kept cutting off limbs. And some of those were inevitable in the long run. A lot of them weren't. A lot of them were just bad execution or, you know, greedy pricing decisions from them. And like, and they just kept cutting off and cutting off and cutting off. So yeah, it's dead now. But the last good Mac Pro was the 2010 Mac Pro 16 years ago. i don't think everything you said supports that i the 2019 was the last good mac pro clearly i know it was a shame that it got apple siliconed but like when it was released there was no apple silicon so we can only you know even though apple knew that it was coming whatever at 29 2019 mac pro was released it was an amazing mac pro now granted the price was higher uh as you noted it was not as economical as it was but i think that's a as the 2008 one but i think that's just a sign of the times and that apple realized we have to move up market because that's where that that's where the people who are going to be able to afford this computer that we want to make exist and the other users will are better served by our lower end stuff you know at that point you could do final cut on an intel laptop still even though it wasn't as good as it is an apple silicon so they went up market with it which is fine you can go up market with a high-end computer uh but the 2019 mac pro is what they promised that 2017 mac roundtable so we're going to make a modular mac pro that's our biggest, baddest computer.

1:15:24How long will it be before Apple ever makes another computer with the capabilities of the 2019 Mac Pro? Just in terms of RAM, it could hold 1.5 terabytes of RAM. Granted, it was slow RAM, not like today's super fast RAM, but it was 2019. Massive RAM capacity. How long has it been, or was it, I'm not even sure, I'll have to look up the stats in this, before there was ever a Mac that could contain as much GPU processing power as the 2019 Mac Pro, which I believe could hold four AMD GPUs, two dual cards in it. It was it was so long before that was bested by any Apple Silicon things. I have to look up the stats to see how close it is to being competitive with the M3 Ultra, which is the current Mac GPU.

1:16:07Marco Arment:Yeah, wasn't it like this year? I think it was recent. I have to look up with it. But like the thing is like, well, what am I going to do with all those GPUs? What the hell is the point of that? I can't play games on it. There's no game that's going to run on four AMD GPUs. again what could i do with a box filled with gpus in today's age i can think of some things you can do with it that aren't gaming you know and their whole relationship with nvidia and how they messed that up as you noted as thing but like the 2019 mac pro when it was released was had capabilities that were so far above any other mac in terms of what it could do and what you could stuff inside it because remember back then you could put in uh you know graphics cards you could buy them from apple you can put in third-party ones you could expand the storage you could expand the ram all inside the big giant tower case like it fulfilled the promise of that again listen to the episode of the last detail here we uh talk about how i love the design of this thing and yeah it did go up market from the 2018 but 2018 was was in an era when power users would buy desktops and by 2019 we were no longer an era where power users were even interested in desktops it was more of like a home thing so who is going to be interested in an ultra powerful thing like this people for whom a$6 ,000 starting price is not a big deal right that's the market they were going after and the 2019 mac pro hit it and then we were just off into the wilderness and how long will it be before at this rate how long will it be before any mac can hold 1.5 terabytes of ram and again i know it's faster ram these days yada yada but like you know how long will it be before you can add as much gpu capability like relative to its time like in 2019 the amount of gpu you can fit in here scale that linearly with uh you know today's standards it what apple is doing with the gpus is amazing but it is non-expandable and you cannot like you know how many watts of heat of gpus could you put inside this if you have that many watts of heat in a mac today just think of the gpus the apple designed first party gpu you could put in there if you had a case that could dissipate it and a gpu that was designed to be in that case they could dissipate it so i will defend 2019 mac pro forever I love the 2008 Mac Pro.

1:18:16You know, I used it for a decade. Like, I love that computer. It was amazing, but it was from a different time. The 2019 Mac Pro fulfilled the promise that Apple made in the 2017 Mac Roundtable. It's just, you know, it got run over by Apple Silicon. Like, it is what it is. Like, I wouldn't want Apple Silicon to have been delayed for the honor of the 2019 Mac Pro. But they released what they said they were going to do. And they did a really, really good job of it. And then we're back in the wilderness. and speaking of the trash can by the way lots of people make this comparison in some ways it's unfair and in some ways it is entirely fair and shows how times have changed the mac studio is the trash can right that's it's it's got its own little thermal corner it's got no expansion it's all in one tiny little thing it doesn't fulfill the same needs as the 2008 mac pro uh but it was fine when it had a bigger sibling next to it or when apple was pretending they were still interested in that market but now they're basically saying like the mac studio is the highest end desktop we're ever going to need it's like saying the trash can is the highest end desktop we're ever going to need and the market said no to the trash can the market said apple we don't like your trash can it's bad you're not able to update it uh you need to do something about that at least apple's able to update the mac studio but i think the market will not put not push back against the mac studio the same way they did about the trash can because it's we're in a different time now so and as you noted apple has so thoroughly abandoned and screwed over the market that would buy a$7 ,000 expandable computer that it will be a long time before they're receptive to anything that Apple introduces.

1:19:43But I still say, essentially, a bigger Mac Studio, I'm not saying it has to be a tower, but like, for the people who would buy a Mac Studio, would you buy one that has two times the computation capability that is also twice as large? A lot of people would say yes, they would say, give me twice as much RAM, give me twice as much GPU, give me twice as much CPU, so I can run my local models that are twice as big, or whatever, whatever, you know, future computations we need. I don't think Apple has anything like that planned, but that's what I'm looking forward to sometime before I die.

1:20:12John Siracusa:You know, to build on something you said earlier, you know, what if you had the ability to be hotter? What if you had the ability to use more padding?

1:20:21Marco Arment:Don't we all want that, Casey?

1:20:23John Siracusa:Tell me about it. But I think there is an interesting thought exercise that we may or may not ever get to experience or see the end of, which is, you know, the A series of chips, I think we can all agree, were designed at the start to be power efficient. Like that was, if not the priority, it was certainly extremely high on the list. And to extrapolate from what John was saying earlier, what if there was a chip where perhaps power efficiency just wasn't, or if it was on the priority list at all, it was way, way, way further down. And what would that look like? And how incredible could that be?

1:21:01John Siracusa:Because these Apple chips, you know, the H-series chips, the M-series chips, they're incredible, incredibly fast, and incredibly power efficient. Imagine what it would be like if they took the gloves off and said, you know what, we're just going to use as much power as we need. I think that's the wrong way to look at it, though, because I think you still want to make the most power efficient thing possible, because think of how many more, how much more computation you can do within a given power envelope like that's the beauty of the ultra and even the max for that matter the the computation that is done in like the m whatever max chips cpu and gpu compared to the heat that is dissipated is amazing when you compare it to a standalone intel or amd cpu and a standalone amd b amd or nvidia gpu like the total amount of heat dissipated by like those discrete components versus the total amount that Apple's chips dissipate to do the same job.

1:21:59That's miraculous. That's why I'm saying you've done it, Apple. You have chips that use very little power and do amazing computation. You're able to put in its thing the size of the Mac Studio something that competes with stuff that requires a giant tower case. Imagine what you could do with that exact same check and just put twice as much in it and make it twice as hot. I don't want them to say throw power to the side. I want them to use their power efficiency to say, this means we can put more GPU cores in. This means we can add, because, I mean, maybe more CPU cores, maybe they've used too many, although Threadripper would say there's still room for more CPU cores.

1:22:34But there's always room for more GPU cores. Like, there's always room for that. And so take advantage of this amazing technology you have, especially with the chiplets. Because before you could say, well, that's all well and good, but they're already making the biggest possible chip they can because they do combine CPU. you know it's because it's an soc they're already at the limit but now they're not at the limit anymore at the very least the limit should have just doubled because they have they're doing separate dives for the cpu and gpu so fingers crossed for the future but i i agree with you casey that i think they could do great things but i think the reason they can do great things is because they're never going to ignore power efficiency and that's an advantage that lets you put more stuff in before you run out of like whatever the maximum wattage is for a tower computer yeah for your your outlet in your north american 15 amp circuit yeah and they're not even close to that like the studio dissipates so little heat like it's so tiny it does have fans but it's so tiny and so quiet and dissipates so little heat and it's like it's competing with like these giant pc towers with a thousand fans and 900 watt power supply or whatever yeah and keep in mind like you

1:23:38Marco Arment:know as like to kind of expand on that like you know efficient like being willing to be less efficient is not the limiting factor here. It is total power use. What gates the performance of modern processors is basically how much power are you allowed to consume and how much heat can you get away from the chip as quickly as possible so the chip can continue generating that much heat. And so efficiency always will lead to better performance. I think the bigger issue with Apple's limitation here is architectural. That, as I was saying earlier, like Apple Silicon, it scales up really well through the whole low end of the performance line and well into the mid-range and well into what most people would consider a high end.

1:24:26Marco Arment:But once you start trying to go past some of these upper limits, again, look at benchmarks of the M3 Ultra versus the M3 Max and certainly versus the M4 Max, which came out at the same time. and the ultra is not very competitive unless you have very specific workloads and even those even the gpu workloads that scale the best for the m3 ultra the m4 max comes very close so the the actual need for the m3 ultra is very very small the main reason to buy it is that it raises other limits like the ram limit although i don't know if you've looked recently at mac studio availability and shipping times yeah you can't even get the high ram ones and you know why by the way if you currently uh if you currently try to spec a 128 gig mac studio it says pick up wednesday august 26th oh my word yeah like that makes perfect sense because uh as we said like you know, Apple does component deals for locked in the price for a long period of time.

1:25:36And that's an advantage to them at the early stages of this terrible RAM crisis. But I can bet you they are not pre purchasing and locking in purchases for the whatever 128 gig RAM modules for the Mac Studio. That's more of a like, if anyone orders this, we'll get it built. Well, guess what? Now they can't get that stuff because like they don't keep they don't you know, when they're selling like the laptops or whatever they're like yeah we're going to pre-order x number of million of these because we know that's how many you're going to sell how many max ram mac studios do they even sell so like ah when those when those orders come in we'll just we'll just buy them at the market price at the time oops now the market price is insane maybe you can't even get those things so

1:26:15Marco Arment:yeah yeah because you can get like the 128 gig uh macbook pro the new macbook pros you can get those like next week yeah because they probably pre-ordered and have a deal for those but not for the next studio yeah yeah but like when you look at the scaling up of apple silicon like the way they've designed the architecture with unified memory and everything being like part of why it's so fast there's a lot of integration here a lot of the controllers are on the die with the processor like there's there's a lot of integration of all the different components that if you look at modern pc architecture there's a lot less of that like some things have moved onto the processor that used to be separate chips like there used to be something called a Northbridge and a Southbridge.

1:26:54Marco Arment:I haven't been paying too much attention to PC architecture recently. On the PC side, they're starting to look a lot more like Apple Silicon these days. Yeah, stuff does, over time, more stuff from what used to be called the Northbridge moved into the processor and stuff like that. But as you scale up the performance of these architectures, integration gets you a lot of wins. But the problem is, the more stuff you integrate into the main SOC or the main CPU die, the less expansion you have to broaden with some of the horizons here unless you make a whole bunch of different parts, a whole bunch of different processors that have a whole bunch of different controllers integrated into them.

1:27:33Marco Arment:And so, like, you will see that on high-end data center CPUs. You're not going to see it from Apple. You're not going to see them make 12 different types of chips. You know, they might have, like, variants here and there. Anyway, so, the Apple Silicon architecture is great for the bottom 90 % of what people need. But for that top 10 % of the market that needs like, again, what if you need the 1.5 terabytes of RAM that the old Mac Pro supported? You still, even before this current RAM shortage, the Mac Studio maxed out at 512. A third of it. And that's years after the 2019 Mac Pro. 2019 Mac, that was, you know, again, that's seven years ago.

1:28:13Marco Arment:And until a few weeks ago, the highest you could get on a Mac Studio was a third of its RAM ceiling. and that's because the architecture is different like because to put that much ram in apple silicon you need you need a whole bunch of the apple silicon chips fused together and because they have all the controllers built in like there's all the level of integration they have it just it's again it's different architectural decisions it is fantastic for all the computers that all the rest of us use but it it precludes them from offering these very high configurations and i don't think the architecture scales well.

1:28:48Marco Arment:So I don't even think that if they did the quad max chip, I don't even think that would be competitive. I'm guessing the reason why they didn't end up shipping that was probably partly because it would cost a lot. But I bet also, I bet it just wasn't worth it for enough people. I bet they learned that when they scaled things up, it didn't scale up enough to be worth that cost to any buyers. I bet that's the real reason. i don't think so you're saying like oh the apple silicon architecture can't scale up or whatever everything you cited for the problems with scaling linearly and stuff that's all to do with the fusion architecture where they take two maxes and stick them together end and right and that is their problem with scaling the first time they did that they screwed something up with the gpu and that didn't even scale linearly and then they fixed that but like that architecture of saying we don't want to build we either don't want to or can't build a chip this big on a single die because we're at the reticle limit or whatever so we're going to take two of these other things to your point, two of these other chips that we already have to build for our laptops.

1:29:46And we're not going to do a whole separate chip, but we're just going to take two of the laptop chips. And we also designed it in a way for them to talk to each other with this, you know, fusion architecture where they stick to each other end to end, just like our old shirt says with the two maxes flipped over and stuck together. That architecture does not scale well, but that's not Apple Silicon. That's whatever it is, the Silicon Interposer fusion architecture thing. i think we're beyond that now as evidenced by the chiplet architecture on the m5 pro and m5 max and so i'm not saying what apple is going to do but if they wanted to they could totally abandon the silicon interposer to max's stuck together architecture and just build a big dedicated cpu on its own die with no gpus and a big dedicated gpu on its own die with no cpus and make them both at the reticle limit and stick them on a chip and there's your m5 ultra i would love for them to do that i don't think they're going to but i would love for them to do that and that is still apple silicon so there's nothing inherent in apple silicon that says you have to do that end-to-end thing and that is the linear scaling barrier that is the thing that is cost cost tons of money but doesn't give you the bang for your buck and so i think the you know the runway is clear for them to do something better especially like two nanometer or whatever they end up going to and like the m6 or m6 i figure when that's going to happen like the next process shrink with the chiplet architecture frees them up to leave behind the ultra the what has been the ultra architecture which is taking two maxes and sticking them together because i think that is a cost savings and it's been cool but it has essentially run its course it was never as good as it could have been but it was a reasonable compromise to get something out better and by the way one of the other big advantages of the ultra has what you see in some of these reviews is and we don't talk about that much because it's not in the world we're in but it has twice as many media units and lots of people who do like these uh you know benchmarks of like oh i'm exporting from final cut and like huh no matter which m chip i use it exports from final cut at like the same speed or like the pro versus the max why is it the same speed because that's all happening on the stupid media engine and the only way to get it to go faster is get the ultra because it's got double the amount of that stuff because it's two maxes stuck together and lo and behold it goes twice as fast again if you built a dedicated high-end chip maybe you'd put four media engines you know what i mean like you'd make different decisions but to your point it's like are they ever going to is it ever going to be economically feasible to make anything custom for the mac studio or is even the mac studio stuck with you just get the leftovers from what the laptops use and maybe we can stick two maxes together and you should be happy with what you get happy with your non-linear scaling whatever it's better than a max it is uh and people who pay for it if you want something better than a max here's our ultra but that's not the promised land of this you know of performance and so i do hope especially with the mac pro gone for good now that they have some better plans for the studio because i do and i you know the silicon interproser thing with the two maxes end end that's also not particularly power efficient and there's a bunch of stuff we talked about there's a bunch of stuff on those chips that is you don't need twice as many of those but you get them because it's two maxes right it would be better if you could consolidate the parts that are in common and yada yada so that's what i'm hoping for and i don't think it's a limitation on apple silicon is just a limitation of the cost-saving approach they have been using for the Ultra since the M1 Ultra.

1:33:06And I agree with you that the quad one that they abandoned, they abandoned it for a reason, not just for the hell of it. They wanted to do it. They tried to do it. They had visions of putting it together. But if the scaling on two of them was bad, the scaling on four would be even worse and the cost would be even higher. And yes, it would have been faster than an Ultra, but it wouldn't have been twice as fast as an Ultra. And so it's a point of diminishing returns. So basically they couldn't pull it off. They couldn't deliver what the 2019 Mac Pro delivered, which is something that has a big multiple of the capability of its contemporary Macs.

1:33:36They just couldn't do it in a feasible way. And so far, they haven't been able to. So fingers crossed that the M5 Ultra is, if it's not, if it's still two M5 Macs are stuck together, that'll still be the most powerful chip that Apple makes. And it'll still be really good. And we'll check the benchmarks and see, is the scaling better than it was, even if it's still not linear? But, you know, I need something to hope for. So I'm here, I'm out here hoping that the M5 Ultra will not simply be two M5 Max stuck together, but I would not put money on it.

1:34:05John Siracusa:So when do you think you're buying a new computer then? It depends on how much the 8 terabyte SSD costs. As soon as they introduce the, I'm going to try to get like a friends and family discount from Apple, from folks that I know who work there. And sometimes the friends and family discount is not available immediately, like you can't pre-order with it or whatever. So it might be a while after the M5 whatever Mac Studios come out. I have to wait around until they show up with the friends and family discount. But when that happens, I will buy one. And will I get the Macs, which would be fine for me?

1:34:38Will I get the Ultra if such a thing exists? We'll talk about that on the show, I'm sure. But that's the plan. M5 something Mac Studio appearing on my desk whenever I can order it. and assuming I can afford the 8 terabyte SSD because that's honestly, I mean, look, it's not a coincidence that I made an app that saves disk space. I am pressing up real hard against the 4 terabyte limit on my SSD. I use my app a lot, like for real. I use it to save disk space. I need more disk space. And the next size up that Apple offers is 8, and it's going to cost so much money. Yes, it is.

1:35:17John Siracusa:Can confirm. Been there. That's what I'm planning for. And, you know, I think months and months ago, I've said I'm planning on getting an M something Mac Studio because I had no faith that the Mac Pro was going to come out. And lo and behold, that is the case. What do you expect to miss about the physical parts of the Mac Pro? Like, because you have like your time machine drive internal to the Mac Pro, don't you? I sure do. Yeah, I have my super duper clone end. Like internal storage is for sure the thing I'm going to do, which sounds silly because my internal storage is incredibly slow. Like it's not, it's like connected to SATA for crying out loud, but they are SSDs.

1:35:49and the the trouble-free nature of internal storage and internal buses i'm gonna miss that the most because it's not like i'm getting the amazing performance out of it it's not like i can't but just it's all inside the box i don't have any wires hanging out they never like unmount weirdly or do weird stuff they don't have their own power supplies my i have an internal time machine ssd and i have an internal super duper clone which and they're ssds but and they're slow but having that inside the case is something that i will miss and i you know that's that's a niche market or whatever but i really do wish there were better external storage options because when i get my mac studio i'm going to want to have uh an ssd time machine drive and i'm going to want to have an ssd uh super duper bootable clone drive and you know i can go try to buy those and buy a Thunderbolt enclosure or something and put a NVMe.

1:36:43It's a whole bunch of, you know, computer nerd hobbyist stuff that I'm actually not that interested in having, you know, bought bare spinning hard drives and put them into enclosures since the days of like the, you know, classic Mac. Like I've done that a lot. I would just prefer to have reliable, trouble-free storage that I can use for this. I mean, hell, I'll probably, oh, I can't. I'll probably take those SSDs. out of this mac pro and use it but i can't because if i have an eight terabyte ssd it's not going to fit on my four terabyte backup drive but oh that's going to cost me even more money after i buy apple's eight terabyte prices then i have to buy then i have to pay actual market prices for non-apple storage to get my uh you know time machine drive my clone drive i might even have to buy new spinning discs for my synology to fit the backup so this is going to be an expensive year and i may have to space i may have to space this out a bit uh we'll see how it goes john you spent 11 billion

1:37:37John Siracusa:billion dollars on the mac that you're using to speak to us right now seven years ago well i'm just saying given that you spent as much as a freaking civic on that and half of that price was the monitor that i'm still going to use all right well all snark aside i am sorry for you i'm glad for you that you have an answer because it is as you mentioned apple style just kind of let this thing fade into the night as it had been doing but i'm glad you have an answer but i'm sorry it's not the answer you wanted i already have the answer like let's like i said the very previous episode, we had the answers.

1:38:08No one was confused about what was going on with the Mac Pro. It was just a question of what are they waiting for, and they were waiting for us to post that episode.

1:38:17John Siracusa:We are sponsored this week by Claude. Let me tell you a true story. A couple of weeks ago, I had a moment that I didn't have locked in for something else, and I thought, you know, I really wish that CallSheet had better support for intents, better support for shortcuts actions, and I wanted to be able to give call sheet or an intent within call sheet, like an ID and a media type, you know, TV or person or whatever, and have it emit information about that movie or show or person. And I have bounced off this API like five times. I just couldn't make heads or tails of it. And so I thought, you know what, let me try Claude Code and see if Claude Code can do it.

1:38:54John Siracusa:And so I set Claude Code off with a very, very basic spec and asked it, hey, can you make this happen? And it and then I was mad with power. And now I've got a whole bunch of intents in beta, all thanks to Cloud Code. So, Cloud is the AI for minds that don't stop at good enough. It's the collaborator that actually understands your entire workflow and thinks with you, whether you're debugging code at midnight or strategizing your next business move. Cloud extends your thinking to tackle the problems that matter. Let me give you another example. I also asked Cloud to generate kind of a plan for me to use Apple search ads.

1:39:26John Siracusa:I haven't done that yet. Well, I have now, thanks to Cloud. but I hadn't done it at the time. And it put together a little information packet and not worksheet, but kind of worksheet of what I should do and what I should look into. And it did it very quickly and very easily. I also, literally an hour ago, had it write a shell script for me to take all the chapters in a video, extract the songs, basically, extract the audio from all these chapters and dump them on disk. It's incredibly, incredibly impressive what Claude and also Claude Code can do. Things like coding, things like developing whole new things, things like doing audits.

1:40:00John Siracusa:It'll do everything. It really, truly is incredible. So if you're ready to tackle bigger problems, get started with Claude today at claude.ai.atp. That's C-L-A-U-D-E.ai.atp. And check out Claude Pro, which includes access to all the features mentioned in today's episode. Claude.ai.atp. Thank you to Claude for sponsoring the show.

1:40:22John Siracusa:All right, let's do some Ask ATP. And Alex Kent writes, You occasionally remind me that running iOS apps on macOS is a thing. I was excited about this capability before it shipped. Years later, I find zero iOS apps that I regularly run on macOS. I can partly blame the allow this app to run on macOS checkbox and App Store Connect that apparently most corporate developers disable out of habit. But I'm also not sure that the experience of running iOS apps on macOS beats using the corresponding web app or picking up my phone. What iOS apps do you find useful to run on macOS? You know, I echo a lot of what Alex is saying here.

1:40:51John Siracusa:I was pretty excited about the thought of doing such things when it was a, you know, the sparkle in Apple's eye. But in reality, I almost never use this. And for a lot of things, I would reach for iPhone mirroring before I would reach for, you know, installing an app on my Mac. That being said, the only thing I can think of offhand is channels, which does have a web app, but it's mostly about management. It's not really about consumption. and if there was any weak part of channels which is a former sponsor and a friend of mine is one of the co-founders but anyways if there's a weak spot about channels it's that the web user interface particularly for the purposes of consumption is really not great and so I've installed the iPad app on my Mac and use it every great once in a while I don't think there's any easy way for me to tell is there what other apps I have installed so I don't even know I can't think of any others off the top of my head

1:41:48Marco Arment:I mean, you might have Overcast installed.

1:41:51John Siracusa:Oh, I did use that all the time. And then I mostly just used my phone, either with the phone speakers, godawful as that is, or occasionally via iPhone mirroring. That's a good point.

1:42:00Marco Arment:This is one of those ideas that when Apple Silicon first came around and this was one of the features of it, I thought this is going to be really useful. And yeah, I think the combination of so many apps opting out of it and also like if you actually use iOS apps on Mac it's not that useful it's not that great if it patches a need that otherwise you would not be able to serve so for instance my Overcast app runs that way I personally don't really use it because I don't listen to podcasts on my Mac but more people than I expected use it significantly more than I expected it. I never would have guessed it would have the users that it has.

1:42:48Marco Arment:But it's still not like the majority of the user base or not even close. Like, you know, most people don't use it, but it does have some utility. But if you actually like try to use iOS apps on the Mac, they don't work that well. It's kind of a weird experience. Lots of things are just worse about them or clunky or don't fit the platform or don't quite work the way you expect them to work. Because when you design apps for iOS, you're designing for touch on a totally different platform. It just, it doesn't translate as directly and as well as I think people want it to. I don't even use like the iPhone mirroring thing that they added last year.

1:43:28Marco Arment:I don't even use that because I find that to be too clunky and I'd rather just pick up my phone out of my pocket and, you know, use it that way. these are features that to me like i will use in a pinch but i don't have that many pinches yeah i don't run any ipad apps on my mac because i can't run ipad oh john uh yeah so i mean obviously i do have apple silicon uh max here um and i i think the only one i think the only ipad app i have knowingly installed is overcast and that was actually to work on uh to correctly file apple's bugs because they had a bunch of bugs related to the apis i was using in switch class when it came to um ipad apps like if you asked to get the icon for an ipad app their api would vend you uh a dimmed icon with like the circle with a line through it buster over it which is not great it's not really what you want uh and they fixed that bug thankfully but it was it was a bug like that for like a year and a half and i had to work around it or whatever there was also some weird stuff with like the way the notification thing that switch glass uses to say like when an app is launched tell me and when an app exits told me because of whatever weird way that ipad apps are handled on mac os the notifications for those apps launching and quitting were different and caught like the timings were different and different events came at different things so i had to work around that so that's why i have overcast installed is essentially to debug my app switcher app when it comes to uh ipad apps if i had an apple silicon mac which i will eventually um i and we were back in time one of the things i would have looked forward to using on it is let's say status board by panic a discontinued app that i have since vibe coded replacement of on the web so i don't need that anymore but i would absolutely use that and that's a perfect use case because as marco said like ios apps or ipad apps on the mac talk about not behaving like a mac app they just absolutely do not behave like a map because they're not they're not it makes sense like but they don't but and they also don't behave like an ios app like if you if you're like an ios developer and you've used the ios simulator you might think the mac apps would behave like they do in the simulator they don't they don't even let you do the option thing to use a little virtual pinchy yeah it's it's even because of course like that would be a little bit weird too so it's even different from the simulator there are so many little things that are different and the supporting apis and the way they behave is also different so you have things like background downloads and background refresh and push notifications and services different kind of extensions lots of like airplay lots of stuff that either doesn't work on the mac at all or works very differently context menus another great example pull to refresh like there are so many things on the mac that just don't work the right way as they do in ios or are absent or don't work at all or have bugs or crash if you try to use them yeah and apple could do things to make those apis better so that they will behave better when they're running on the mac but like things like status board are ideal because it's essentially like a big widget it's mostly read only it's an app that didn't exist on the mac at all and there used to be no web version of it before i had my thing slapped together um and so i would i absolutely would run status board on my mac as an ipad app and just had it up there as a little dashboard of stuff just like the old just like i used to run dashboard but that's the type of app that i would do like if there was some kind of app if there was a social media app that didn't have a native mac version maybe i would use that but i vastly prefer to just have like you know ivory on mac as opposed to you know iphone mirroring ivory on my phone or something like that but like that's that's where i'd be using it like you said it's gap filling this isn't available on the mac and using it uh using the ipad version you know or the phone version mirrored is better than nothing and there are apps where that is true a status board totally would have been the case before i had the web version so yeah and developers not opting into it it's like you know there's a million incentives for them not to do it and apple didn't really make it attractive for them to do it like there was there were plenty of sticks and not a lot of carrots for doing that and so i don't really blame developers for not checking that box but you know it is what it is and this is the argument for like you know these are separate platforms with separate strengths and being able to use them use any app anywhere is convenient but it doesn't suddenly make the platforms the same they're just not like fundamentally one is a touch-based thing that you carry around with you and one uses a much more precise pointer on bigger screens and yeah like those differences will never go away if there was a grand unification and there was one apple os those platform differences would still exist and that one apple os would have to handle those platform differences one apple os can't be like, oh, it's a big phone everywhere.

1:48:13That's not going to work.

1:48:15Marco Arment:My reminder app that I occasionally remember to work on that I'm using. You should put a reminder for that. Yeah, right. I use it constantly. But, you know, so anyway, so I'm having it. I'm having this is my VibeCode of reminder app. I'm having it be cross compiled with Swift UI. It compiles on Mac and it compiles on iOS. And many of the things that I have had the AI do for me is customize the behavior between the two and have it behave differently and look differently and have different controls between the Mac version and the iOS version because they're just different. One of the things I added the other day was like a swipe to snooze or swipe the other way to delete on iOS on the row of the reminder.

1:49:01Marco Arment:Standard iOS swipe actions. And on the Mac, of course, it dutifully added that on the Mac where it makes no sense because you never want to swipe your mouse over a table row on the Mac and have it swipe to delete accidentally. You never want that on the Mac. That's the wrong thing on the Mac, and it's easy to accidentally do and hard to intentionally do.

1:49:22John Siracusa:Tell me that you are a mouse user without telling me you're a mouse user because as a trackpad or touchpad, whatever it's called, user, I don't have any problems with doing swiping on the Mac.

1:49:33Marco Arment:Well, lucky you. But anyway, on the Mac, that is not a useful or that is not like the standard correct metaphor for actions on the Mac. The standard metaphor for quick actions on an item is usually a context menu, keyboard shortcuts or visible controls. And so I'm having different like on the Mac. I have the right click menu to do those same actions or, you know, at some point I might add keyboard controls like you develop software and you develop interfaces differently for these two platforms. because they work differently. So anything that tries to run the same code, the same app with the same interface or almost the same interface on both of those platforms, it's going to result in mediocrity.

1:50:18Marco Arment:And that's what this is. It's a well-intentioned system that doesn't work very well and therefore isn't that compelling most of the time. All right.

1:50:28John Siracusa:And then finally tonight, Michael Bresher Jr. writes, if you were in charge of the Mac and iPhone software teams, how would you all balance new features needed to keep up with the industry with keeping design native to people who have used the Mac for over 30 years like John? Also, would you all move to a one and a half or two year cycle for the OSs? This is a tough thing, right? And this is what makes engineering so fun and so challenging is figuring out the right balance. And I don't know, I feel like for me, if things are new, that doesn't immediately bother me. So as a silly example of this, Volvo just released for Aaron's car, a new version of the infotainment software that really aggressively rejiggers where everything is on screen.

1:51:14And in some ways,

1:51:16John Siracusa:I think that the rearranged deck chairs are an improvement, right? And it's for the best. But I can tell you for Aaron, it's been very frustrating because she, as the primary driver of that car will mash on a particular portion of the giant touchscreen in the center dash in order to see her like backup camera, for example. And now that button somewhere else. And that's frustrating for her. And I get that. Liquid glass pissed off most of the community. And I don't, I mostly don't agree with that. I mostly think liquid glass is actually pretty good. But it certainly made everyone upset and it definitely moved things around.

1:51:52And so I don't

1:51:54John Siracusa:necessarily think new is bad. And I'm kind of of the opinion, like actually a great example of this, God help me, I'm going to get so much email, please email someone else.

1:52:02Marco Arment:But the icons,

1:52:04John Siracusa:the icons in the menu bar, like I don't think they're particularly additive, but I am not deeply morally offended by them. Like it seems most of our peers are. That doesn't make me right for the record. I'm just telling you my personal opinions about it. And so I, I don't think that new is necessarily bad, where it becomes frustrating to me, and this is not a particularly hot take, is new for the sake of new. And I think Michael is touching on that. New features needed to keep up with the industry. Well, a great example of that is let's AI all the things. And I think there's appropriate times for AI and there's inappropriate times for AI.

1:52:44John Siracusa:And do I really need what? I don't even remember what the stupid thing is called. It was like something playground. what's the image playground image playground i couldn't even remember the darn thing because i've used it like twice but it's here but it's here baby i cannot i have dismissed image playground is here 48 times yep exactly um so you see where i'm going with this right is that i i don't think adding just because you're keeping up with the joneses is necessarily the right idea in apple in the past had usually been very good about yes that's wonderful for you but we don't need that It's not additive for our users.

1:53:17John Siracusa:It's not helpful for our users. So I'm kind of meandering around the question rather than answering it because I think every decision would be different. It would have to be on a case-by-case basis. But I think generally Apple's thrust to move forward isn't by itself unhealthy. I think the problem is that sometimes there's a little too much new shiny and not enough improvement of what was already there. And I think what makes it even more aggressive or worse is when the new shiny is just for the sake of new shiny rather than actually additive. But I don't know. Maybe that's a non-answer.

1:53:55Marco Arment:So I think the big challenge here is if we're in charge of the Mac and iPhone software teams and we're trying to figure out keeping up new features to keep up with the industry, are we talking about 10 years ago, 5 years ago, or now? because right now the hot new stuff admittedly you're right casey like keeping up with with hot new stuff is it can be wasteful and it can be misguided um but right now the hot new thing is ai integration and ai based features and that is no matter what you think of ai uh there's there There are three things that are real. Number one, it's not going away. Number two, it is not as bad as you think it is.

1:54:42Marco Arment:And number three, it is not as good as you think it is. It is a really, really big deal, despite all of those dynamics going on. It's here. It's going to define and direct a huge amount of what we use computers for, what they can do, what we can do. It's a big deal. It's a very disruptive and promising area of technology. And so right now, trying to figure out like, hey, how do we how do we like add new features to like the finder or something like that's that is something that people should occasionally be considering. But like if you're the head of Mac and iPhone software, that's that kind of thing is the last thing you should be considering.

1:55:32Marco Arment:Whether we put icons in the menu items, that's a terrible design decision to have them there, especially the choices they've made, etc. But that's not a good use of anyone's focus right now. There's huge shifts happening in the tech industry. There will be huge shifts happening for the next number of years. No matter what you think of AI, it is going to be as disruptive as things like the internet, mobile. It is on that level of the amount of disruption that this is going to cause to our business and many other businesses. When the internet first came out, if a leading computer company maybe spent a lot of time doing a lot of other features and things that didn't matter as much and maybe missed the internet, that could be a problem.

1:56:25Marco Arment:They might have to like really scramble and catch up later. Let's say maybe that same company, maybe a little while later, missed mobile. And mobile was exploding. The smartphone revolution was exploding and they just weren't in that game and stumbled around for a while and kind of never got in that game and it reshaped their business forever and they missed out on that entire thing. That's Apple right now with AI. Apple has missed the main part of it so far. They're scrambling. We'll see where they get. But right now, that needs to be their focus. And we can debate all we want about the crappy Tahoe redesign they did to macOS, which they did.

1:57:06Marco Arment:And it needs attention. But if you're the head of software, that kind of thing is not what you should be paying a lot of attention to right now. Because the entire industry has this massive earthquake shaking it all up. and you're worried about whether your corner radius should be smaller or not. No, that's not your job. Delegate that to the designers. Hopefully they have better leadership now and they have better heads on their shoulders and hopefully they'll fix this problem. But your job as the head of software needs to be, what do we do to embrace AI, to make good features with this, to keep ourselves competitive and to move our platforms forward in the actual real modern place that we find ourselves and hopefully can catch up soon.

1:57:51I don't think I would concentrate quite that much on the AI stuff, especially considering how badly Apple is doing there. But yeah, this is one of those things where it's like it would have been better if 10 years ago they had struck a better balance between, you know, new features and keeping up with industry trends because that would put you in a better position. And I would argue that paying attention to the plumbing, especially on the Mac, but even on the iPhone or whatever, like really just, you know, polishing up the basic functionality, release after release is super essential in a world where you're going to drop in a bunch of agentic ai because the cleaner you can provide interfaces to your functionality like that you have stuff that it works it's proven it's reliable putting you know allowing ais to start controlling that or having access to it like you know for example what apple has done with security like they weren't doing it for ai but having a good security architecture allows them you know potentially because we don't know what they're actually going to do but in theory puts apple in a good position to have ai features integrated with the phone in a way that is privacy preserving let's say they didn't do anything with privacy and then ai came along and they said oh well we should really just concentrate on ai now it's like no you should have been concentrating over the past 20 years on privacy stuff because if you want this you know unreliable lm things to control any aspect of the phone you need to have a good privacy architecture in place and they do thankfully right and that's that same philosophy i think is what they should be doing over all their platforms which is it behooves you to maintain and polish and improve the basic architecture because whatever the next thing is it will be easier to integrate it into your stack if you're not building on a house of cards if you didn't neglect privacy for 20 years if you didn't let's say neglect the finder for the entire history of mac os 10 it would be way easier to hook up an into that if it wasn't like this terrible fragile house of cards it's buggy nobody likes anymore and so you know you can't go back in time and fix that but like to answer michael's question like the balance between like new features and keeping up with the industry every one of my like apple report cards have ever done i've said that apple is not correct not striking the right balance between essentially maintenance stuff like you know taking existing features and eliminating bugs and improving the design and new features that continues to be the case it is exacerbated by what marco said which is like oh guess what there's a bunch of really new stuff that you should get on right now it's a lot easier to strike that balance when you're kind of in a quiet period when you're not that's when you're you know all those chickens come home to roots like oh we totally should have been keeping mac you know more ship shape for all those years because now we're going to try to hook up this agent stuff to it and it's just it's you know it's terrible setting aside that's probably not what app's problem is their problem seems to be just with the basic ai stuff or whatever but uh yeah i would i would you know this is the type of where like if you keep making a mistake in one direction you keep uh overcooking something just try to undercook it just try to undercook it once like go in the other direction because people like correct for it they're like oh we're doing too far to a we need to go to b and they go like slightly less far to a and it's like can you make them can you at least make a mistake in the other direction uh and you know the mistake in the other direction being like you know snow leopard no new features or whatever let's actually do this as a marketing push and that was kind of a one-time thing so i think they need to massively correct currently in the direction of shoring up existing functionality which is exactly the opposite of what marka was saying it's like you just got to ignore stuff because you're behind that ai but like i think those two things are not mutually exclusive and that gets to the second point here which is would you change to a one-time-five or two-year cycle or whatever i have said for years and continue to say based on my 25 years experience in the software industry and releasing products on schedules and deadlines that it is absolutely absolutely possible to have annual releases by smartly choosing what goes in those releases or not.

2:01:34You can stagger them. You can have two-year plans, three-year plans. It's not like, oh, you can only do things that fit in one year. No, this is a scheduling possibility. You can have things take three years to come to fruition and still release on a yearly schedule. It's just a question of being smart about picking what's in it and what's not. Apple's problem is not the annual schedule. Apple's problem is their inability to correctly pick stuff that will fit in a year. that's there's always been their problem that is you know that's most organizations problem like i don't know how many times i've had this exact discussion in my jobby jobs about like deadlines and schedules and what can fit and resources available or whatever like it's like casey was saying it's a trade-off it's like we have this many people we have this much money we have this much time we have these people and who have these skills like it's not you know everyone's not just fungible right you have this set of people and this set of skills and this set of resources in this time or whatever, what can you fit?

2:02:25Can you get X by this date? Can you get Y by that date? If you have to have something out by this date, what can you fit in? Don't be ambitious and say, we're going to be done by this date and we're going to have all the stuff. You know, like you're not going to have all the stuff. So you have to start to cut things. And Apple has been really, really bad. And I feel like worse over the past several years at figuring out what can we actually fit into an annual release and instead they just like they overstuff it then they trickle stuff out over the course of the next year which bleeds into which pulls resources and people who should be working on the next year's release they're too busy releasing the rest of this year's release which trickles out over the year that's their problem their problem is one of you know it's not just them it's every everyone has difficulty of like estimating software projects and figuring out what will fit it within a schedule so i don't think it's the annualness of it if you did a two-year they would do the exact same mistake they would just do it over two years and it would be an even bigger disaster or three years or four years so that's the challenge i would not necessarily like i think mac os could use a two-year cycle but it wouldn't remove the problem because their problem is if given a finite period of time choose features that you'll be able to be done with by the by the end of that cycle and apple is not very good at that and and nobody is it's a difficult problem but it is frustrating to see apple continue to have difficulty getting that done and like i said if you know the only time they didn't have this problem was when they didn't have any regular cadence whatsoever which generally is a sign of an unhealthy organization if you're like well we can't build to a deadline it's just done when it's done oh nice you know nice work if you can get it but that's not usually not the real world usually you can't just have that luxury in the early mac os 10 days it was like we're working on another version of mac os but we're doing a lot of other things too and it'll be done when it's done and Apple was smaller then and people were paying less attention and it wasn't that big of a deal.

2:04:16And the phone pulled resources off the Mac and macOS releases would take 18 months, 16 months, 13 months, like whatever, whenever it's done, whenever it's ready, no big deal. But that is not the sign of an efficient, effective organization. We can't build to a schedule. We just do stuff and when we finish it, we release it. That is a little bit too touchy-feely for whatever trillion dollar corporation. so I think it behooves Apple to continue to try to get better at actually hitting deadlines with a set of realistic features that you can finish not just barely finish so it barely works but finish to a degree of quality that you would be proud of within the allotted time and being honest with yourself about what you can actually do if only

2:05:01Marco Arment:thanks to our sponsors this episode Masterclass, Factor, and Claude and thanks to our members who support us directly, you can join us at atp.fm slash join. One of the many perks of ATP membership is our weekly bonus topic, ATP Overtime. This week on Overtime, we're going to be talking about, we have an update to the egg. What is happening with Johnny Ives' OpenAI product or products? Updates to that in Overtime. Join to listen, atp.fm slash join. Thank you so much, everybody, and we'll talk to you next week. Now the show is over. They didn't even mean to begin. Cause it was accidental. Oh, it was accidental.

2:05:45Marco Arment:John didn't do any research. Marco and Casey wouldn't let him. Cause it was accidental. Oh, it was accidental. And you can find the show notes at ATP.fm. And if you're into Mastodon, you can follow them at C-A-S-E-Y-L-I-S-S So that's Casey Liss M-A-R-C-O-A-R-M-N-G Marco Arment S-I-R-A-C-U-S-A-C-R-A-C-U-S-A-C-U-S It's accidental They didn't mean to Accidental Tech Podcast So long There's a historical contention between Marco and I about naming things on Long Island. But there is one particular naming thing that he's been doing for years that has just been bothering me. And he did it again in this show.

2:06:49And I figured we need an after show. So let me see if I can convince him to stop doing this.

2:06:52Marco Arment:Oh, my God. I know the argument will probably be that this is not just a you thing. This is just a common thing. And you're just doing what other people do. And so you're fitting in with the culture and yada, yada, yada. And in general, I am a descriptivist, not a prescriptivist, but I can't feel like this one. Maybe I can convince you. Oh, I can't wait. You always talk about going over to the mainland. Yes. But the definition of the word mainland reading from Merriam Webster here is a continent or the main part of a continent distinguished from an offshore island or sometimes from a Cape or Peninsula.

2:07:27Well, guess what? long island is not a continent or the main part of a continent it is in fact an offshore island thus the name and the surrounding of it by water so you're not going to the mainland you're going to another different island this is one of those things that is

2:07:47Marco Arment:technically correct but the purpose again yeah you're right descriptivist versus here's my first question do other people say mainlanders is just a you thing everyone on fire Island when referring to Long Island. I'm not shocked. Those people have no idea what's going on over there. Wow. Most, okay. Most Long Islanders don't even know Fire Island is there and they've never even been there. What are you talking about? Everybody Long Island knows about Fire Island. It's just not an obscure thing. Slow down.

2:08:15John Siracusa:I don't have a horse. On Long Island? Would you relax for a second? I actually don't have a horse in this race, but let's just remind ourselves that your whole oeuvre with regard to Long Island is 30 years out of date? 40 years out of date? Sure, but I will, let me assure you that most people on Long Island don't know Fire Island exists. It's false. Most people on Long Island absolutely know Fire Island exists, and I will wager that most people on Long Island have been to Fire Island. Not only do they know it exists, they've been, but it's not that far away.

2:08:45Marco Arment:And here's why. Because Long Island, which partly undermines your point here, is really big, really big. It's just long. It's full of lots. I mean, that's the main dimension in which it is big, but it's really big. So first of all, everyone on Fire Island refers to Long Island as the mainland. Well, they're all wrong. I think I would love for you. Why don't you come to Fire Island and tell everyone they're wrong and see what they think? I'll just take the dictionary definition and print it out at the ferry terminal. i'll just have it there and just just fyi mainland is not the correct term for another i mean technically north america is an island no north america is not an island it's just a really big one isn't every isn't pangia really an island because it's surrounded by water and the answer to that is no yeah it's not continent has a definition it's all about relative scale what's a continent if not a giant island it's not that's not how this language works i'm sorry i i can go descriptivist with you but like there's a reason we have the word content is australia an island uh no it's okay what about new zealand i know well isn't isn't that crazy it's like a robot or not episode can you imagine that there's a definition depends on depends on the size of things that a very small thing is like a pebble in the water is not an island but then like pangea is not an island but something in between is an island how does that work that's just the nature of the language we use to describe things and all i'm saying is mainland is the wrong word you are wrong well no i think everyone in fire island is just using the wrong word but i understand that it's a cultural thing and they like to say the mainland is just they just happen to pick the wrong word because they don't have another word it would be interesting to see if like someone from hawaii could tell me whether they use mainland to refer to like the big island or whatever probably not because they do call it no well because and but the size differences between the hawaii island the big island hawaii compared to those little tiny ones it's probably a similar size difference yeah well i mean but but i mean they they use the term the big island because that's more descriptive yeah i i don't actually know anything about hawaii so i can't i can't comment on but it's it's similar in nature in that like when you're going from some tiny little speck to something that's massively larger and you call that the mainland and of course obviously people in hawaii have like you know the continental united states to refer to as the mainland but i just want to say that in addition to all the other things that you do nomenclature wise related to fire island and uh the referring to long island as the mainland is nonsensical and I hate it.

2:11:13But apparently it's everyone in Fire Island, not just you.

2:11:15Marco Arment:If you were on Long Island and you referred to the mainland, nobody would have any idea what you were talking about. No one on Long Island refers to the mainland. Right. Because it doesn't make sense in that context. You're making my point for me. I'm not making my point. I'm just saying that no one uses mainland on Long Island. We don't use some other incorrect term instead. Oh, my God, John. You know, like I'm saying, it's not a thing that gets referred to. We could talk about Manhattan, but that's also an island. Right. And nobody in Manhattan would ever say, like, if they were going to New Jersey, I'm going to the mainland.

2:11:53No.

2:11:54Marco Arment:Like, they would never say that. That's true. They wouldn't. And yet it would be the right term if they chose to do that. It would be correct. I literally earlier today walked from Manhattan to New Jersey. I walked to the mainland, John. No one said it. You could have said that correctly in that instance, but you didn't because it's not a thing that's done. But if you did say it, it would be correct because that is a continent or the main part of a continent distinguished from an offshore island. Oh, my God. Yeah. Next time I'll pull all the bikers that are whizzing past me to see, hey, are we going to the mainland?

2:12:21Marco Arment:Where are you going? Is it the mainland over there? They would be confused, but you would say technically I'm using the right word.

2:12:29John Siracusa:I think the thing that cracks me up most about John's ownership of Long Island is that you haven't been able to claim any ownership of Long Island in 30 years. When you grow up there, that's the claim. That's the claim, is growing up there makes all the difference. I get that, but like, maybe it's unfair because I moved so much when I was growing up, and I was only ever in a place for like two, maybe, maybe four years at a time. Yeah, you don't feel like you have an ownership of a place like that. Right, but to me, like, I don't own the podunk little town in western Connecticut that I grew up in, not in a literal sense, you know what I'm saying.

2:13:03Well, Connecticut doesn't have any culture, so.

2:13:05John Siracusa:Oh, all right, now your New York is definitely showing. Except for rich people in Greenwich who consider themselves New Yorkers. I mean, the whole left side of the state thinks they're New York.

2:13:16Marco Arment:Casey, in the last 36 hours, I have stepped foot on all four of these places. Fire Island, Long Island, Manhattan, and the mainland. And John's still going to override me on this.

2:13:27John Siracusa:I know, it's so true. I mean, it's just, it's funny to me how devoutly, John, in particular, you believe in this balls to bones. and yet it's so unequivocally outdated information. Now, it may still be accurate, but it's... The definition of mainland is not an outdated information. It is a definition in the dictionary. Like, maybe they'll change the definition someday. Maybe they'll say, you know, the existing definition and also Long Island. But it doesn't currently say that. So keep working on that. No, my point is that the colloquial usage may be wildly different today. No, it may not be. It may be the same as when you were there.

2:14:04I'm not saying it was any different when I was there. I'm saying if it was like that when I was there was also wrong because it's not the definition of the word. And if they're working on changing the definition, they got more work to do because it hasn't happened yet. That literally figuratively people are doing better. That got in the dictionary. Literally can also mean figuratively. Can you imagine they got the opposite definition in there yet? Mainland is not budging so far.

2:14:26Marco Arment:Like biweekly. Yeah, well, biweekly I think happened before we were all born. By the way, for the record, I'm aware that New Zealand is actually two islands. I just, I didn't even have time to go back and correct it then.

2:14:39John Siracusa:There are so many times on this show that I know, and it's not just me, that I know something that, you know, we either I misspoke or one of us misspoke or perhaps spoke in a not entirely accurate way. And I can hear the of all the emails coming, but the conversation has moved on. And it's, and then what ends up happening because it's me is I forget to come back and be like, oh, by the way. You can insert it in the end. That helps. Well, either way, it's just this happens all the time. And there's so many times that I'm like, oh, I got to make sure I say blah, blah, blah. And then one of the great parts being on the show is that I get to listen to the show as it's happening.

2:15:14John Siracusa:And so sometimes the conversation will move on. And I've completely lost the plot with regard to going back and filling in whatever blank we need to fill in. Marco will occasionally move those things back. If I say, oh, yeah, like, oh, I forgot to say during this section, he'll put it back in that section, which saves me a lot.

2:15:29Marco Arment:By the real-time follow-up, New Zealand is like a million islands. It turns out there's two big ones and a lot of small ones.