PodSearch

ATP

687: You Can Bend This Line

 

00:00:00   Marco, we've covered this on the show before, but hi, it's me and I have no memory. You are using, I almost said iTunes, you're using the music app as your playback app of choice, is that correct?

00:00:10   That makes it sound like I like it more than I do.

00:00:13   Fair.

00:00:14   Yes, I am using it to play music.

00:00:17   All right, here's what I need from you. I need you and me, and I think we probably have some mutual friends that would be interested as well.

00:00:24   And even though this is a terrible idea, it would be an amazing idea. I need you and me and maybe others to work on like a, I guess it would kind of be vaguely similar to a classical music app, but an app for people who are just binging the same like five artists nonstop forevermore. And there's 800 albums per artist.

00:00:44   This is part of the jams app.

00:00:46   I'm telling you, you need to do it.

00:00:48   So here's the thing with the jams app. So this is an app that I intend to make probably with the help of AI at some point. It's one of those apps that like it would never be worth the time to code the whole thing from scratch without something like AI helping.

00:01:01   Because again, like as we've talked about before, like the market for an app that caters to jam band users and jam band users who buy a bunch of the downloaded music and also want an app like this.

00:01:17   We're talking five people like it's every one of those just like it strips it down and down and down further.

00:01:22   And the thing is like there are other music listening apps out there and, and, and, and even like the jam band crowd, um, there's a lot of overlap between like the, the like audio file slash like power user of music crowd and jam band listeners.

00:01:41   So there are already apps. Um, I think is this one called Rune? I think is one of them. Um, there's apps that like specialize in things like, you know, high, high sample rate, lossless playback and go into like fancy DAX and, and DSD and stuff like that.

00:01:56   Like there's all these like high end audio file, um, you know, formats and apps that specialize in catering to those formats.

00:02:04   That world already exists and we wouldn't be any, like anybody who wants that world, we wouldn't satisfy their, their desire.

00:02:15   And, and so like we're, we're going to rule out people who don't care about jam bands.

00:02:20   We're going to rule out people who don't buy all the jam band recordings or otherwise acquire them.

00:02:25   Like people who just want streaming, we're going to rule out people among who's left.

00:02:29   We're going to rule out people who, who actually want like those like high end, you know, high bit rate, audio file, high fidelity, you know, kind of things.

00:02:37   And then we're going to limit it just to people who are on Apple's platforms because, you know, screw that.

00:02:41   And then also then we're, it's going to be like people who, who like our choices and our taste.

00:02:47   And so I think it would be a very small market, which is why it is probably worth doing it with AI,

00:02:54   but probably not any other way.

00:02:56   Well, the reason I bring this up, as you well know, but our listeners may not, is that it is currently goose season.

00:03:02   And I believe you had told me before the show that fish season is coming soon.

00:03:06   And there is a, a really great service called nugs and UGS, uh, that will let you stream or purchase.

00:03:14   I guess you can only stream from nugs actually, but it will let you stream.

00:03:18   I think you can buy downloads.

00:03:19   Maybe you can, but one way or another, you can certainly stream, um, concerts in, in some cases, as they're being aired.

00:03:24   Like I could be watching goose right now, but I'm talking to you two numbnuts.

00:03:27   Uh, and so anyways, you know, as I'm amassing more and more goose shows, as each passing day goes on, I am finding myself more and more in want of a bespoke solution to manage all of this.

00:03:38   And so maybe I can convince you to let me work with you on the jams app or something.

00:03:43   I don't know.

00:03:43   What is the maximum amount you can charge for an app?

00:03:46   Is it a thousand bucks?

00:03:47   If we, if we get, I don't know, a hundred people at a thousand bucks, that math might almost work out maybe.

00:03:53   None of them would pay a thousand bucks for our app.

00:03:55   Based on Marco's description, I feel like you left out the last part, which is that it turns out Marco's taste is different than Casey's taste.

00:04:01   And so now you have to make your own separate vibe code.

00:04:04   So now you really have an audience of one free chap.

00:04:06   And we'd be competing for the search keywords, you know, from, for all six people who would ever search for it.

00:04:11   Um, you know, I am mostly snarking and mostly joking.

00:04:14   However, I will say the thought of you and me in any capacity working on some sort of like actual development project together,

00:04:21   I think would be a beautiful disaster and would make incredible material for the show.

00:04:26   So I'll just put that out there.

00:04:27   That's I, would it be, would it be good for our friendship though?

00:04:30   No, not at all.

00:04:33   Not even a little bit.

00:04:34   It would be so bad.

00:04:36   Somehow, I, you know, I was thinking recently that somehow this show has been going on with 13 years and I'm still dear friends with the two of you idiots.

00:04:42   And yet I think 13 days of Marco and I working on this together, it would be the end of everything.

00:04:48   Yeah, I think like, you know, there are certain, you know, certain friendships that are, that can be ruined by say becoming roommates or friendships that can be ruined by becoming romantic partners.

00:04:58   And I, I think our friendships between the three of us would be ruined with like shared coding projects.

00:05:06   I mean, in, in your defense, our defense and someone's defense, maybe John's defense, he's done a lot of work in your PHP, my friend.

00:05:12   Yeah, because he took it over.

00:05:13   Cause like if he and I were working together, he would have killed me like two years ago.

00:05:19   That's also very fair.

00:05:20   Or vice versa, because, yeah.

00:05:22   Probably that direction though.

00:05:23   Probably you would kill me.

00:05:24   Well, I can tell you, you probably have a lot more opinions about PHP than I do.

00:05:27   I just want to get the hell out of there most of the time.

00:05:29   Before we derail this entire episode, let me tell you, the listeners, about a couple of fun things we have going on.

00:05:37   First of all, we have, I'm going to do this in reverse order, which John's going to yell at me for.

00:05:41   We have a new member special, uh, we have a new member special ATP dev nuggets of wisdom, which is different from nugs of wisdom.

00:05:48   Speaking of nugs, it might be a different route meaning there.

00:05:51   Yeah.

00:05:52   Also fair.

00:05:53   I didn't even consider that.

00:05:54   Uh, anyways, uh, nuggets.

00:05:55   Really?

00:05:56   Wait, you didn't get that?

00:05:57   No, I didn't.

00:05:58   Why a jam band site would be called nugs.

00:06:01   No, never even cross.

00:06:02   Cause I was never in that scene in no small part, because I think if I ever, ever, ever tried it for more than a second, I would never stop.

00:06:08   And so ignorance is lists, my friend.

00:06:11   Anyways, uh, John, tell us about the ATP dev nugs slash nuggets of wisdom.

00:06:15   Yeah, it's just what it says.

00:06:16   Uh, as usual, we went in, uh, kind of blind on this one, because this is the first episode like this that we've done.

00:06:21   Uh, we are all developers, uh, and most of our development history is in the days before LLM.

00:06:26   So we have a lot of, uh, things that we've learned over the course of writing all the code that we've written and figured we would do a member special with, uh, you know, tips and tricks or, you know, lessons or, you know, nuggets of wisdom.

00:06:37   And, uh, my idea was to come in with, uh, you know, things that we've learned and that no idea was too small.

00:06:44   Uh, and that even that instruction turned out to have many loopholes.

00:06:48   So you'll see how it went.

00:06:49   Um, there are nuggets and it is some wisdom.

00:06:53   And if people like it, you know, if people like this episode, we will have, we'll probably do another one because we didn't get through all of our lists and there's surely more that we can add.

00:07:01   And even now we're getting follow-up on it and everything.

00:07:03   So, um, if you are a developer for sure, check this out, especially if you want to hear from a bunch of old people who are developers.

00:07:09   And if you're not a developer, maybe you'll learn something about development or just hear us yell at each other.

00:07:13   So, uh, yeah, there you go.

00:07:14   ATP dev nuggets of wisdom.

00:07:16   Yeah, it was fun.

00:07:17   It was fun.

00:07:18   Uh, and yes, it's very small, has very different definitions between the three of us.

00:07:22   And then, uh, we will make our repeat plea for the ATP store.

00:07:26   You can find that at ATP.fm slash store.

00:07:28   If you are in line for coffee, if you are walking the streets of Manhattan, if you're driving, pull over, step out of line, do whatever you need to do and go to ATP.fm slash store and get yourself some sweet, sweet ATP merch.

00:07:41   John, would you like me to do the nickel tour or would you like to?

00:07:44   I will do it.

00:07:45   I just want to remind everybody the sale ends Sunday, April 26th.

00:07:48   This is the second of three shows that we will talk about it on.

00:07:51   Just do it now because if you don't listen to the third show immediately, by the time you hear it, the sale will be over.

00:07:56   So this is it.

00:07:57   This is actually when you should, if you're going to want something to start, get it now.

00:08:00   Um, what we've got is we've got our ATP Neo shirts, which are in the colors of the MacBook Neo, which we'll talk about more in a little bit.

00:08:08   Um, we've got a Mac Pro Memorial shirt, which is just like the Mac Pro Believe shirt.

00:08:12   But instead of saying Believe, it has a birth date and year date on the bottom.

00:08:16   It's very sad where to make Apple executives guilty.

00:08:18   We've got the T568A crossover and B crossover shirts, which are like the T568A and B Ethernet wiring shirts that we sold in a past sale.

00:08:29   Only now there are the fully crossed crossover cable for both of those standards, making the colors even more mixed up, even nerdier than before, even harder to explain to people when you wear it.

00:08:39   I love this so much.

00:08:40   We have, of course, the M5 Pro and M5 Max shirts.

00:08:43   People ask, did you ever sell a plain M5 shirt?

00:08:45   Yeah, we did when the plain M5 was out on.

00:08:47   And then people are like, oh, but I want to get the plain M5 one now.

00:08:50   Well, here's the thing.

00:08:51   We sell these CPU shirts when the CPU, like whatever the sale is after the CPUs came out.

00:08:57   And then we tend not to sell them again, except for the on-demand ones, which you don't want those.

00:09:01   They're not as good, although they are cheaper.

00:09:03   But anyway, someone's like, oh, I just got an M4 Max.

00:09:06   I want, where can I get an M4 Max shirt?

00:09:08   I always say this.

00:09:09   It's like the old cliche of, you know, dress for the job you want.

00:09:12   Buy the shirt with the chip that you want, not the one you have.

00:09:15   Aspirational.

00:09:16   These are aspirational shirts.

00:09:18   Maybe you don't have an M5 Max.

00:09:20   Maybe someday you think you're going to own one.

00:09:21   Get the shirt now because we won't sell it again.

00:09:24   So M5 Pro and M5 Max shirts.

00:09:26   We like to bring back an old shirt.

00:09:28   This time we brought back ATP Pixels, a very popular shirt, which is the colorful ATP logo

00:09:33   rendered in extremely non-retina pixels, big chunky pixels that are very difficult to print,

00:09:38   but we figured out a way to do it.

00:09:39   That's very cool.

00:09:40   Yeah, the Pixel shirt, by the way, I think among all of the like regular-ish ATP shirts we've ever made,

00:09:47   the Pixel shirt might be one of the nicest.

00:09:49   It's a really, like I really enjoy the complexity of the pattern, like the implementation details of it.

00:09:56   It's a very nice shirt.

00:09:58   That's part of why it has almost no profit margin because it has very expensive printing methods to

00:10:04   get all the sharp pixels and all the different colors, but it is worth it.

00:10:07   Like if you just want a fun, nerdy ATP shirt, that's a good one to get.

00:10:11   And fun fact, I made all those little squares.

00:10:13   It's like one of those things where because I'm not a graphics designer, I have no idea how to use the

00:10:18   tools.

00:10:19   So I'm like, well, I can brute force this.

00:10:21   And I did.

00:10:22   They're all little squares.

00:10:23   It's ridiculous.

00:10:24   It's terrible.

00:10:24   Someone who knows how to do this would look at the file and say, why didn't you just do X,

00:10:27   Y and Z and just automatically cut up your pattern into squares.

00:10:30   It's like, I don't know how, but anyway, they're squares.

00:10:34   We've got a regular ATP shirt, our hoodie, our polo shirt, which we sell in the warmer weather.

00:10:39   It's got a collar on it and embroidered ATP logo.

00:10:41   And of course, our hat.

00:10:42   One other thing people are asking about how the Neo sales, the colors are doing because

00:10:49   we have like a indigo, blush, citrus, and silver to match the MacBook Neo colors.

00:10:54   What are, I don't know if either one of you guessed, seen it.

00:10:58   If you've looked at the sales figures, don't say anything, but you haven't looked at the

00:11:01   sales figure.

00:11:01   What is your prediction for like ratios of the colors that are available?

00:11:04   All right.

00:11:04   I have not looked at the sales figures.

00:11:06   I'm going to guess the blue one.

00:11:08   I'm guessing blue is number one and then it outsells that whatever's number two, two to

00:11:12   one.

00:11:13   And number two, I would guess probably is the, the lemon color one.

00:11:18   Citrus, please.

00:11:21   Casey, you got anything?

00:11:22   So I didn't see numbers.

00:11:24   However, I will say that I, we, we get emails once, once we've crossed the threshold of,

00:11:30   oh, this will be printed by Cotton Bureau.

00:11:31   And I feel like the Neo one came in just lickety split.

00:11:35   So clearly.

00:11:36   Marco would have seen that email too.

00:11:37   So it's kind of cheating.

00:11:38   Oh, I don't read those emails.

00:11:39   So I knew that the Neo was selling the best.

00:11:42   I would say, I think Marco just said this as well, by a factor of two, maybe even three

00:11:47   to one would be my guess.

00:11:48   Cause I haven't seen numbers.

00:11:49   I just know that it.

00:11:50   Wait, are you talking about Neo versus the other shirts or the colors within Neo?

00:11:53   Yeah, you know, I'm sorry.

00:11:54   I misspoke.

00:11:54   I think Indigo is probably two, maybe even three X over the others.

00:11:58   I would say Neo is probably second.

00:11:59   God, I did it again.

00:12:00   I would say Citrus is probably second.

00:12:02   Yeah.

00:12:02   I'd say pink number three.

00:12:03   And then silver, I would say silver's third and blush is fourth.

00:12:08   I'm kind of surprised that you two didn't make the prediction I thought you would make,

00:12:12   which is that, you know, the sort of stereotypical, like when, when someone who, uh,

00:12:19   is like us is buying an iPhone or something, we just get the gray one or the black one.

00:12:23   Right.

00:12:23   Well, but I figured like, if you're going to have like, if, if, if one of these colors

00:12:27   was just black, I think that would outsell by, by a pretty good margin.

00:12:31   But like is Indigo your stand in for that?

00:12:33   Yeah.

00:12:33   In the same way, like with the iPhone 17 that they made like the dark blue one to, for people

00:12:37   who really want like just the boring gray or black one.

00:12:40   Yeah.

00:12:40   Anyway, um, as you know, with these, like these sales where we take a bunch of orders and then

00:12:43   they print the shirts, you have to order a certain number of them before they print the shirts.

00:12:47   And what Casey was talking about is when we crossed the threshold of like now enough people

00:12:51   have ordered this, that we're actually going to print it.

00:12:53   We get an email when that happens.

00:12:54   And so that's where he was getting the info from.

00:12:56   Um, and I can tell you the number required to print one of these shirts is not high.

00:13:00   It's 12, 12 people need to buy a shirt for it to get printed.

00:13:04   If fewer than 12 people buy the shirt or pre-order the shirt, they'll just get their money refunded

00:13:09   and the shirt will not be printed because it's not worth the time for the printing company to

00:13:12   do that because of the scale of things.

00:13:14   Uh, here's the thing.

00:13:15   Uh, Indigo is the most popular as Casey noted because that email came first and they're, they're

00:13:20   not that distant from each other.

00:13:21   The second place is citrus and it is, I don't think it's double, but it's, you know, it's,

00:13:25   it's a pretty big gap.

00:13:26   All right.

00:13:26   But here is the shocker.

00:13:28   We are a week into the sale.

00:13:30   You know how many silver shirts we've sold?

00:13:32   One.

00:13:34   Oh, wow.

00:13:35   Holy jamolis.

00:13:37   We have sold one silver shirt.

00:13:40   Nobody wants that.

00:13:42   I mean, I granted it is the color that is the least accurate to the laptop.

00:13:45   So that's true.

00:13:46   I said, I said it in the, the, the description, I think when I, when I was posting, tooting about

00:13:50   it or whatever, that it's, I thought it felt like it was more important to get the same

00:13:54   shirt from the same manufacturer for all the colors rather than having one be an entirely

00:13:57   different shirt.

00:13:58   And this is the close, this is the best gray that they had to match the silver one, but it's

00:14:02   also very neutral in plain and kind of gray scale.

00:14:04   So I'm like, yeah, maybe people will pick it just because they don't want a brightly colored

00:14:07   shirt.

00:14:07   Indigo is the most popular and Indigo blush and citrus are all going to be printed because

00:14:12   more than 12 people have ordered them.

00:14:13   Although sometimes not much more than 12 silver one shirt.

00:14:17   So whoever, if you're out there and you ordered ATP, Neo and silver, you're not going to get

00:14:22   your shirt unless 11 other people buy that shirt.

00:14:25   I'm not encouraging people to buy it because hey, if you don't want a gray shirt, don't

00:14:28   get the gray shirt.

00:14:29   But I'm just warning that one person who ordered the silver shirt, it's not looking good for

00:14:33   you.

00:14:33   Maybe change your order to Indigo.

00:14:36   It's very popular.

00:14:39   A reminder that ATP members get 15% off.

00:14:42   So if you are a member, go to your member page, get your discount code for 15% off.

00:14:45   If you're not a member, it's worthwhile for you to sign up and get 15% off because you can

00:14:50   make up the cost of one month of membership easily by ordering a couple of things.

00:14:54   So there you go.

00:14:55   And you don't have to cancel after the month is over.

00:14:58   Just put it out there.

00:14:59   You have a lot of specials like ATP Dev Nuggets of Wisdom that you can check out as well.

00:15:03   All right, let's do some follow up.

00:15:06   First of all, Marco made what I presume was just an off-the-cuff judgment call, which given

00:15:11   the information you had made perfect sense.

00:15:13   And oh boy, did we get feedback about it.

00:15:15   So Marco, would you like to handle this?

00:15:18   Yeah.

00:15:19   So when I was talking about in the after show, I was talking about the Suunto watch I was

00:15:22   trying as compared to Garmin watches for sports and fitness watches.

00:15:27   And I referred to Suunto as a young or as a new company, which I think was generally accurate

00:15:35   for their presence in the sport watch market relative to Garmin, I think.

00:15:41   However, everyone wrote in to tell me that Suunto was actually founded in 1936.

00:15:52   They made like compasses at first and then dive computers later.

00:15:56   So eventually, like they did get to the like the like the modern smartwatch.

00:16:01   But the actual company itself is like 100 years old.

00:16:06   So whoops.

00:16:08   So Henry Sivanen writes, on the last episode, Marco referred to Suunto as a younger company

00:16:12   relative to Garmin.

00:16:13   According to the Finnish language Wikipedia, Suunto was founded in 1936, making it 53 years

00:16:17   older than Garmin.

00:16:17   Suunto started making dive computers in the 80s.

00:16:19   The first PC connectivity for those was in the early 90s.

00:16:22   And Suunto introduced a wrist computer in 2004.

00:16:25   We'll put a link to English language Wikipedia in the show notes.

00:16:28   And then Fugu writes, the only thing Suunto seems to have done later than Garmin is to

00:16:32   release a GPS watch in 2012.

00:16:34   The Garmin watch from 2003 already had GPS.

00:16:37   However, it's quite possible that the current Suunto software stack is much younger than Garmin's,

00:16:42   which seems to be much more fragmented.

00:16:44   All right.

00:16:45   Last episode, we were talking about color spaces and we were gushing about Eric Portis' really

00:16:51   great webpage about that, which is linked in the prior show notes.

00:16:55   We were talking about Bartosz Szymanowski's, hopefully I got that close, and basically everything

00:17:02   that he has ever done is incredible.

00:17:04   And John, you had made an off-the-cuff comment about how there was a really good thing that

00:17:10   you had seen recently that you probably couldn't put your finger on.

00:17:12   Well, Ben Landsberg figured it out.

00:17:15   This is what both John and myself were thinking of.

00:17:17   It's ASCII characters are not pixels, a deep dive into ASCII rendering by Alex Hari, which

00:17:23   we'll put a link to that in the show notes for sure.

00:17:26   And if I remember, I'll put the other links back in as well.

00:17:28   Yeah, that's less of an explainer, more of like I was doing the software project and here

00:17:32   is how I progressed through the project.

00:17:34   It's really fun.

00:17:35   Again, with lots of interactive examples and explaining like the reasoning and the different

00:17:40   things that Alex tried to do this somewhat fanciful and whimsical thing.

00:17:45   that he's doing, which is basically like, can you render things using only ASCII characters,

00:17:49   which is something that you're very familiar with if you're a Unix nerd from way back, but

00:17:53   people are still doing it.

00:17:55   All right, is Apple running low on the BINN 18 Pros?

00:17:59   Uh-oh.

00:18:00   Joe Rossingall from MacRumors writes,

00:18:03   The all-new MacBook Neo has been such a hit that Apple is facing a, quote, massive dilemma, quote,

00:18:07   according to Taiwan-based tech columnist and former Bloomberg reporter Tim Kulpan.

00:18:12   Kulpan says the MacBook Neo is selling so well that Apple's supply of the BINN 18 Pro chips

00:18:18   with a five-core GPU will run out before the company is able to fully satisfy demand for the laptop.

00:18:23   Apple's initial plan was to have suppliers build around five to six million MacBook Nios

00:18:27   before ceasing production of the model with the A18 Pro chip, Kulpan said.

00:18:31   But it sounds like demand is so strong that Apple might run out of the A18 Pro chips

00:18:34   before the second-generation MacBook Neo with an A19 Pro chip is ready next year.

00:18:39   A18 Pro chips are manufactured with TSMC's N3E process,

00:18:43   and Kulpan said that the N3E production lines are currently operating at maximum capacity.

00:18:48   As a result, Apple may have to pay a premium to restart A18 Pro chip production for the MacBook Neo,

00:18:54   which would lower its profit margins.

00:18:55   Alternatively, Apple could reallocate some of its chip production that was originally planned for other devices,

00:18:59   but the cost would still be higher than what Apple paid for its initial batch of A18 Pro chips

00:19:04   that were presumably binned.

00:19:05   In either case, Apple would have to disable a GPU core in these chips

00:19:10   just to ensure that they have only a five-core GPU like all the other MacBook Neo units sold to date.

00:19:14   Yeah, I'm not sure that's the case.

00:19:16   So the binned one, this is like a way to, you know, make lemonade out of lemons.

00:19:22   So you're making a bunch of A18 Pros.

00:19:23   Some of them have a bum GPU.

00:19:25   Why don't we just save them?

00:19:26   We'll just put those off to the side.

00:19:28   Don't throw those away.

00:19:29   We might find a use for those, and they use them in the MacBook Neo.

00:19:32   So all the ones with the bad GPU core they can use in the MacBook Neo.

00:19:35   I don't know if that is the entire source of their A18 Pros,

00:19:38   or if they also took a bunch of fully working A18 Pros and disabled one of the GPU cores.

00:19:43   But either way, when you buy a MacBook Neo from Apple, you get an A18 Pro.

00:19:47   One of the GPU cores is either disabled or didn't work to begin with.

00:19:51   Okay, so that's the deal.

00:19:52   If they're running out of those chips, yeah, it kind of sucks for them now.

00:19:56   They have to start up the manufacturing line and start making more of them,

00:19:59   and that costs them money, and they got to take capacity from other things

00:20:01   and all that other stuff, and it lowers their margins.

00:20:04   But they could also just say, okay, well, when we start up the line now,

00:20:09   you will take all the ones that have all the working GPU cores,

00:20:14   and then just update the Neo spec to say, now when you buy a MacBook Neo,

00:20:18   it's like a rev, you know, not a rev2 or whatever, but like a second version of it,

00:20:21   where now all the GPU cores work.

00:20:23   It just seems like a waste to take a perfectly working six GPU core thing

00:20:27   and disable one of them, and people are like, well, you can't do that

00:20:30   because then the people who bought the first one with the core that didn't work,

00:20:33   they're going to be mad.

00:20:33   It's like it happens all the time.

00:20:34   Apple releases a new revision of a product that has some spec slightly better.

00:20:38   That's life.

00:20:40   You buy the product you buy, and if in the future a slightly better version

00:20:44   of that product comes out, you can't say, but I bought one last year,

00:20:47   and it only had five GPU cores, and now you're selling it for the same price with six.

00:20:50   Well, that's technology for you, so they could do that,

00:20:53   or they could pull forward the A19 Pro product or whatever.

00:20:56   So all I can say is this is a good problem to have.

00:20:59   They made a product that everybody loves.

00:21:01   It's selling like hotcakes so much so that even their probably very accurate,

00:21:05   usually very accurate planning, underestimated the demand.

00:21:09   You'd rather have this than the opposite problem,

00:21:11   which is you buy a HomePod on its last day for sale,

00:21:14   and it turns out it was manufactured three years ago,

00:21:16   and those are all the HomePods they ever manufactured.

00:21:18   Yeah, I think, though, like, you know, in this case, you know,

00:21:21   if they do have to make more A18 Pros, I'm like, yeah, if they have to, they will.

00:21:25   They'll work it out with TSMC, and yeah, they might eat a little bit of margin,

00:21:28   but they'll work it out.

00:21:30   I think whether they, you know, hardware disable one of the GPUs,

00:21:35   even if it would have worked, or whether they do this kind of, you know,

00:21:38   quieter revision where they just all of a sudden now they have six core GPUs,

00:21:42   I don't think it really matters either way because the difference of that one GPU core

00:21:48   is not a huge difference.

00:21:49   We're not talking about, like, it'd be different if, like, you know,

00:21:51   the number of PCORs on the CPU was cut in half.

00:21:55   That's not the case here.

00:21:57   This is five of six GPUs or six of six GPUs.

00:22:01   They scale almost perfectly proportionally, so you can figure out the, you know,

00:22:04   the math on that, you know, what is that, like a 16% drop or difference in performance?

00:22:09   It's not a huge difference here, so I think in reality, whichever one of these they do,

00:22:15   it shouldn't and probably won't be, like, a big deal to basically anybody.

00:22:19   The one thing they can't do, or probably, not can't do,

00:22:22   the one thing they probably shouldn't do is manufacture a bunch of new ones

00:22:26   and then just put them into the NEOs, whether they have five or six,

00:22:31   because then you don't know what you're going to get.

00:22:33   And, you know, people are always waiting to, you know, file class action lawsuits again to Apple, right?

00:22:38   So if you decide that you're going to do the six, then they all have to have six,

00:22:42   and you need to update it on your spec page and say it's a new revision of the MacBook Neo.

00:22:45   They all have six now, right?

00:22:47   Because no one can really sue you about that.

00:22:49   You just made a better version of a product later.

00:22:51   But if you don't update the specs at all, but some of them have six and some of them have five,

00:22:56   I don't know if it would be a valid lawsuit,

00:22:58   but I can almost guarantee you somebody would sue and say,

00:23:02   oh, I paid the same price as this person and they got six and I got five

00:23:05   and the spec page says five, doesn't say anything about six.

00:23:07   What's the deal, Apple?

00:23:08   Someone would definitely sue over that.

00:23:09   Again, I don't know if they would win or if that is even remotely valid,

00:23:12   but people love to sue Apple over stuff like that.

00:23:15   So that's the only path they really can't take.

00:23:17   Any other path is open to them, including disabling the working ones

00:23:21   or just selling all the Mystic score.

00:23:22   I'm guessing they would actually just disable the working one to make it five

00:23:26   because I think that is the less problematic version of that.

00:23:30   And there's a lot of precedent for that in the chip business.

00:23:32   Well, you wouldn't waste the ones with only five then

00:23:35   because if you disable the other ones, then you can use every chip.

00:23:37   Well, not every chip, any chip, every chip with five or six you can use.

00:23:40   But if you only use the six ones, the ones with five,

00:23:43   like what are you going to save them for?

00:23:44   Is there going to be another product that's going to want a five GPU?

00:23:47   I'm saying like I think the best course of action, if they need to manufacture more A18 Pros,

00:23:52   I think the best course of action is basically lock them all to five GPUs

00:23:57   and continue to have the product be described and performing the same way.

00:24:02   And there is a lot of precedent for that.

00:24:03   Like back in the old days, there were all these hacks about like overclocking certain Celerons

00:24:09   because I believe Intel had a kind of a similar issue where like they were selling the Celeron

00:24:17   as like their discount chip, but their manufacturing at that time was pretty good.

00:24:22   And so they didn't really have enough like crappy, low performing stuff or whatever it was.

00:24:28   Like their production was so good that they had to basically like soft limit the chips

00:24:33   just to make enough of them to sell.

00:24:36   And your luck in overclocking the chip was kind of in part due to just like

00:24:41   however good yours happened to be above the spec they were actually selling it at.

00:24:45   I bet they got sued by somebody about that too.

00:24:47   I doubt it.

00:24:49   But because like people can sue for anything.

00:24:51   I'm not saying they won the case, but like that's the type of thing that someone would

00:24:54   sue over and say that shouldn't be a random lottery, you know.

00:24:57   Yeah, but it performed to what they advertised it as.

00:25:00   Yeah, I know.

00:25:01   I know.

00:25:01   I'm not saying they won the lawsuit.

00:25:02   I'm just saying that's exactly the type of thing people sue about.

00:25:05   We are sponsored this week by Lisa.

00:25:09   Lisa makes interesting and great mattresses.

00:25:12   Mattresses that are a combination of materials that you wouldn't have thought,

00:25:15   but work together extremely well.

00:25:17   So for example, Lisa was kind enough to send a mattress to us

00:25:21   and Michaela was in desperate need of a mattress after the cheapo Amazon one we got her

00:25:25   when she moved to a big bed was no good.

00:25:28   And so we landed on the natural hybrid mattress.

00:25:31   And this is really cool because among other things, it has breathable organic cotton

00:25:35   and moisture wicking New Zealand wool, which provides cool and comfortable slumber.

00:25:39   And let me tell you, when the mattress showed up, I thought, you know what?

00:25:42   I'm going to sleep on it for a few nights in a different bed and I'm going to see what I think.

00:25:47   And it is really good.

00:25:49   I've really enjoyed how comfortable it was.

00:25:51   I'm very picky about my mattresses.

00:25:53   And to be honest with you, I wasn't sure if I was going to like it.

00:25:55   And I really, really did.

00:25:56   It's really quite good.

00:25:58   You know, Lisa in general has a lineup of all sorts of different kinds of mattresses.

00:26:01   And so many, in fact, that I would advise taking the Lisa sleep quiz.

00:26:05   And you can find your perfect mattress match in just under two minutes.

00:26:09   Their mattresses are meticulously designed and assembled in the United States for exceptional quality.

00:26:13   And they back it all up with free shipping, easy returns, and a 120-night sleep trial.

00:26:18   It's really good stuff.

00:26:20   They've been awarded many times, for example, Best Hybrid and Best Memory Foam Mattresses by Wirecutter.

00:26:25   And they've been featured by West Elm as their go-to mattress provider.

00:26:28   So if you are interested, and you should be, like I said, I got the natural hybrid mattress,

00:26:34   but there's a zillion other options you can try.

00:26:36   And what you can do is go to Lisa.com for 20% off select mattresses,

00:26:40   plus get an extra $50 off the promo code ATP, exclusive for our listeners.

00:26:44   That's L-E-E-S-A.com, promo code ATP, for 20% off select mattresses, plus an extra $50 off.

00:26:51   Support our show and let them know we sent you after checkout.

00:26:54   Lisa.com, promo code ATP.

00:26:56   Thank you to Lisa for sponsoring the show.

00:26:59   All right.

00:27:03   With regard to Apple IDs and pass keys, Vitor writes,

00:27:06   the situation is even more bananas than you described last episode.

00:27:11   Apple added a pass key to my account without my consent with my original Apple ID email.

00:27:15   I've since changed that email, but every time I need to log in,

00:27:18   they still push me to do so with the pass key associated with the old email.

00:27:22   There's no way to remove or update that pass key.

00:27:24   I'm not alone with this problem, and we'll put in one link to Reddit,

00:27:27   the Apple Help subreddit, and a second link, in this case, to the pass key subreddit,

00:27:32   where Apple claims, or so claim somebody on Reddit,

00:27:35   that the fix would be made soon, two years ago.

00:27:39   Whoops.

00:27:40   Yeah, I don't quite understand the problem here,

00:27:42   but clearly it is a real problem of, like, Apple creating pass keys for you,

00:27:45   but they're, like, like everything else having to do with Apple IDs or Apple accounts,

00:27:48   that it's not normal.

00:27:49   It's its own special, unique snowflake that works differently than all the other stuff,

00:27:54   so I guess it won't show up in the passwords app.

00:27:56   And again, surprise, Apple implemented a supposed, you know,

00:28:00   sort of quality of life convenience feature,

00:28:02   where now you can change your Apple ID email.

00:28:03   I bet that won't cause any problems,

00:28:05   just, like, transferring purchases and all the other things they do that it seemed,

00:28:09   they seemed like a good idea until you actually try to use it,

00:28:12   and it turns out, you know, oh, I did, I used that feature where you could change your Apple ID email,

00:28:15   and now you're forever prompting me for a pass key I no longer have access to, apparently.

00:28:19   So, yeah, that whole system is so creaky.

00:28:21   Talk about, you know, again, making a John Turner's to-do list or whatever.

00:28:25   It's like, jeez.

00:28:26   Just there's so many parts of Apple that have not faced any competition in so long,

00:28:32   and they have not gotten any better, that are just filled with garbage,

00:28:35   and the Apple ID system is one of those.

00:28:37   Oh, yeah, especially around pass keys, around developer accounts being separate,

00:28:42   like, all of those conditions.

00:28:43   And it's funny, because, like, you know, the developer account issue is, like,

00:28:46   this is, like, every Apple developer hits this issue.

00:28:48   And the login process for a developer is so awkward,

00:28:52   because you have to first, like, hit cancel,

00:28:54   no, I don't want to use my pass key for my personal Apple ID,

00:28:57   and then log in for the developer, use the password,

00:29:00   and then if it prompts you for 2FA,

00:29:01   then you have to, even if the machine that you're on is trusted,

00:29:06   it will put up the dialog box with the six-digit code,

00:29:10   and the input box on the website that you're logging into

00:29:14   is centered perfectly in the middle of the browser window,

00:29:16   well, then the six-digit code dialog box is centered in the middle of the screen.

00:29:21   So it covers up.

00:29:23   If your browser window is centered on your screen,

00:29:24   then the code that pops up covers up where you have to enter it.

00:29:28   So you have to drag the code window off to the side a little bit

00:29:31   to be able to then re-access on the page the six-digit codes.

00:29:35   You can type it in, which it went on autofill.

00:29:37   Like, it's the worst experience.

00:29:38   Like, Apple has made logins so much nicer for everyone else except Apple.

00:29:44   It's very, very weird.

00:29:46   And you're forgetting my favorite game,

00:29:47   which now I find myself doing unconsciously,

00:29:49   which is, you know, when it sends me the little, you know,

00:29:53   Apple ID login thing that shows the map,

00:29:55   it'll send it to my phone or whatever,

00:29:56   and it'll say, you know, allow, deny,

00:29:58   and it shows the little map.

00:29:59   I'm like, yeah, that's me.

00:30:00   I hit allow.

00:30:01   After hitting allow, it then shows the six digits.

00:30:05   But sometimes, as I've complained about in the past,

00:30:07   sometimes it will show the six digits for a fraction of a second

00:30:09   before they disappear.

00:30:10   So now, when I hit that allow,

00:30:12   I prime my brain to memorize the six digits as they flash on the screen.

00:30:17   Because sometimes, as I've said when I was setting things up,

00:30:20   sometimes there's just no way to do, like, it will never stay, you know,

00:30:23   unless it comes up on my phone, I hit allow,

00:30:25   the six digits appear, and they immediately disappear.

00:30:27   And I have to memorize them in that fraction of a second that they appear,

00:30:30   so I get good at it.

00:30:31   It's ridiculous.

00:30:32   Ridiculous.

00:30:33   All right.

00:30:34   And then finally for tonight,

00:30:35   some updates with regard to stealing television.

00:30:37   This is the overtime discussion we had in 679 about the super box

00:30:43   and the other ways that people are stealing TV.

00:30:45   And Nye Hughes writes,

00:30:47   this reminded me of a Guardian article that suggests,

00:30:50   in the UK at least,

00:30:51   modded Amazon Fire TV sticks are being used to log keystrokes

00:30:54   and passwords to enable fraud.

00:30:56   A recent survey, this is from the article,

00:30:58   a recent survey from B Streamwise,

00:31:00   a UK initiative established to counter the problem,

00:31:02   found two out of every five people who used illegal streaming were defrauded.

00:31:06   They lost an average of almost 1,700 pounds each as a result.

00:31:11   Whoopsie-dipsies.

00:31:12   The potential for fraud happens the minute you connect the dodgy stick,

00:31:15   or so it's called,

00:31:16   with a laptop or a TV,

00:31:18   according to Rob Shapland of Psionic Cyber.

00:31:21   Quote,

00:31:22   alongside the stream of TV or sports,

00:31:25   will also install some malware onto your computer

00:31:27   and give the criminal direct access to your computer

00:31:29   so that they can use it as if they were sitting there,

00:31:31   he claims, says, whatever.

00:31:32   Quote,

00:31:34   or they can install key loggers,

00:31:35   which will record any password you are typing.

00:31:37   So when you're accessing online banking,

00:31:38   it will record your banking passwords.

00:31:39   You are essentially volunteering to have your laptop hacked in many cases.

00:31:42   Whoopsie-dipsies.

00:31:43   And then additionally,

00:31:45   there was a Krebs on Security article

00:31:47   where Brian Krebs writes,

00:31:48   Ashley, an engineer at Census,

00:31:49   a cyber intelligence company,

00:31:51   found that the Superbox devices

00:31:52   immediately contacted a server

00:31:54   at the Chinese instant messaging service Tencent QQ,

00:31:56   as well as a residential proxy service called Grass.io.

00:32:00   Also known as GetGrass.io,

00:32:03   Grass says it is a decentralized network

00:32:05   that allows users to earn rewards

00:32:06   by sharing their unused internet bandwidth

00:32:08   with AI labs and other companies.

00:32:09   Quote,

00:32:10   buyers seek unused internet bandwidth

00:32:12   to access a more diverse range of IP addresses,

00:32:14   which enables them to see certain websites

00:32:15   from a retail perspective,

00:32:16   the Grass website explains.

00:32:18   By utilizing your unused internet bandwidth,

00:32:21   they can conduct market research

00:32:22   or perform tasks like web scraping to train AI.

00:32:25   I would just love to meet,

00:32:26   I think the person who wrote this copy

00:32:27   did such a good job

00:32:28   because it's like,

00:32:29   okay, here's what we're doing.

00:32:31   We need to appear to be legitimate people.

00:32:33   So we're going to come

00:32:34   and do your legitimate people's internet connection.

00:32:36   So basically we'll be indistinguishable

00:32:37   from legitimate people.

00:32:38   Basically we want to turn people

00:32:39   into a giant bot that we can control.

00:32:40   All right, now you write copy for that

00:32:42   and make it sound not nefarious.

00:32:43   They want to access

00:32:45   a more diverse range of IP addresses,

00:32:47   which enables them to see certain websites

00:32:49   from a retail perspective.

00:32:51   They did it.

00:32:52   They did it.

00:32:54   You made it not sound terrible,

00:32:55   but it is.

00:32:57   It's really terrible.

00:32:57   It sure is.

00:32:59   So finally from Krebs on security,

00:33:01   it may be that many Superbox customers

00:33:04   don't care if someone uses their internet connection

00:33:06   to tunnel traffic for ad fraud

00:33:07   and account takeovers.

00:33:08   For them-

00:33:08   Oh, I forgot about ad fraud.

00:33:09   Yeah, making it seem like people

00:33:10   are clicking on ad banners.

00:33:11   They're coming from legit IP addresses.

00:33:14   For them, it beats paying

00:33:16   for multiple streaming services each month.

00:33:18   My guess, however,

00:33:19   is that quite a few people

00:33:20   who buy or are gifted these products

00:33:21   have little understanding

00:33:22   of the bargain they're making

00:33:23   when they plug them into an internet router.

00:33:24   Some of the other,

00:33:26   I didn't pull the quotes,

00:33:27   but some other sections of the same article,

00:33:29   I believe it was the Krebs article,

00:33:30   talk about how they use some sort of hack.

00:33:33   I'm outside my comfort zone,

00:33:34   but basically they force a device

00:33:37   that's already on your local network

00:33:39   to crash or what have you.

00:33:41   Again, my terminology is wrong.

00:33:42   And they will then slurp up that IP address

00:33:45   so it looks even more legitimate

00:33:47   to your own network.

00:33:48   Like, it's crazy

00:33:49   some of the shenanigans these things do.

00:33:51   Basically, do not touch.

00:33:52   And then finally, with regard to this,

00:33:54   David Brownman writes,

00:33:56   the DocNet Diaries podcast

00:33:57   did a pretty detailed episode

00:33:59   on the cybersecurity side of the Superbox,

00:34:01   and we'll put a link to that in the show notes.

00:34:03   Yeah, so we did touch on this in the overtime

00:34:05   that there's, you know,

00:34:06   lots of these boxes are shady

00:34:07   and people claim that the software on them

00:34:09   is attacking people or whatever,

00:34:10   but this is like a concrete study

00:34:12   from the UK saying that people

00:34:14   are getting their money stolen for real,

00:34:15   like immeasurable large amounts

00:34:18   by using these products.

00:34:19   And, you know,

00:34:20   the attraction of getting TV for free

00:34:23   is apparently enough

00:34:24   to keep people coming back,

00:34:26   but they're bad news.

00:34:27   Indeed.

00:34:29   All right.

00:34:29   Let's talk topics.

00:34:31   And maybe I should have poured a drink

00:34:33   for this episode

00:34:33   because we got to talk about ads

00:34:35   in Apple Maps, baby.

00:34:36   For real this time.

00:34:38   So Emma Roth,

00:34:39   the Verge wrote,

00:34:40   I don't know,

00:34:40   I think a few weeks ago

00:34:41   as we record this,

00:34:42   Apple will soon allow businesses

00:34:44   to buy advertisements

00:34:45   in its Maps app.

00:34:47   In an announcement on Tuesday,

00:34:48   Apple says ads may appear

00:34:50   at the top of your search results

00:34:51   in Maps,

00:34:51   as well as in a new

00:34:52   suggested places list.

00:34:54   They will arrive in-app

00:34:55   in the US and Canada

00:34:56   this summer.

00:34:57   Marco,

00:34:58   if you wouldn't mind,

00:34:59   would you prepare your bleep gun?

00:35:01   Because this is such

00:35:03   fucking bullshit

00:35:04   and I hate it so much.

00:35:05   I haven't even seen it

00:35:06   and I hate it.

00:35:07   Oh,

00:35:07   it makes me so mad.

00:35:08   It makes me so mad

00:35:10   because I,

00:35:11   and I think many like me,

00:35:12   came to Apple

00:35:14   because we were so used

00:35:16   to getting PCs

00:35:17   filled with bullshit malware,

00:35:20   just overflowing with it

00:35:22   everywhere.

00:35:23   By the way,

00:35:24   on that topic,

00:35:25   and I know this is not

00:35:26   in the notes of this week

00:35:26   and maybe we'll talk

00:35:27   about it next week,

00:35:28   but there was a recent interview

00:35:29   that Jaws and Ternus

00:35:31   did with Tom's hardware

00:35:32   and they were,

00:35:33   they were fielding

00:35:34   a bunch of softball questions

00:35:35   about like what's great

00:35:36   about Apple products

00:35:37   or whatever.

00:35:37   And,

00:35:38   and they went through

00:35:39   everything that's great about it,

00:35:40   but then they were on

00:35:40   to the next question

00:35:41   and ATP style,

00:35:42   one of them,

00:35:43   I forget who it was,

00:35:43   says,

00:35:43   oh yeah,

00:35:44   and also,

00:35:44   don't forget,

00:35:45   Macs aren't loaded up

00:35:46   with bloatware.

00:35:47   That was their point

00:35:49   that like,

00:35:49   unlike the competing PC products,

00:35:51   the products that compete

00:35:52   with the MacBook Neo,

00:35:53   when you get a Mac,

00:35:54   it's not filled with like,

00:35:55   you know,

00:35:55   like you said,

00:35:56   a bunch of crappy software

00:35:57   that you don't want

00:35:58   that's thrown a bunch

00:35:59   of stuff in your face

00:36:00   and I felt like reaching

00:36:00   through the YouTube screen

00:36:02   and saying,

00:36:02   but it is,

00:36:03   increasingly it is.

00:36:04   Yeah.

00:36:05   Have you seen the defaults,

00:36:06   the doc?

00:36:08   God,

00:36:08   I just had my,

00:36:09   blue screen.

00:36:10   Obviously when it's Apple stuff,

00:36:11   that's not an ad,

00:36:12   that's Apple stuff.

00:36:13   No,

00:36:13   I know,

00:36:14   exactly,

00:36:14   but like the default.

00:36:15   The new version of Pages.

00:36:16   So I actually have seen

00:36:17   the default doc

00:36:18   because,

00:36:19   when I was setting up

00:36:21   48 Mac minis

00:36:22   without doing any kind

00:36:23   of MDM,

00:36:24   one of my steps

00:36:25   was taking all of those

00:36:27   out one by one.

00:36:28   Whoosh.

00:36:28   Whoosh.

00:36:28   Whoosh.

00:36:29   You can automate that

00:36:30   by just deleting the,

00:36:32   changing the P list.

00:36:32   Well,

00:36:33   I didn't know that

00:36:34   at the time.

00:36:34   I didn't want to ask you

00:36:35   because then it would

00:36:35   spoil the bit.

00:36:36   So I did it.

00:36:37   I,

00:36:37   every single,

00:36:38   for 48 Mac minis,

00:36:41   one of my setup steps

00:36:42   was dragging out

00:36:43   all like,

00:36:43   you know,

00:36:44   15 of the doc icons

00:36:45   out of the doc,

00:36:46   except for,

00:36:46   did you get really good

00:36:47   at figuring out

00:36:48   where the remove thing

00:36:49   appears?

00:36:49   Yeah,

00:36:49   because it's not,

00:36:50   you know,

00:36:50   you have to go up

00:36:51   a little bit,

00:36:51   like,

00:36:52   you know,

00:36:52   it can't just be like,

00:36:53   you know,

00:36:53   the first pixel.

00:36:54   You got to go up

00:36:55   like,

00:36:55   you know,

00:36:55   200 pixels up.

00:36:56   Remember when it used

00:36:57   to have the poof?

00:36:57   Yeah,

00:36:58   exactly.

00:36:58   Well,

00:36:58   and it still makes

00:36:59   the same sound.

00:37:00   Those are the days.

00:37:00   And that's just,

00:37:01   anyway,

00:37:03   sorry to derail you,

00:37:03   Casey.

00:37:04   No,

00:37:04   no,

00:37:04   no,

00:37:04   not filled with bloatware.

00:37:05   I'm not sorry.

00:37:06   So,

00:37:08   anyways,

00:37:09   you know,

00:37:09   so many of us,

00:37:10   and I think especially

00:37:11   those of us who had grown

00:37:12   up with PCs,

00:37:13   like Marco,

00:37:13   like me,

00:37:14   came to Apple

00:37:15   because especially

00:37:16   circa,

00:37:17   you know,

00:37:17   mid-2000s,

00:37:18   my recollection,

00:37:19   maybe I'm just looking

00:37:20   with rose-colored glasses,

00:37:21   but my recollection was

00:37:22   I had been using PCs

00:37:24   for so long,

00:37:25   they were utter crap.

00:37:26   They were crap

00:37:27   in ways that I knew

00:37:28   and crap in ways

00:37:29   that I didn't even

00:37:30   understand yet.

00:37:30   And they were crap.

00:37:31   And then finally,

00:37:33   and John,

00:37:34   I mean,

00:37:34   the timeline may not

00:37:36   be exactly right,

00:37:36   but the way I recall it

00:37:37   is that it was around

00:37:39   the mid-aughts

00:37:39   that Mac OS X

00:37:42   was really starting

00:37:42   to get its junk together.

00:37:43   Now,

00:37:43   you could argue

00:37:44   it was great

00:37:45   from the beginning,

00:37:45   but it was like

00:37:45   properly good

00:37:46   at this point.

00:37:47   And so,

00:37:48   I moved to the Mac

00:37:49   and I think it was 2008,

00:37:50   if I'm not mistaken,

00:37:51   maybe it was 2007.

00:37:52   No,

00:37:52   it was 2007.

00:37:53   And it was such

00:37:55   a breath of fresh air.

00:37:56   Everything worked.

00:37:57   It was intuitive.

00:37:59   It was beautiful.

00:38:00   There was no garbage.

00:38:01   I didn't have to

00:38:02   put my wrist

00:38:03   on an Intel inside sticker

00:38:04   for the rest of time

00:38:05   and just watch it

00:38:07   peel away slowly,

00:38:08   but surely.

00:38:09   And everything

00:38:10   was so amazing

00:38:11   and so wonderful.

00:38:12   And I feel like

00:38:14   in the last five,

00:38:15   maybe ten years

00:38:15   in particular,

00:38:16   all of that goodness

00:38:18   has gone away

00:38:19   in various

00:38:20   different ways.

00:38:21   Software isn't as stable

00:38:24   as it was.

00:38:25   The Apple software

00:38:25   isn't as stable

00:38:26   as it was.

00:38:26   And it does

00:38:28   so many things.

00:38:29   And I think

00:38:31   Apple is spread

00:38:32   so thin

00:38:33   that they lose

00:38:34   track of things.

00:38:35   Like we were just

00:38:36   lamenting a moment ago,

00:38:37   like Apple IDs.

00:38:38   But I feel like

00:38:39   the software's been

00:38:40   spread so thin,

00:38:41   there's so many

00:38:42   edge cases

00:38:42   that Apple

00:38:42   no longer conquers.

00:38:44   Now,

00:38:44   in the defense of Apple,

00:38:45   it's also gotten

00:38:45   way more complicated,

00:38:46   which is hard.

00:38:47   It's very hard.

00:38:48   But it doesn't

00:38:49   just work anymore.

00:38:51   and that bums me out.

00:38:53   Additionally,

00:38:54   the search for

00:38:56   ever more money

00:38:57   and ever more services

00:38:58   revenue

00:38:58   means that

00:38:59   instead of just

00:39:00   being a company

00:39:01   that makes their money

00:39:02   off of hardware,

00:39:03   now they have to

00:39:04   make their money

00:39:05   off of recurring revenue

00:39:06   from me paying them

00:39:07   for things

00:39:08   that sometimes I think,

00:39:09   yes,

00:39:10   they deserve money

00:39:11   for this.

00:39:11   And sometimes I'm like,

00:39:12   what is going on here?

00:39:13   For example,

00:39:13   iCloud storage.

00:39:14   I should probably

00:39:15   pay them

00:39:15   for some amount

00:39:16   of iCloud storage,

00:39:17   but should that

00:39:17   iCloud storage

00:39:18   still be 5 gigabytes

00:39:19   in the year 2026?

00:39:22   I think not.

00:39:23   And it's just frustrating

00:39:26   that this is

00:39:26   yet another example.

00:39:28   This is yet another example

00:39:30   of things just getting

00:39:32   ickier

00:39:32   and getting more

00:39:34   Windows-like.

00:39:35   And honestly,

00:39:35   I haven't used Windows

00:39:36   in over a decade.

00:39:37   And it is,

00:39:39   I genuinely wonder

00:39:40   at this point,

00:39:41   is Windows

00:39:43   less garbaged up

00:39:45   than Apple stuff?

00:39:46   No,

00:39:46   Windows is getting worse.

00:39:47   They're putting ads

00:39:48   in the start menu and stuff.

00:39:49   All right,

00:39:49   well,

00:39:49   that makes you feel

00:39:50   a little bit better.

00:39:50   And this is part

00:39:52   of the shame of it.

00:39:52   And this is kind of

00:39:53   part of what I was hoping,

00:39:54   what I was trying to get at

00:39:55   with my letter to John Ternas

00:39:57   blog post.

00:39:57   Apple is getting

00:39:59   significantly worse

00:40:01   in certain ways

00:40:02   in areas that are

00:40:03   very important to us.

00:40:04   But no one else

00:40:07   is getting better at it.

00:40:08   It's not like Windows

00:40:09   is amazing at this.

00:40:11   Yeah,

00:40:11   Windows is worse

00:40:12   than it's ever been

00:40:12   at this,

00:40:12   arguably.

00:40:13   Yeah.

00:40:13   Industry-wide,

00:40:16   there is a slow

00:40:17   process of,

00:40:19   I won't use

00:40:21   Cory Doctorow's word

00:40:22   necessarily,

00:40:22   because I don't think

00:40:23   it quite fits here,

00:40:23   but there's a slow

00:40:25   process of just

00:40:26   erosion of respect

00:40:29   for users' time,

00:40:31   attention,

00:40:32   and resources.

00:40:33   Motivated by

00:40:34   the possibility

00:40:35   of making more money.

00:40:36   Right.

00:40:36   And even,

00:40:37   and if it's,

00:40:38   the reality is,

00:40:39   like,

00:40:40   in many industries,

00:40:41   their margins

00:40:42   get super squeezed,

00:40:43   and they get

00:40:44   desperate.

00:40:44   And they start

00:40:46   doing things

00:40:47   that are crappy

00:40:47   because they kind

00:40:49   of have to

00:40:49   to make any money

00:40:50   at all.

00:40:51   And the really sad

00:40:52   thing is,

00:40:52   that's always been

00:40:54   Apple's competitors.

00:40:54   That's never been

00:40:55   Apple's situation.

00:40:56   Apple,

00:40:57   I mean,

00:40:57   not never,

00:40:57   but, you know,

00:40:58   certainly not in the

00:40:59   modern era.

00:41:00   Apple's modern era

00:41:01   is,

00:41:02   they have more money

00:41:03   than they know

00:41:04   what to do with.

00:41:05   They don't need

00:41:06   to become desperate

00:41:08   to find ways

00:41:09   to make 0.5%

00:41:11   more money.

00:41:11   They can just

00:41:13   make better products.

00:41:13   Look,

00:41:14   we were just talking

00:41:15   about the MacBook Neo

00:41:16   doing really well,

00:41:18   like,

00:41:19   blowing away

00:41:19   expectations

00:41:20   because they made

00:41:22   a really good product

00:41:22   and people like it.

00:41:23   It fits the market well,

00:41:24   it's there

00:41:26   at a good time,

00:41:26   you know,

00:41:27   the opportunity

00:41:27   to take market share

00:41:28   from PCs right now

00:41:30   is really huge

00:41:31   for lots of reasons.

00:41:32   And so,

00:41:32   what a great product.

00:41:33   Look at the success

00:41:35   of the iPhone 17 Pro.

00:41:36   They made the cool orange one

00:41:38   and it's a great product.

00:41:40   They, like,

00:41:40   so they had this cool color

00:41:41   that did the marketing

00:41:43   for them largely

00:41:43   and then they made

00:41:45   this amazing pro phone

00:41:46   that took what we wanted

00:41:48   about iPhones

00:41:48   and removed certain pain points,

00:41:50   made things better,

00:41:51   like,

00:41:52   it's just,

00:41:52   the iPhone 17 Pro

00:41:53   is an amazing product.

00:41:55   Look at the modern Macs.

00:41:56   The modern Macs

00:41:57   beyond just the Neo.

00:41:58   They're amazing.

00:42:00   They're so amazing

00:42:01   you can't even buy

00:42:02   the desktops anymore

00:42:03   right now

00:42:03   because they're all sold out.

00:42:04   Like,

00:42:05   the modern Mac lineup

00:42:07   is so good.

00:42:08   There is not a single

00:42:10   bad laptop

00:42:11   that Apple sells anymore.

00:42:12   What a thing

00:42:13   to be able to say

00:42:14   after, you know,

00:42:15   some of the kind of

00:42:15   mid-2010s

00:42:17   kind of years

00:42:18   there got kind of bad.

00:42:19   But,

00:42:20   what an amazing state

00:42:22   the products are in

00:42:23   and they can make

00:42:24   more money

00:42:24   by expanding

00:42:26   the market

00:42:27   of those products

00:42:28   or by making people

00:42:29   upgrade

00:42:30   when they're

00:42:30   compelling upgrades.

00:42:31   And they do.

00:42:32   That is still

00:42:33   how they make

00:42:33   most of their money.

00:42:34   They can even have

00:42:35   this whole thing

00:42:36   called services

00:42:37   even though

00:42:38   what services

00:42:39   makes most of us

00:42:40   money from

00:42:40   is rents

00:42:41   and, you know,

00:42:41   but let's,

00:42:42   we'll put that aside

00:42:43   for this particular

00:42:44   part of this discussion.

00:42:45   You know,

00:42:46   services

00:42:47   makes tons of money

00:42:48   based on

00:42:49   delivering

00:42:51   additional

00:42:52   revenue sources

00:42:53   to their products

00:42:54   most of which

00:42:55   don't annoy

00:42:57   or badger

00:42:58   or harass

00:42:59   their users.

00:42:59   Most.

00:43:01   And then they

00:43:03   started selling ads

00:43:04   first in the app store

00:43:05   and, you know,

00:43:07   now coming to Maps.

00:43:08   And there's a few

00:43:10   things about this

00:43:11   that irritate me greatly.

00:43:12   Number one,

00:43:13   it's just not

00:43:15   that much money

00:43:16   relative to

00:43:17   the services category.

00:43:19   especially since

00:43:19   Apple's really bad

00:43:20   at selling ads

00:43:20   against anything.

00:43:21   Well, see that,

00:43:22   yeah, that's,

00:43:22   that's the second part

00:43:23   that annoys me.

00:43:24   So, like,

00:43:24   is there,

00:43:24   what is the potential

00:43:25   upside for Apple?

00:43:26   Not what is the

00:43:27   potential upside

00:43:27   for Google

00:43:28   because Apple's

00:43:28   not Google.

00:43:29   Yeah, the second

00:43:30   part of this

00:43:31   that annoys me

00:43:31   is that

00:43:32   I have been

00:43:33   an Apple search

00:43:34   ads customer

00:43:35   since search ads

00:43:37   launched.

00:43:37   I actually,

00:43:37   I just paused

00:43:38   my search ad

00:43:39   campaign

00:43:40   for the first

00:43:41   time in years

00:43:42   just about a month

00:43:43   ago because I'm

00:43:44   mad at them.

00:43:45   But the Apple

00:43:46   search ads

00:43:47   product

00:43:47   is

00:43:48   awful.

00:43:49   The only thing

00:43:51   holding it up

00:43:51   as being anything

00:43:52   that is worth

00:43:53   anyone ever

00:43:53   spending money

00:43:54   on is that

00:43:55   there are still

00:43:56   a lot of people

00:43:56   in the apps

00:43:56   that are searching

00:43:57   for things

00:43:57   and now they've

00:43:58   created this

00:43:58   awful situation

00:43:59   for all of us

00:44:00   which is

00:44:01   if you don't

00:44:02   buy like

00:44:03   your own name

00:44:04   and your own

00:44:04   keywords

00:44:04   everyone else

00:44:05   will buy them

00:44:06   and you will

00:44:06   have to fight

00:44:07   and now you

00:44:08   have to pay

00:44:08   Apple extra

00:44:09   just to get

00:44:10   to your own

00:44:10   customers

00:44:11   that you were

00:44:11   already going

00:44:12   to get before

00:44:12   and

00:44:13   they're still

00:44:15   taking their

00:44:15   30%

00:44:16   of course

00:44:17   of all your

00:44:17   in-app purchases

00:44:18   and stuff

00:44:18   so it just

00:44:19   feels a lot

00:44:20   like double

00:44:20   dipping

00:44:20   but it's

00:44:23   a crappy

00:44:24   ads product

00:44:25   and it makes

00:44:26   the app store

00:44:27   worse

00:44:28   for everybody

00:44:29   except Apple

00:44:29   Apple makes

00:44:31   a little bit

00:44:31   more money

00:44:32   from that

00:44:32   great

00:44:33   good job

00:44:33   everything else

00:44:34   is worse

00:44:35   it's worse

00:44:35   for customers

00:44:36   it's worse

00:44:37   for most

00:44:38   developers

00:44:38   and it's

00:44:39   mostly just

00:44:40   a way

00:44:40   for big

00:44:41   companies

00:44:41   to pour

00:44:42   a bit

00:44:42   more money

00:44:42   into Apple's

00:44:43   pocket

00:44:43   with high

00:44:44   bids on

00:44:44   keywords

00:44:44   and now

00:44:47   they're bringing

00:44:47   that same

00:44:48   benefit to

00:44:48   Maps

00:44:49   great

00:44:50   adding ads

00:44:51   in places

00:44:52   like this

00:44:52   to make

00:44:53   reasonably

00:44:54   small amounts

00:44:55   of additional

00:44:56   revenue

00:44:56   sells out

00:44:58   the user

00:44:58   experience

00:44:59   yes

00:45:00   a little

00:45:00   bit

00:45:01   yes

00:45:01   it sells

00:45:02   it out

00:45:02   because

00:45:02   what Apple

00:45:03   what their

00:45:04   actions say

00:45:05   here is

00:45:06   it matters

00:45:07   less to us

00:45:08   to have

00:45:08   a great

00:45:09   high-end

00:45:10   premium

00:45:10   user

00:45:11   experience

00:45:11   we would

00:45:12   rather make

00:45:13   1% more

00:45:13   money

00:45:13   and they

00:45:14   have

00:45:14   seemingly

00:45:15   run out

00:45:15   of ideas

00:45:16   to make

00:45:16   1% more

00:45:16   money

00:45:17   in other

00:45:17   ways

00:45:17   I guess

00:45:18   or they

00:45:18   just don't

00:45:18   care

00:45:19   and this

00:45:19   is like

00:45:19   the hallmark

00:45:20   of the

00:45:20   Tim Cook

00:45:21   era

00:45:21   is like

00:45:22   product

00:45:23   direction

00:45:23   by numbers

00:45:24   and there's

00:45:26   a place

00:45:26   for numbers

00:45:27   try the new

00:45:28   version of

00:45:28   numbers

00:45:29   oh my god

00:45:29   oh that's

00:45:31   a whole other

00:45:31   don't even

00:45:32   get me

00:45:32   started

00:45:33   on that

00:45:34   oh I'm

00:45:34   so mad

00:45:35   about that

00:45:35   but again

00:45:37   but that's

00:45:37   that's actually

00:45:38   a similar

00:45:39   problem here

00:45:39   it's like

00:45:40   nothing is

00:45:41   sacred to them

00:45:41   anymore

00:45:42   in the user

00:45:43   experience

00:45:43   they are

00:45:44   willing to

00:45:45   put promos

00:45:45   which are

00:45:46   ads

00:45:46   and actual

00:45:48   ads

00:45:48   all over

00:45:49   their products

00:45:50   now

00:45:50   there is no

00:45:52   more premium

00:45:53   experience with

00:45:53   Apple products

00:45:54   because now

00:45:55   it's all full

00:45:55   of ads

00:45:56   because

00:45:56   what are you

00:45:57   gonna do

00:45:58   where are you

00:45:58   gonna go

00:45:59   there's

00:45:59   nowhere else

00:46:00   so they can

00:46:01   now make

00:46:01   1% more

00:46:02   money from

00:46:02   you

00:46:02   old Apple

00:46:04   wouldn't

00:46:05   have done

00:46:06   that

00:46:06   you know

00:46:07   you don't

00:46:08   I don't like

00:46:08   to invoke

00:46:08   Steve Jobs

00:46:09   that often

00:46:09   I know

00:46:09   there were

00:46:10   a couple

00:46:10   of cases

00:46:10   where Steve

00:46:11   Jobs

00:46:11   floated the

00:46:11   idea of

00:46:12   putting ads

00:46:12   in different

00:46:12   places

00:46:12   but

00:46:13   under

00:46:14   previous

00:46:15   leadership

00:46:15   Apple

00:46:17   would not

00:46:17   have been

00:46:18   putting ads

00:46:18   in the places

00:46:19   they're putting

00:46:19   ads now

00:46:20   because I

00:46:21   think the

00:46:21   previous

00:46:22   leadership

00:46:22   had more

00:46:22   appreciation

00:46:23   for a

00:46:24   premium

00:46:24   user

00:46:24   experience

00:46:25   and when

00:46:26   you start

00:46:27   being managed

00:46:27   as you know

00:46:28   just bean

00:46:29   counter

00:46:29   only numbers

00:46:30   matter

00:46:31   and nothing

00:46:31   else matters

00:46:32   if this thing

00:46:33   can increase

00:46:34   numbers by

00:46:34   5%

00:46:35   why wouldn't

00:46:35   we do it

00:46:36   obviously we

00:46:36   must do it

00:46:37   because the

00:46:37   numbers will

00:46:37   go up

00:46:38   that is a

00:46:39   pretty rapid

00:46:41   path to

00:46:42   mediocrity

00:46:42   and that's

00:46:44   never been

00:46:45   what has

00:46:45   made Apple

00:46:46   stand out

00:46:47   they have

00:46:48   certainly had

00:46:49   many mediocre

00:46:50   products and

00:46:51   services and

00:46:51   times in the

00:46:52   past no

00:46:52   question but

00:46:53   where Apple

00:46:54   succeed

00:46:55   needs is

00:46:56   making nice

00:46:57   products with

00:46:58   nice experiences

00:46:59   and ads

00:47:00   don't have a

00:47:01   place in that

00:47:01   and I wish

00:47:02   Apple's leadership

00:47:03   would realize

00:47:04   where their

00:47:04   own success

00:47:05   comes from

00:47:05   like you said

00:47:06   before like

00:47:07   it's not

00:47:07   you were saying

00:47:08   like have they

00:47:08   run out of

00:47:09   idea other

00:47:10   ways to make

00:47:10   money it's not

00:47:11   that they've

00:47:11   run out of

00:47:11   other ideas

00:47:12   for incremental

00:47:13   revenue or

00:47:14   other ideas

00:47:14   to innovate

00:47:15   it's that

00:47:15   in addition to

00:47:17   those ideas

00:47:18   these ideas

00:47:18   which used

00:47:19   as you noted

00:47:19   used to be

00:47:20   get shot

00:47:20   down

00:47:20   no longer

00:47:21   get shot

00:47:22   down and

00:47:22   that is the

00:47:22   dysfunction

00:47:23   in the

00:47:23   organization

00:47:24   someone in the

00:47:25   org is always

00:47:25   going to say

00:47:26   hey we can

00:47:27   make more

00:47:27   money if we

00:47:28   did this

00:47:29   that would

00:47:29   annoy users

00:47:30   slightly

00:47:30   but what are

00:47:30   they going to

00:47:31   do like

00:47:31   that idea

00:47:32   will always

00:47:33   exist will

00:47:33   always be

00:47:34   presented is

00:47:34   always compelling

00:47:35   it's a compelling

00:47:37   business case to

00:47:38   Apple because you

00:47:38   can say look

00:47:39   I can show you

00:47:39   how this will

00:47:40   make us more

00:47:40   money it will

00:47:41   make us more

00:47:41   successful in

00:47:42   this way and

00:47:43   what the what

00:47:44   Apple needs to

00:47:45   do and what

00:47:46   they're getting

00:47:46   less good at

00:47:47   doing is

00:47:48   saying yes

00:47:49   yes but

00:47:50   our brand

00:47:52   promises that

00:47:53   we don't do

00:47:53   crap like

00:47:54   that and

00:47:54   long term

00:47:55   even though

00:47:55   your idea

00:47:55   will make

00:47:56   a little

00:47:56   money now

00:47:56   it will

00:47:57   erode that

00:47:58   brand promise

00:47:59   the value

00:47:59   proposition

00:48:00   it will

00:48:00   in the long

00:48:01   run it will

00:48:02   make us less

00:48:02   money it's the

00:48:03   exact opposite

00:48:03   of the thing

00:48:04   that made

00:48:04   Apple successful

00:48:05   which is do

00:48:06   the difficult

00:48:06   thing that

00:48:07   everyone thinks

00:48:07   is a dumb

00:48:08   idea because

00:48:08   if you do

00:48:08   it long

00:48:09   enough or

00:48:09   hard enough

00:48:09   and it

00:48:10   makes users

00:48:10   happy you

00:48:10   will make

00:48:11   more money

00:48:11   than them

00:48:11   in the long

00:48:12   run and

00:48:13   Apple does

00:48:14   still fight

00:48:15   that they're

00:48:15   not saying yes

00:48:15   to every one

00:48:16   of those

00:48:16   ideas but

00:48:17   they're saying

00:48:17   yes to way

00:48:18   too many of

00:48:18   them in

00:48:19   particular saying

00:48:20   yes to like

00:48:20   their own

00:48:21   internal things

00:48:22   like what this

00:48:22   part of the

00:48:23   org says we

00:48:24   need to you

00:48:25   know do

00:48:26   upsells inside

00:48:27   the iWork

00:48:28   applications because

00:48:29   that's the only

00:48:29   way people will

00:48:30   pay for this and

00:48:30   we need to do

00:48:31   this and we

00:48:31   need to put

00:48:32   like we need

00:48:33   to advertise

00:48:34   the you know

00:48:35   new Apple care

00:48:35   plans inside

00:48:37   system settings

00:48:37   because I know

00:48:39   the people who

00:48:39   are writing

00:48:40   system settings

00:48:40   they don't care

00:48:41   about that but

00:48:41   this part of

00:48:42   the org cares

00:48:42   about selling

00:48:43   Apple care and

00:48:44   we need a way

00:48:44   to sell okay

00:48:45   hey we're all

00:48:45   Apple here

00:48:46   same team

00:48:46   so let's do

00:48:47   it and that's

00:48:47   what they're

00:48:47   saying yes to

00:48:48   they're saying

00:48:48   yes to you

00:48:50   know letting

00:48:51   themselves advertise

00:48:52   their own crap

00:48:52   everywhere which is

00:48:53   already terrible

00:48:54   and then in these

00:48:55   cases like search

00:48:55   ads and map

00:48:57   ads they're saying

00:48:57   we can get money

00:48:58   from third parties

00:48:59   here too because

00:48:59   we have all these

00:49:00   customers we have

00:49:00   their eyeballs

00:49:01   they're doing

00:49:01   something related

00:49:02   to maps they're

00:49:03   in the app store

00:49:03   searching like you

00:49:05   said Marco it's a

00:49:05   it's a win for

00:49:06   Apple they think

00:49:07   because it's like

00:49:08   look we annoy the

00:49:09   user and we suck

00:49:11   more money from

00:49:12   developers and we

00:49:12   annoy the developers

00:49:13   so everything is

00:49:15   worse the only

00:49:16   arguable possible

00:49:17   user benefit is

00:49:18   if you're searching

00:49:19   for an app you

00:49:20   get to see some

00:49:21   possible alternatives

00:49:22   to that app like

00:49:23   if it's a really

00:49:23   well-known app you

00:49:24   would otherwise

00:49:24   never find the

00:49:24   alternatives that is

00:49:25   the only benefit I

00:49:26   can think of for

00:49:27   that for the user

00:49:27   but mostly it's a

00:49:28   it's a it's a net

00:49:29   negative for everybody

00:49:30   but I was like yeah

00:49:31   but we make more

00:49:31   money so yeah that's

00:49:33   the problem they

00:49:33   keep saying yes to

00:49:34   that and then this

00:49:34   the ads and maps

00:49:35   like I don't even

00:49:37   think of all the

00:49:38   things they could do

00:49:38   this is that

00:49:40   egregious other than

00:49:40   the fact that they're

00:49:41   going to do a

00:49:41   terrible job at it

00:49:42   except for the

00:49:44   fact that as far

00:49:46   as I know I don't

00:49:47   this is is this a

00:49:48   rumor I think it's a

00:49:48   rumor I don't know if

00:49:49   it was announcement

00:49:49   or when I was

00:49:50   announcement announcement

00:49:51   from Apple but

00:49:51   anyway we don't know

00:49:52   the details but here's

00:49:53   the thing so you're

00:49:54   going to put ads and

00:49:55   stuff at very least

00:49:57   allow people to pay

00:49:59   money to make the

00:50:00   ads go away how

00:50:01   about the people who

00:50:02   are already paying

00:50:03   you for let's say the

00:50:04   Apple one bundle

00:50:05   because like I'm not

00:50:07   saying oh it's fine

00:50:08   Apple put ads

00:50:09   everywhere and just

00:50:10   make people pay to

00:50:10   get rid of them

00:50:11   but if we have to

00:50:12   live in this world

00:50:12   if we're gonna have

00:50:13   these ads forced on

00:50:14   us like to Casey's

00:50:15   point if I'm already

00:50:16   paying you like the

00:50:17   maximum amount a

00:50:18   person can pay you

00:50:19   like I buy your most

00:50:20   expensive Apple one

00:50:21   bundle that you want

00:50:22   me to get I buy all

00:50:23   your products or

00:50:23   whatever is there no

00:50:25   way that me paying

00:50:27   money will make ads

00:50:28   go away anyway I

00:50:29   would pay to get rid

00:50:30   of search ads in the

00:50:31   app store if I could I

00:50:32   would certainly pay to

00:50:33   get rid of ads in

00:50:34   Apple Maps even though I

00:50:34   mostly use Google Maps

00:50:35   like and I'm not

00:50:37   saying like obviously

00:50:38   the number of people

00:50:39   who are gonna pay to

00:50:39   make ads go away is

00:50:40   tiny like we know

00:50:41   we have we have a

00:50:42   podcast where people

00:50:43   pay to make the ads

00:50:43   go away most people

00:50:45   don't like it but we

00:50:45   get that but like just

00:50:48   let let that's it's

00:50:49   such an easy out for

00:50:50   them to continue to be

00:50:51   crappy like again I'm

00:50:52   not saying this is the

00:50:53   solution but hey Apple

00:50:54   if you want to continue

00:50:55   to erode your your

00:50:56   own value you would

00:50:58   just you would silence

00:51:00   more of Casey's cursing

00:51:01   if you just at least

00:51:02   allowed allowed the

00:51:04   relative handful of

00:51:05   people who are willing

00:51:06   to pay not to see ads

00:51:07   to let them do that

00:51:08   they would grumble

00:51:09   about it and they

00:51:09   would be pissed but it

00:51:10   would make them slightly

00:51:12   more quiet than the

00:51:13   current case of just

00:51:14   like we don't care how

00:51:14   much money you give us

00:51:15   there's no escape from

00:51:16   these ads yep couldn't

00:51:18   agree more it's just

00:51:20   this is the sort of

00:51:21   thing now I'm showing a

00:51:23   lot of biases and I'm

00:51:24   bringing a lot of my

00:51:25   own priors as Merlin

00:51:26   would say to the table

00:51:27   but I've always found

00:51:29   it frustrating that it's

00:51:33   in my mind I filed it

00:51:35   as a West Coast thing

00:51:36   that there are no

00:51:37   decisions made there's

00:51:38   only A-B testing I know

00:51:40   I'm being hyperbolic and

00:51:41   I know it's not quite

00:51:42   that simple but I feel

00:51:43   like everything is A-B

00:51:44   tested and everything is

00:51:46   about numbers and at

00:51:47   some point you have to

00:51:49   understand that there is

00:51:51   there are things that are

00:51:52   unmeasurable and there

00:51:54   are things that are just

00:51:55   a vibe and the vibe that

00:51:57   I get from Apple these

00:51:59   days and I don't think

00:52:00   I'm alone is that oh

00:52:02   they're just as shitty as

00:52:04   everyone else except just

00:52:06   a teeny bit less so if

00:52:08   Windows is an 11 Apple is

00:52:10   an 8 or a 9 but that's

00:52:12   still shittier than it

00:52:14   needs to be and I'm sorry

00:52:15   that I don't have more

00:52:16   grown-up language or maybe

00:52:17   less grown-up language to

00:52:19   express my feelings but I'm

00:52:20   so frustrated that this is

00:52:25   what this is and there's

00:52:27   nowhere else to go I think

00:52:28   that's the other problem

00:52:29   like yes I suppose I

00:52:30   could make this the year

00:52:31   of Linux on the desktop for

00:52:32   me but I don't want to do

00:52:33   that nobody really wants to

00:52:34   do that please don't

00:52:35   redeem me even NPM throws

00:52:36   ads at you now they're

00:52:37   not really ads but like

00:52:38   hey these projects need

00:52:39   funding because they've

00:52:41   got you at the prompt and

00:52:42   how else are you going to

00:52:43   discover that the software

00:52:44   you're that you've been

00:52:45   using for free forever

00:52:45   needs funding like I

00:52:47   think I don't begrudge

00:52:48   NPM that but I'm just

00:52:49   saying like everybody that

00:52:50   the idea has always

00:52:52   existed which is right

00:52:53   where the person's using

00:52:54   your software can we put

00:52:56   ads there someone will

00:52:58   always have that idea

00:52:59   the idea is not new what

00:53:01   is the difference now is

00:53:02   people who would

00:53:03   formally say of course I'm

00:53:04   not in my command line

00:53:05   Unix utility I'm going to

00:53:07   put an ad like I

00:53:08   understand open source

00:53:09   needs funding but this is

00:53:10   not the way to do it but

00:53:11   eventually the generation

00:53:13   that grew up in the

00:53:14   current environment gets

00:53:15   to control open source

00:53:16   projects and said oh yeah

00:53:17   totally we should put that

00:53:18   into NBM and so now it's

00:53:19   there yeah and I think

00:53:20   like what what irritates me

00:53:22   about Apple doing it and

00:53:24   this is granted this is

00:53:25   less about maps now now I'm

00:53:26   talking about the platform

00:53:27   is like if inside of an

00:53:30   app like maps they put

00:53:32   ads okay well we can use

00:53:35   different apps and yeah I

00:53:36   mean the other ones have

00:53:38   ads too but at least they're

00:53:40   better so you know whatever

00:53:42   there's different options for

00:53:44   mapping apps you can make

00:53:45   your own mapping app if you

00:53:46   really want to nobody can

00:53:48   really anyway but you know

00:53:50   you know what I mean but when

00:53:52   the system has ads in places

00:53:55   that are required to interact

00:53:57   with or privileged things like

00:54:00   the settings app having promos

00:54:02   for all of Apple's upsells and

00:54:04   crap the app store having ads

00:54:07   when you are looking to buy to

00:54:08   get software in the only way you

00:54:09   can get software on your iOS

00:54:11   device there's no way to escape

00:54:14   that so you you don't have a

00:54:16   choice as a customer you are

00:54:17   locked in and you are forced to

00:54:19   see those ads and this is not

00:54:21   like you know a severe like

00:54:25   human rights issue or anything

00:54:26   but it is Apple kind of

00:54:29   mishandling their position of

00:54:31   power abusing the power they

00:54:33   have as the platform to make

00:54:36   things a little bit crappier

00:54:37   under the guise of well

00:54:39   everyone's doing this as you

00:54:40   were saying like everyone's

00:54:41   doing this so therefore we can

00:54:43   do it too and too bad for you

00:54:45   if you don't like it that

00:54:47   doesn't make it better and it

00:54:48   totally works like to Casey's

00:54:50   point about AB testing I can

00:54:51   guarantee you they sell more

00:54:52   Apple care now that they push

00:54:53   it to you in system settings

00:54:54   like there's no question about

00:54:56   that that is that's the you

00:54:57   know that's why you don't let

00:54:58   it happen because you know if

00:54:59   you do let it happen of course

00:55:01   it's going to give you a

00:55:02   return on investment like it's

00:55:04   the one and only system

00:55:05   settings and you get to put ads

00:55:06   there and that will that will

00:55:08   sell more Apple care and the

00:55:09   people who came up with ideas

00:55:10   to see I told you and the

00:55:11   people against the idea says

00:55:12   we're never arguing that it

00:55:13   wouldn't work we're arguing

00:55:14   that we shouldn't do it

00:55:15   because it erodes the value of

00:55:17   Apple as a brand and a

00:55:18   company and it just and like

00:55:20   even even if you can't get

00:55:21   them to quantify the brand

00:55:24   damage it seems like you also

00:55:27   can't get a lot of these higher

00:55:29   ups there now to believe in the

00:55:32   promise of just making things

00:55:34   because they're better this way

00:55:36   instead of adding ads somewhere

00:55:38   because it makes some number go

00:55:39   up maybe you can have the

00:55:42   prestige and self-control to

00:55:46   show some restraint and not take

00:55:48   every single opportunity to make

00:55:50   a bit more money if it makes the

00:55:52   product worse that's what I was

00:55:54   talking about with the turnist

00:55:55   thing it's like that if you're

00:55:56   Steve Jobs you say no that makes

00:55:58   it crappy I'm not going to do it

00:55:59   and everyone disagrees you but if

00:56:00   you're John turnist and you say

00:56:01   that then now you have people

00:56:02   saying well he turned down this

00:56:03   idea that we prove could give you

00:56:05   X amount more money in Apple

00:56:06   care subscriptions and he said no to

00:56:07   it for an unquantifiable benefit

00:56:10   we can show you the numbers this

00:56:12   is how much more money we make on

00:56:13   Apple care if we do this and all

00:56:14   they had to say was some mumbo

00:56:16   jumbo about brand value like

00:56:17   whatever show me the numbers on

00:56:19   that dummy and because you're not

00:56:20   Steve Jobs and didn't found the

00:56:22   company and everybody worships you

00:56:23   that it becomes a ding against you

00:56:25   and it's like that's that's what

00:56:27   makes people do the wrong thing

00:56:28   because they don't feel secure

00:56:30   enough to say we're not doing

00:56:31   that because it sucks like and I'm

00:56:33   sure sometimes they do again it's

00:56:34   not like it's not like they're

00:56:35   just you know bending over and

00:56:36   doing every bad idea that exists

00:56:38   but there you can in this battle

00:56:41   within Apple you can see the people

00:56:43   who want to junk stuff up are

00:56:44   winning more of these fights and

00:56:46   that's the trend against the trend

00:56:48   line I would still say it's you

00:56:49   know Casey said that like Windows is

00:56:51   an 11 and Apple is an 8 I think the

00:56:52   range is much greater Windows is in

00:56:54   a way worse situation but like you

00:56:58   know we're not Windows users so we

00:56:59   don't care I'm just looking at the

00:57:00   trend lines that's why we're even

00:57:01   though every one of these little

00:57:02   things it seems small and it is

00:57:04   small but like the line is going in

00:57:06   the wrong direction if we just

00:57:08   extrapolate bad things happen you

00:57:11   know Apple does have an area that

00:57:14   they are a little bit better at

00:57:16   defending and that's privacy privacy

00:57:18   in some ways I mean again it's a

00:57:22   little bit different because you know

00:57:23   when Apple says privacy is a human

00:57:25   right I think that is that is about

00:57:28   as important as it is I don't think

00:57:29   that's that's hyperbole and I do

00:57:31   think that the that the right to use

00:57:34   your computer without seeing ads is

00:57:36   not a human right but they do things

00:57:39   in the name of privacy now granted

00:57:40   every not every time they invoke

00:57:43   privacy is for good reasons many times

00:57:46   they will invoke privacy in things like

00:57:48   you know defenses against monopoly

00:57:50   accusations and bad behavior basically

00:57:52   as a as a BS defense to try to excuse

00:57:56   their bad behavior that's really that

00:57:57   they're really doing for other

00:57:58   reasons but not every time many of

00:58:00   the times that Apple invokes privacy

00:58:01   are real and genuine and beneficial

00:58:05   and Apple really does care about

00:58:07   privacy more than just the numbers

00:58:09   alone would would you know dictate

00:58:11   that they should that is an area that

00:58:14   it's it's it's considered a core value

00:58:16   to the company and that's and credit

00:58:19   to Tim Cook that is something he has

00:58:21   brought to the table and he cares a

00:58:22   lot about it isn't that Steve Jobs

00:58:24   didn't care about privacy but Tim Cook

00:58:25   seems to care a lot more and I give him

00:58:27   credit for that is a core value that

00:58:30   Apple defends privacy in their product

00:58:32   choices even when it's really

00:58:34   difficult most of the time but under

00:58:39   Tim Cook that same consideration has

00:58:42   not been held for the quality of the

00:58:45   user experience that's where you know

00:58:47   Steve Jobs really did hold that as a

00:58:48   high value the same way Tim Cook values

00:58:51   privacy Steve Jobs valued a premium

00:58:54   nice user experience and Tim Cook

00:58:56   doesn't and modern Apple doesn't they

00:59:00   happen to have enough people there who

00:59:02   are left who still are able to create

00:59:04   good user experiences most of the time

00:59:06   but there isn't any pressure in the way

00:59:11   that there is so much pressure to

00:59:12   preserve privacy at Apple there is no

00:59:14   such pressure to preserve the quality of

00:59:17   the user experience and there really

00:59:19   needs to be because having all these

00:59:21   little ads and promos and little paper

00:59:23   cuts everywhere each one of those is an

00:59:26   erosion of quality and an erosion of the

00:59:29   premium brand and the experience and

00:59:32   they don't seem to think so or at least

00:59:34   those people are not able they're not

00:59:36   empowered to defend the quality of your

00:59:39   experience when oh well if but if we

00:59:41   ruin this thing a little bit more we can

00:59:43   make a bit more services revenue and I

00:59:46   think the thing that makes me one of

00:59:49   the things that makes me so

00:59:49   exasperated about this is that there's

00:59:53   almost no way to put this back away

00:59:56   right what's what's the turn of phrase

00:59:57   I'm looking for there's no way back in

00:59:59   the bottle there is a way you just

01:00:00   make different decisions in the future

01:00:01   you can bend this line you need a new

01:00:03   CEO to do it but you know fingers

01:00:05   crossed but even still are they ever

01:00:06   going to take away a revenue stream

01:00:08   really that's that's what I'm yeah

01:00:10   some Steve Jobs did things like that

01:00:12   all the time you see good CEOs it's

01:00:14   called leadership you can't say I'm

01:00:16   never going to do anything that's

01:00:17   going to cause any kind of kind of bad

01:00:19   result you know for because there's a

01:00:22   long-term benefit that's called

01:00:23   leadership that's what Apple needs

01:00:25   good CEOs make decisions like that all

01:00:27   the time it is absolutely possible

01:00:28   will they do it I don't know but I

01:00:30   I totally believe it's but we see it

01:00:32   happen all the time we saw it happen

01:00:34   with turning around Apple itself when

01:00:35   Steve Jobs came back and did it he

01:00:37   destroyed tons of stuff before

01:00:38   things got better well also it's not

01:00:41   even that much money yeah that's the

01:00:43   other thing the other thing is not

01:00:44   it's not like you're saying stop

01:00:45   selling the iPhone come on like it's

01:00:47   you'll be okay like we're talking

01:00:49   about like a drop in the bucket we

01:00:51   even within the services category

01:00:53   yeah which itself I mean granted the

01:00:55   services category is pretty large now

01:00:57   as a percentage of their profit and

01:00:58   that's why they keep pushing it it's

01:00:59   the growing one and the stock market

01:01:01   won't like it and yada yada like

01:01:02   there's a million reasons but it's

01:01:04   it's not insurmountable or end it's

01:01:07   not that much money like I mean

01:01:09   obviously in absolute terms like

01:01:10   it's probably more money than all

01:01:11   all of us will ever see but a huge

01:01:13   amount of money absolute terms but in

01:01:14   relative terms yeah in relative terms

01:01:16   relative to their their services

01:01:17   revenue we're not talking about

01:01:19   getting rid of the entire app store

01:01:21   cut and the entire Google search deal

01:01:23   and like you know the things that

01:01:24   actually make big differences we're

01:01:26   not talking about that we're talking

01:01:28   about drops in the bucket especially

01:01:29   like ads and maps the way Apple

01:01:33   handles ad sales and the systems they

01:01:36   built and the the primitive garbage

01:01:38   ranking algorithms they use those

01:01:41   aren't going to be that successful

01:01:42   financially like because they're just

01:01:44   they're not going to be that good

01:01:45   the app store search ads probably make

01:01:48   substantially more money than anything

01:01:51   that will ever be made in maps and

01:01:53   the app store search ads are also

01:01:55   terrible but at least those are more

01:01:56   like you know there's more kind of

01:01:58   of like an intent to buy there's a lot

01:02:01   probably a lot more search volume if I

01:02:02   had to guess you know compared to

01:02:04   monetizable maps transactions like or

01:02:06   maps sources so I think this is going

01:02:08   to be we're talking about crapping up

01:02:11   a core system app for not that much

01:02:14   money and that that's what makes it

01:02:15   even more frustrating that it's not

01:02:17   like they're like betting the whole

01:02:18   company on some huge thing that's

01:02:20   going to turn a huge amount of money

01:02:21   over no it's not they're just adding

01:02:23   another paper cut to make one of their

01:02:26   core apps worse for everyone to make a

01:02:29   drop more money that sucks and by the

01:02:32   way that's exactly what they did to

01:02:34   numbers pages and keynote and screw

01:02:35   them for that I'm I'm so annoyed at

01:02:37   what they did to those apps no

01:02:39   argument I think the one thing that I

01:02:41   will say with regard to that's

01:02:42   leadership I agree however if we think

01:02:45   about the situation with Apple when

01:02:47   Steve came back like they were on

01:02:49   death's door or at least so it so I've

01:02:51   understood I mean I wasn't I wasn't

01:02:53   around for this I wasn't an Apple person

01:02:55   for this but they were on death's door

01:02:56   90 days from bankruptcy they say yeah

01:02:58   exactly so of course Steve came in and

01:03:02   you know destroyed everything and

01:03:03   turned everything upside down not only

01:03:05   was he one of the founders and that

01:03:06   kind of gave him the clout to do it but

01:03:08   what else are they gonna do he did that

01:03:10   when they were successful too I mean

01:03:11   the iPod mini and getting killed for

01:03:14   the iPod nano even though it was their

01:03:15   most successful iPod he was killing the

01:03:16   successes that were helping them come

01:03:18   back from where they were like he was

01:03:19   not afraid to do what he thought was

01:03:21   right like that's what I'm saying it's

01:03:23   leadership like you can Tim Cook would

01:03:25   not make that choice and would make a

01:03:26   very different very strong argument

01:03:28   to keep selling the iPod mini because

01:03:30   it's the most popular iPod they ever

01:03:31   made and just sell the nano to the

01:03:32   people who want that one to see the

01:03:33   jobs like no that was better kill the

01:03:35   iPod mini and I get that all I'm

01:03:37   saying is yes it is leadership but I

01:03:39   think that in a lot of ways Steve was

01:03:42   set up for success set up for doing the

01:03:45   things that had to be done because

01:03:46   clearly whatever was going on wasn't

01:03:48   working whereas right now that you

01:03:50   know whoever takes over be that

01:03:52   turnist or whoever else they're

01:03:53   entering the cockpit of a rocket ship

01:03:56   and to slow that rocket ship down

01:03:58   even though I again I concur that it's

01:04:00   relatively not a lot of money that's

01:04:02   still a tougher it's a tougher bigger

01:04:04   ask is all I'm saying yeah what they

01:04:06   would argue is like oh it's not going

01:04:07   to lose us that much money but our

01:04:08   share price is going to go down so our

01:04:09   market cap is going to lose ten

01:04:10   billion dollars or whatever and Steve

01:04:11   Jobs would say so what like it'll be

01:04:13   fine but like I said it's not just

01:04:15   when things are going bad the iPad

01:04:16   iPod mini was going great that was a

01:04:18   number that was going up and that was

01:04:20   a case where you're making a

01:04:22   decision that people would say

01:04:23   you're you're you're risking the the

01:04:25   rocket ship by making that change

01:04:28   but Steve Jobs wanted what he wanted

01:04:29   and most of his decisions were right

01:04:32   in that direction so anyway we'll see

01:04:33   we'll see what happens I'm not sure

01:04:35   this trend line is going to change at

01:04:36   all because I see no deviation or even

01:04:38   acknowledgement that everything we're

01:04:40   talking about is even a thing when

01:04:41   Apple executives talk like do you even

01:04:43   see them acknowledging this like I

01:04:44   said with a turnist and jaws saying

01:04:46   and of course Macs aren't junked up

01:04:48   with anything I'm like well they are

01:04:50   now they're they're like they don't

01:04:51   see it they don't if they do say it

01:04:53   they're not saying they're not

01:04:54   voicing it in any way so it's really

01:04:56   difficult to expect a different

01:04:57   decision when there is not even any

01:04:59   public acknowledgement that this issue

01:05:01   exists let alone where Apple is on the

01:05:02   spectrum so

01:05:03   we are sponsored this episode by Zapier

01:05:08   Zapier is how you actually deliver on

01:05:10   your AI strategy not just talk about it

01:05:12   you can put AI to work across your

01:05:14   company right now for real with all the

01:05:16   tools that you already have Zapier ties

01:05:19   them all together you can bring the

01:05:20   power of AI to any workflow with Zapier

01:05:23   whether it's AI powered workflows

01:05:25   autonomous agents chatbots or whatever

01:05:26   you can imagine in fact one thing I

01:05:28   tried with Zapier I thought you know

01:05:30   what let me type in they have this

01:05:31   wonderful prompt on their site and it

01:05:33   said you know what do you want to

01:05:34   make and I said why don't you email me

01:05:36   every morning at 7 a.m. whenever the

01:05:38   weather is going to be five degrees

01:05:40   different than it was yesterday or any

01:05:42   important weather events like

01:05:43   precipitation and in these two

01:05:45   different zip codes that I care about

01:05:46   and I watched as Zapier's AI agent

01:05:49   created an entire workflow based on my

01:05:52   prompts it asked me a couple of

01:05:54   clarifying questions when necessary it

01:05:55   asked me to approve things that need

01:05:57   human approval which I thought was

01:05:58   very nice very respectful and it

01:06:00   tested things out and it showed me how

01:06:02   things were working and it built the

01:06:03   whole thing at any point I could have

01:06:05   jumped in and customized whatever I

01:06:06   wanted and but it knew what data

01:06:08   sources were it knew how to tie them

01:06:09   together it was an incredible thing to

01:06:11   see not only is this automation

01:06:13   possible on Zapier and you can help

01:06:16   tie AI tools together with Zapier but

01:06:18   you can even use their built-in AI to

01:06:20   create the AI tools which I thought was

01:06:23   really quite a quite a fun thing to see

01:06:25   so Zapier you know you don't have to be a

01:06:26   tech expert to do it you can just tell

01:06:28   it what to do and it'll do it or if you

01:06:30   are a tech expert of course you have

01:06:31   to all the all the control and and

01:06:34   options and everything that you might

01:06:36   want you can build things yourself you

01:06:37   can have AI build things you can have a

01:06:39   combination of both it can do whatever

01:06:41   your team needs with AI powered

01:06:43   workflows so join the millions of

01:06:45   businesses transforming how they work

01:06:47   with Zapier and AI get started for free

01:06:49   by visiting Zapier dot com slash ATP

01:06:52   that's Zapier spelled Z-A-P-I-E-R

01:06:56   Zapier dot com slash ATP thanks to

01:06:59   Zapier for sponsoring our show

01:07:01   let's talk about AI generated art and

01:07:08   Emma Roth again the verge Emma Roth and

01:07:11   the verge on March 2nd writes the US

01:07:14   Supreme Court has declined to hear a

01:07:16   case whether over whether AI generated

01:07:18   art can obtain copyright the decision

01:07:20   comes after Stephen Thaler a computer

01:07:23   scientist from Missouri appealed a

01:07:25   court's decision to uphold a ruling

01:07:27   that found AI generated art can't be

01:07:28   copyrighted in 2019 the US copyright

01:07:31   office rejected dollars request

01:07:32   copyright an image called a recent

01:07:34   entrance to paradise on behalf of an

01:07:36   algorithm he created the copyright

01:07:38   office reviewed that review that

01:07:40   decision in 2022 and determined that

01:07:42   the image doesn't include quote human

01:07:44   authorship quote disqualifying it from

01:07:46   copyright protection after Thaler

01:07:48   appealed the decision the US district

01:07:49   court judge barrel a howell ruled in

01:07:52   2023 that quote human authorship is a

01:07:55   bedrock requirement of copyright quote

01:07:56   that ruling was later upheld in 2025 by

01:07:58   federal appeals court in Washington DC

01:08:00   last year the copyright office issued new

01:08:02   guidance that says AI generated artwork

01:08:04   based on text prompts isn't protected by

01:08:06   copyright the US federal circuit court

01:08:07   similarly determined that AI systems

01:08:10   can't patent inventions because they

01:08:11   aren't human which the US patent office

01:08:13   reaffirmed in 2024 with new guidance

01:08:15   stating that while AI systems can't be

01:08:17   listed as inventors on a patent people

01:08:18   can still use AI powered tools to

01:08:20   develop them UK Supreme Court made a

01:08:22   similar determination in a case brought

01:08:24   forward by the same fellow so this is a

01:08:27   background on just establishing what what

01:08:30   least the US law has been saying about

01:08:32   copyright and AI art and a few other

01:08:34   things because the actual topic here is

01:08:37   AI generated code and copyright and when I

01:08:40   put this in the notes which was a while

01:08:41   ago at this point lots of people online

01:08:43   were stating flat out as a fact that hey

01:08:47   don't you know because of the Supreme

01:08:48   Court decision in the US any code you're

01:08:51   generating with these coding agents this is

01:08:52   back when coding agents were just sort of

01:08:54   becoming popular so you know a little while

01:08:56   ago now any code you're generating with

01:08:58   those coding agents is not copyrightable

01:09:00   therefore you don't you know that you

01:09:03   there's no ownership or protection or

01:09:05   copyright protection for the code that

01:09:06   you're doing for this so if you're if you

01:09:07   are coding with one of these agents and

01:09:08   you're putting that code into your

01:09:09   product what was previously copyrighted

01:09:12   code that you owned is no no longer is

01:09:13   because you can't copyright AI generated

01:09:15   stuff here's what the Supreme Court says

01:09:16   and most of the time people just let that

01:09:21   exist as a thing that people say and

01:09:23   they'd be sure I didn't know that I

01:09:24   didn't know you can't copyright AI

01:09:25   generated code but I'm not a lawyer or a

01:09:28   judge but my understanding from reading

01:09:30   all these stories is this has yet to be

01:09:32   adjudicated I know this doesn't help but

01:09:34   so like so many things having to do with

01:09:35   LLMs and AI stuff it's an open question

01:09:39   the the Supreme Court has ruled on AI

01:09:42   generated art but I wouldn't just simply

01:09:45   assume that because they made a decision

01:09:47   about AI generated art and patents that

01:09:49   that same decision exactly applies as is to

01:09:53   AI generated code because historically art

01:09:56   and code and patents and things like that

01:09:58   have all been treated differently by the

01:10:00   courts so as far as I'm aware this is yet

01:10:04   another thing involving AI that is like

01:10:06   hmm I don't know probably I mean it's

01:10:09   it's probably fine maybe I maybe not I

01:10:11   don't know what do you think but like it

01:10:13   hasn't it hasn't been decided yet there's

01:10:15   I haven't even heard any cases coming up

01:10:17   about this to you know someone claiming

01:10:20   that hey because you used AI to generate

01:10:22   that code it's not copyrighted therefore you

01:10:24   don't know therefore blah blah blah blah

01:10:25   hasn't even copied I'm sure it will it

01:10:28   will come up just like we have all those

01:10:30   cases about you know from the New York

01:10:32   Times and Disney and and all these other

01:10:33   companies suing the people suing ChatGPT and

01:10:36   Anthropic and all these other companies for

01:10:38   like passages from their books appearing and

01:10:40   output like that's already happening like the

01:10:42   copyright you know LLM's being trained and

01:10:44   training that but this particular frontier like

01:10:45   I feel like coding agents have not been

01:10:48   around and when wide use long enough for

01:10:50   these laws these big famous lawsuits from

01:10:52   big giant companies to appear but what I

01:10:54   will say is human nature being what it is

01:10:58   and I know this is awful and doesn't sound

01:11:00   good and very many things in our legal system

01:11:03   are you know are this way but human nature

01:11:06   being what it is and the power structures of

01:11:10   the world being what they are as in governments

01:11:13   and rich people and rich companies have lots of

01:11:15   power and other people don't the more big

01:11:20   powerful entities use AI generated code in their

01:11:24   products the more likely it is that it will

01:11:27   magically be deemed to be okay and 100% legal

01:11:30   by the courts because everything in the system

01:11:33   is like there is to your phrase that you were

01:11:36   reaching for before our case you all the genie's

01:11:38   out of the bottle these big powerful companies

01:11:40   want it to be okay and if you if there had been a

01:11:46   court case early on maybe it could have gone

01:11:47   either way but I feel like now no matter what

01:11:50   happened I'm not saying this is even the right

01:11:52   decision I'm just saying this seems like it's gonna

01:11:54   gonna be what happens because so many people are

01:11:56   generating so much code with coding agents that it

01:12:01   will be very difficult for any court to decide and then

01:12:05   be able to enforce the concept that AI generated

01:12:09   code is not copyrightable in any way because how do you

01:12:13   even disentangle it from all the human written code that

01:12:15   it's mixed in with how do you decide what has been AI

01:12:18   assisted like the patent cases versus totally AI generated

01:12:22   how do you decide about human authorship is for code

01:12:25   depending on how long and complicated the prompt is and

01:12:27   it's incredibly like simple human nature it's like the if it

01:12:31   looks right it flies right BS that I've talked about before

01:12:33   about Lockheed planes and stuff it's like big powerful

01:12:38   companies have been doing it too long every day that

01:12:40   passes every day that passes makes it much more likely this

01:12:43   potentially unjust and unjustifiable and unjustified

01:12:47   decision will be become fact and reality because it's just

01:12:52   the easiest thing to do the big powerful companies want it it's

01:12:55   too late now oh well and again I'm not saying that's what I

01:12:59   agree with but that's how I feel whenever I see this topic come

01:13:02   up and everyone who says AI generated code is not copyrightable

01:13:05   you're all doomed if you do this in your apps I'm like maybe but

01:13:09   probably not all right so with that in mind can coding agents

01:13:15   re-license open source through a clean room quote-unquote

01:13:18   implementation of code question mark Simon Wilson writes

01:13:22   Chardet a Python character encoding detector was created by Mark

01:13:26   Pilgrim back in 2006 and released under the LGPL or new lesser public

01:13:31   lesser general public license Mark retired from public internet life in

01:13:35   2011 and Chardet's maintenance was taken over by others most notably

01:13:37   Dan Blanchard who has been responsible for every release since 1.1

01:13:41   in July 2012 two days ago which was I think in early March

01:13:46   Dan released Chardet 7.0.0 with the following note in the release notes

01:13:51   ground up MIT licensed rewrite of Chardet same package name same public

01:13:55   API drop-in replacement for Chardet 5.x or 6.x just way faster and

01:14:00   more accurate yesterday Mark Pilgrim opened issue 327 colon no right to

01:14:05   re-license this project in which he writes first off I would like to thank

01:14:09   the current maintainers and everyone who has contributed to and improved this

01:14:11   project over the years truly a free software success story however it has been

01:14:15   brought to my attention that in the release of 7.0.0 the maintainers claim to

01:14:18   have the right to quote re-license quote the project they have no such right doing

01:14:22   so as an explicit violation of the LGPL license code when modified must be

01:14:27   released under the same LGPL license their claim that it is a complete rewrite is

01:14:31   irrelevant since they had ample exposure to the originally licensed code i.e. this is

01:14:35   not a clean room implementation adding a fancy code generation generator into the

01:14:40   mix does not somehow grant them any additional rights so among all the ways

01:14:44   that AI and LLMs are have the potential to destroy open source here is a new

01:14:48   innovation in that area which is hey what if I take an open source project and I

01:14:53   point an LLM at it and say see this this is a existing project written in

01:14:58   whatever language it's got a test suite it's got a spec it's got documentation

01:15:01   can you write me a new version of that maybe you say to do it in a different

01:15:06   language maybe you say to do it in a new version of a language maybe just say just

01:15:09   do it in the same language or just say here is your here is your here's an existing

01:15:12   thing I want another implementation of that and no lines of code are shared it

01:15:17   writes entirely new code that passes all the test suite in the other thing or

01:15:21   whatever and like that one was GPL license now you make it MIT license GPL tends

01:15:25   to be more restrictive what you can do with it MIT tends to be more laissez-faire

01:15:29   giving people more rights or whatever and like done and done I generated with my

01:15:34   code generator this code which is totally copyrightable as far as I'm concerned

01:15:38   right I have the copyright to this code and no code to share with the old one

01:15:43   don't worry about that that was LGPL or GPL nope the new one is MIT license and in

01:15:48   fact I'm just going this is the new version of that old project so now there's a new

01:15:52   project also called chardette with a new version number that has a different license

01:15:56   it doesn't share a single line of the code with the old one it is a quote-unquote clean

01:16:00   room re-implementation and I'm done um and everyone in open source is like wait what

01:16:06   like is that that's not a thing is it and they start arguing based on the only precedence

01:16:12   they have which is like when they talk about clean room re-implementation it's uh the most

01:16:15   famous one in our industry is uh I believe it was compact whoever made the first uh IBM pc clone

01:16:20   uh they this is a great story if you ever want to read about old people doing computer stuff um

01:16:26   IBM made the original personal pc IBM pc personal computer and it uh it worked a particular way uh and

01:16:32   a company wanted to make a computer that could run all the same software as the IBM pc so they set up a

01:16:37   team and said we are going to isolate you and we are going to give you a spec of how the thing you

01:16:46   should build should work and that spec is basically like how does the IBM pc work but you're not going

01:16:51   to see an IBM pc we're not going to give you an IBM pc you're not going to look at an IBM pc you're

01:16:55   not going to look at any IBM documentation we're just going to say please build this thing

01:16:58   and you even though we know what an IBM pc is and are looking at it right now you aren't you're just

01:17:04   given a spec and you create a thing that conforms to a spec and when you're done you have a pc clone

01:17:10   that can run IBM pc software but you've never even seen an IBM pc you didn't read their manual you didn't

01:17:15   look at their documentation you don't have the chips in front of you to test or whatever

01:17:18   you've now cloned the pc in a quote-unquote clean room re-implementation and it's clean because the

01:17:25   people who made it didn't have any knowledge of the thing that they were making you know what i mean

01:17:29   and that held up in court and that's why that's why the IBM pc clone exists that's why we're all not

01:17:34   we're all weren't using the IBM computer that's why the whole Wintel duopoly happens it's a whole big

01:17:38   thing but anyway that particular thing legally held muster in court so yep you just made a thing according

01:17:45   to a spec and you didn't copy any of their stuff and this is this is the precedent that people are

01:17:50   citing for hey i pointed my llm at an open source project and said write that but

01:17:55   over here and it did it and lo and behold it passes all the tests and now it's all a clean room

01:18:00   implementation that's what the argument about is in mark bear room coming coming out of uh internet

01:18:04   retirement and saying uh this is not clean room first of all the people who are prompting the llm

01:18:09   were the previous maintainer of this project so they've been exposed to the code constantly and

01:18:13   second of all as many people pointed out in the very long argument thread in this github issue that i

01:18:19   invite you to look at if you want to see it people are like but you know but the the llm got to see all

01:18:24   of the source code i was like well what if the llm just saw the test suite and like well the test suite

01:18:27   is part of the source code too what if i just described the project to it and i made sure the

01:18:31   llm couldn't see that it's like but you can't control what an llm sees by just telling it not to do stuff

01:18:35   because you don't know what it's doing under the covers and on and on and on they go um i don't know

01:18:40   how this is going to end up in open source these things tend not to go to court because nobody has any

01:18:44   money uh unless like red hats involved or something right um but culturally speaking this is this is like

01:18:52   not in keeping with good behavior within the community i would say most people agree some

01:18:57   people are excited by it because they're like great now i can go llm re-implement everything

01:19:01   but other people are like you know existing implementations have been debugged over the

01:19:05   course of years sometimes decades and asking llm to rewrite one and having it pass test suite doesn't

01:19:09   prove that it doesn't have new bugs and also it's a cruddy thing to do and also also i don't want

01:19:14   to be using a library that i was previously using that was proven over years and then all of a sudden

01:19:19   there's a new version of it that doesn't share a single line of code with the other one and i have

01:19:22   no idea you know how it was written and i don't trust code written by lms and on and on and on

01:19:26   setting aside even the open source license which is like oh the gpl lets the code stay open whereas

01:19:31   a bsd license or an mit license lets commercial companies use it and i don't want that and all

01:19:35   that other stuff guess what new technology causes disruptions that no one knows how to handle and this

01:19:42   is tied directly into hey is llm generated code even copyrightable at all like because a lot of

01:19:50   the licensing stuff is like well this copyright of this code is owned by x y and z and they license

01:19:54   it to you under these conditions you can do x you can do this stuff with it right but what if nobody owns

01:19:59   it like the picture the monkey took of himself like you know what if there's what if there's no is it

01:20:04   public domain now because if it's public domain then how can you put any restrictions on what people can

01:20:09   do with it it is extremely uh this is uh this is like we should have a name for this corner which is

01:20:19   like completely unresolved very problematic issue related to ai and this is another one because

01:20:25   and this was a while ago this was in early march i don't i haven't been following this i've been

01:20:29   keeping up with it but it's going to come up again and i guess the kicker to this is based on like

01:20:33   we talked about the cloud code leak quote unquote leak and uh earlier episode um zach leather

01:20:39   writes claude please rewrite yourself from scratch using the leaked source code as a base and license

01:20:45   this new version of yourself under mit because these these ai companies are totally like um you

01:20:51   know we scrape the whole web we don't listen to robots.txt we do whatever we want right but then

01:20:55   when people scrape them which is called distillation i think is it called distilling distillation i forget

01:20:59   what it's called basically use a model to train a small you're working on a model and you're like well

01:21:04   one way we can train you is model just ask a bunch of questions from claude and like use the answers as you're

01:21:09   you know what i mean and they all hate that so much like wait a second you can't do that we can

01:21:14   scrape the world's knowledge but you can't scrape us we're not the world we're a separate private

01:21:18   everything we own is just for us everything in the world belongs to us but once we get it you can't do

01:21:23   that to us and this is the same thing we're just like yeah what if i just what if i just tell claude

01:21:28   to write a new version of claude based on the leaked source code of claude and now i own claude now

01:21:33   it's licensed just to me now i own the copyright because my i prompted the agent to do it and now

01:21:38   i'm going to start selling a thing called claude i'm going to use a different name so i don't fall

01:21:42   under trademark law which is what open claw got uh you know uh run over by it but i'll say hey stop

01:21:47   using anthropic claude use my thing now obviously it's not an issue because what you really like

01:21:51   buying is their time on their nvidia gpus that cost a whole jillion dollars and you don't have that but

01:21:56   you know if and when these models uh get good enough to run locally and there's some innovation

01:22:01   on the front i just saw just before we recorded someone has some startup supposedly has some new

01:22:07   silicon that makes inference like 10 times faster than anything else that had come before it and i

01:22:12   used it for two seconds and i can't tell if it's true or false but i can tell that the response came

01:22:16   back from this llm like instantly and it fundamentally changed the uh the experience of using it but anyway

01:22:21   uh someday something like this could actually be a threat to them and i bet if you did do this

01:22:27   suddenly anthropic would have very strong opinions about whether you now own the copyright to code

01:22:32   that was created with an llm and they would be conflicted because like yeah pay us for claude code

01:22:37   uh you'll totally own the copyright of the code you make unless you ask it to clone anything that we

01:22:43   produce and then you don't own that at all because we own that just fyi yeah i don't know i don't i don't

01:22:47   know how i feel about this but i do feel like it seems pretty gross and i don't think i like

01:22:53   this approach at all especially it's doubly gross that it was the person who was or one of the

01:22:59   maintainers for such a long time like i don't know i how can you say with an honest i mean because he

01:23:05   feels like he owns the product he's been like decades he's been doing this it's like well yeah mark

01:23:09   pilgrim wrote this but that was two decades ago or 15 years ago whatever it was i can't do the math but

01:23:13   like this is my project i can do whatever i want with it but then everyone's saying like even if

01:23:18   that was true which is probably not but even if it were true it's a betrayal of your users like they

01:23:23   don't they were using a thing that they trusted for reasons they thought they understood and now every

01:23:28   single line of it is replaced with entirely new code which is ai generated which is itself fraught

01:23:33   and there's a different license that's not good yeah that's just not good bob like like setting

01:23:39   aside the legalities this is like as they would say back in the day a party foul like you you have

01:23:43   broken the social norms you have done something that the the crowd you are walking in does not accept as

01:23:49   you know good behavior even if it turns out to be legal but the legal questions are also just a

01:23:55   gigantic can of worms like regardless of how the is llm code copyrightable or not thing turns out and

01:24:02   that's gotta get to the courts eventually because the uncertainty is also something that the rich and

01:24:07   powerful companies don't like they want some kind of certainty and so i think that issue will get

01:24:13   forced sooner rather than later but it's a complicated issue it's not obvious what the right thing to do is

01:24:18   like i all these things i people talk about i bet as a listener you would feel more comfortable if we all

01:24:25   could tell you this is exactly how it should be but i personally i don't i don't actually know

01:24:29   like i don't know what the right call is i think they made the right call on copyrightable art i think

01:24:35   that one is more cut and dried because it's the human creation of it is so tied up in it but

01:24:39   code i don't know i could have told like i i agreed with the decision on the oracle like um

01:24:45   java oracle like re-implementing java like that apis aren't copyrightable i agree with that decision

01:24:49   that you could you could should be able to just look at the api and write your own implementation

01:24:52   that's fine um but this one i don't know what the right answer is like i i do know that it's a crappy

01:24:58   thing to do in this context but i don't know about the legality we are sponsored this week by quince

01:25:05   you know here in the northern hemisphere it's basically springtime now and let me tell you the pollen all

01:25:11   over my car can confirm it but that's not the point the point is is that springtime is

01:25:15   one of the times anyway that a lot of us like to go through our wardrobe our closets our dressers

01:25:19   etc and figure out hey what's working and what isn't and maybe some things are too big or maybe

01:25:24   some things are too small and you want to just revamp your whole situation right well regardless

01:25:30   of your gender or your age you should and can check out quince because quince makes high quality

01:25:36   everyday essentials using premium materials and their pricing is great so as an example uh last year

01:25:44   they were kind enough to send me some men's linen pants and they are super duper comfortable i really

01:25:48   enjoy them and i'm getting on the time that i think it's appropriate to wear them now i think i don't

01:25:52   know i'm i'm not great with fashion all i know is if i feel good when i'm wearing it and that's all that

01:25:57   really counts right uh the best part about quince is that their prices are usually between 50 and 60

01:26:01   percent less than similar brands and you might wonder well how that's because quince works directly

01:26:06   with ethical factories and cuts out the middlemen so you're paying for quality not brand markup

01:26:11   everything is designed to last and make getting dressed easy you know recently quince sent me

01:26:16   another batch of stuff including a 100 cotton poplin printed long sleeve button down for declan which he

01:26:22   really enjoys and looks smashing in i have to tell you it's uh really really great and i didn't realize

01:26:28   at first that they have all these different sizes and all these different options they also have uh stuff

01:26:33   for your house they have like pillows and things like that you would be surprised how much quince has and

01:26:38   how really affordable it is so what are you going to do you're going to refresh your wardrobe with

01:26:42   quince go to quince.com slash atp for free shipping and 365 day returns it's now available in canada to

01:26:49   go to q u i n c e.com slash atp for free shipping and 365 day returns that's quince.com slash atp

01:26:56   thank you to quince for sponsoring the show

01:26:59   let's do something ask atp and jason writes when marco was doing his long walks while wearing two

01:27:07   fitness watches were was there any difference in the measurements for example does the heart rate

01:27:11   monitoring look the same are the differences in the estimated calories used up what about distance i

01:27:15   know both have gps tracking but did they both record the same distance yes there are differences and

01:27:21   no they did not record the same of everything so wearing the two watches it's very similar like if

01:27:26   you if you've ever changed apple watches sometimes your calorie counts will be somewhat far off on the

01:27:33   new one and it seems like it seems like calorie estimation is just that it's an estimation i believe

01:27:41   the way calorie counting usually works on fitness watches is they basically extrapolate for like the

01:27:47   heart rate your body measurements like you know your body weight and height your age and the heart rate

01:27:52   and maybe the type of activity and they extrapolate based on that how many calories you are probably

01:27:57   burning based on your heart rate in that scenario i think i could be wrong about that but i think that's

01:28:02   roughly what they're doing and the thing is there's a bunch of estimations involved there and so any of

01:28:08   those any of that data that's different between different watches or any of those assumptions that are like

01:28:14   weighted differently or based on different modeling will result in different calories being reported

01:28:20   generally speaking the calorie counts reported by smart watches should basically be considered

01:28:27   an approximation of the effort you expended relative to other similar workouts that you've done on the same

01:28:35   watch so uh you know like like when i do my my three times a week trainer workouts the calorie counts

01:28:42   there give me a rough approximation of like yeah this was this was a hard one or this was this was a less

01:28:47   intense one based on you know what what i've seen from other workouts with that same watch of that same

01:28:53   type um but the calorie counts between the suunto and the garmin and the apple watch ultra and regular

01:29:00   apple watches they're all different from each other so the calorie counts themselves those are not

01:29:04   absolutes they're all estimates and again they should only be used in terms of relative comparison

01:29:09   between themselves doing the same kinds of activities um for gps that's a little bit different gps has a

01:29:15   real answer and so you can kind of there's more of a more of a quality judgment there and and there is

01:29:22   kind of a a truth to be found there what i have seen is that every smart watch and phone for that matter

01:29:31   reports gps a little bit differently and the main reason i think based on what i know about gps and what i've

01:29:38   learned so far um gps if it's if it's inaccurate if you have like a little bit of a of like a signal

01:29:46   weakness or you know the antennas pick up like a reflection of a signal off a building instead of

01:29:52   the direct signal um what you tend to get with that is kind of a wobble you'll you'll have gps that

01:29:59   seems like maybe you're like it'll it'll look on the map kind of like you're zigzagging back and

01:30:04   forth within the path more than you really did or if it's if it's like city streets it'll show you

01:30:08   crossing the street sometimes and you really didn't cross the street and part of the reason why

01:30:14   i like dual band gps watches which in the apple world is just the apple watch ultra but in in you

01:30:20   know the other fitness watches world they have many more models with dual band gps part of the reason

01:30:25   why dual band gps is better is that they use both bands to try to figure out like what was the more

01:30:32   likely path this person actually took because if there's like a problem or if there's a weakness or

01:30:37   reflection with the signal from one of the bands odds are the other one won't have it quite the same way

01:30:42   and so they can kind of like average them out or try to you know try to filter them in clever ways

01:30:46   um but gps data is something that needs filtering to to really make it make sense like sometimes if

01:30:53   you have like a bad reading it could show it as being 100 feet away or something or 10 feet away

01:30:58   or something and so different watches will also not only have different you know antennas and things

01:31:03   to pick up the signals but they also have different algorithms for how they're filtering that data

01:31:07   and even be even within the apple watch you'll have different apps that have that have different

01:31:12   filtering algorithms because the apple watch delivers you what it thinks is you know accurate

01:31:17   gps data but it also has wobbles and imprecision built into certain certain signals that come in

01:31:22   and different apps will will sample them differently and will will filter them differently so gps is also

01:31:29   an approximation but what 10 so what i have seen is the suunto watch so far seems to be it seems to

01:31:38   under report distance compared to the apple watch ultra by about five to ten percent

01:31:43   and now with gps if you're thinking about like that wobble effect if if you're walking in a straight

01:31:50   line and your gps points kind of show you wobbling left and right a little bit then the wobbled data set

01:31:56   is going to be a little bit longer distance because you're you're burning up some fake distance with

01:32:02   those fake wobbles um so what i tend to think is that with gps measurements on smart watches i think

01:32:08   the shortest measured distance is probably the more accurate one and in this case so far what i'm seeing

01:32:15   every time is that the suunto is a little bit shorter than the apple watch ultra so i think the

01:32:22   suunto is a little bit more accurate based on that my experience with my dog's gps is the opposite of that

01:32:27   where when signal gets crappy or has some kind of interference the the line on like where did you

01:32:33   walk your dog will suddenly have this perfectly straight segment and i know that segment is wrong

01:32:38   like there's actually there was wobble the path does wind but it was like well we had bad signal we got

01:32:42   you here we got you there it's two points we connect him with a line so that could that's the other case

01:32:47   of like it could be you're getting shorter measurements because you can check on pedometer plus plus to see

01:32:51   well i guess you can't but like it'll be good to see the actual data because if you see wobbly line

01:32:55   versus straight line and there's the same number of points fine but if you see wobbly line versus

01:32:59   straight line and the straight one only has two points uh that's not more accurate yeah and that's

01:33:04   probably a smoothing artifact that's probably like if the path was actually shaped wobbly and you know

01:33:09   it's possible that the algorithm that was taking those measurements might have thought this was

01:33:13   probably really a straight line and so they simplify it or maybe they didn't have more than two data

01:33:17   points like maybe they couldn't get satellite signal for these 10 feet and now that's all they've got

01:33:21   is this point and that point and what can they do they got to connect them with a line

01:33:24   yeah that's that's also very possible because like you know with with gps like you have a certain

01:33:29   amount of precision with the signal and like the weaker or or more kind of dirty uh or unreliable the

01:33:36   signal is the less certainty the radios can derive about your actual location you know the so it's also

01:33:43   possible that like between if between different points the kind of circle of doubt around your point

01:33:49   around your location where like they're they think you're somewhere in here but they their precision

01:33:53   has temporarily decreased because of conditions like that can also cause the same kind of wobble so

01:33:59   it it's it's an imprecise thing um and and so again kind of like calorie counting like

01:34:04   it's more of an estimate but gps is is much more precise than calorie counting well you know the

01:34:09   solution underscore would tell you more watches yeah so far i'm on there just average them all together

01:34:14   all right and then steven goes all right what's the difference between marco putting all of his

01:34:19   many servers in a data center versus just putting them in his home business class fiber service has

01:34:24   no data caps so the data center doesn't really have anything over regular business fiber you can have

01:34:28   backup electricity at home so it's not about the data center power based on marco's story it doesn't

01:34:32   seem like the people in data in the data center will physically help him like if he needs to restart a

01:34:37   machine they're not going to walk down the aisle and push the button for him so what makes the data

01:34:40   center hosting more professional than just hosting them at home great question uh and for a while i was

01:34:46   hosting them at home and so i can tell you exactly the difference um so what what steven says is correct

01:34:51   that most most residential fiber services like you know verizon files i have here prohibit running servers

01:34:58   now what does that mean that's that's a very vague thing generally i i think what they mean by that is

01:35:04   using a lot of upstream bandwidth to serve live traffic from people making requests inbound to your ip address

01:35:10   for some kind of open port that you run services off of um in practice that's going to be really hard

01:35:16   to enforce and i've never heard of it actually being enforced against anybody but that is officially

01:35:21   against their terms of service to run servers like that there's a few things that that i think

01:35:26   wouldn't have triggered this for me one is that the usage pattern of these particular servers running

01:35:31   transcripts are mostly downloading data they so they download a bunch of data and they upload a little

01:35:38   data because they download the mp3s and listen to them and transcribe them and they upload the transcripts

01:35:43   which are mostly text so they're small that usage pattern looks like residential usage that kind of just

01:35:49   looks like somebody has is running netflix on the tv um so i don't think that would really you know flag a bunch

01:35:56   of attention to it um now the question is what if you just get business internet service you can get a business

01:36:03   fiber line i have one at the restaurant the main reason why not to do that is business internet

01:36:09   service is just much more expensive for the exact same service like that's it's the standard kind of

01:36:13   enterprise pricing model it's like oh you're a business slash wedding you know whatever it is

01:36:18   these keywords that make things more expensive baby yeah exactly business is one of those and so business

01:36:24   internet service costs two to three times as much in most places for the same amount of bandwidth

01:36:30   but that being said you know if i was if i wanted to run this like on the up and up i could do business

01:36:35   internet there is the question though of redundancy so a data center like they have redundancy built in

01:36:42   they usually have multiple internet connections from multiple providers possibly on multiple physical

01:36:47   routes so like you know if a line gets cut somewhere the entire data center isn't just out of luck for a

01:36:51   while that's certainly a big advantage there on the internet service now let's talk about power

01:36:56   at a restaurant having redundant power having like a big backup generator or something actually in new

01:37:05   york does not make a lot of sense because as i learned from my health code class you are required to

01:37:12   close if you lose power from the utility company for any amount of time like even if it's just a hiccup

01:37:19   if it's like a brief blip i that's fine but like if you if you have like a total power outage like a

01:37:24   sustained power outage you are required to close the restaurant and you know have all the customers

01:37:28   leave yeah but like a generac or whatever would would paper would like make that not an issue right

01:37:34   technically you are correct but it is against the rules the health code rules are if you lose utility

01:37:40   power you must close the restaurant and you can see why they would have that yeah totally like well

01:37:45   you know i have a generator so i don't have to worry about this but of course the generator doesn't run

01:37:49   the whole restaurant just the essential things and what i decide is essential is perfectly fine and so

01:37:54   there's no chance of food borne illness because i decided one thing wasn't actually essential what it

01:37:58   turns out it was like the law is a lot easier to just say look you lose power tough luck maybe they'll

01:38:02   have to update it someday when every place has their own solar panels powering them but for now i

01:38:06   understand this law entirely yeah like that's it yeah you can exactly see why because it's like well are

01:38:10   your are all your free walk-in fridges and stuff on that like you know what about emergency things like

01:38:15   you know are your bathroom exit signs like you know there's tons of stuff where like if you know if a

01:38:20   restaurant just has power to a few key things that might not necessarily be enough for the health

01:38:24   department um so there is actually a pretty strong incentive for restaurants not to have much

01:38:28   redundant power i of course have like ups's on my network equipment just you know for brief blips to

01:38:34   not disrupt things but you know that's so there's not redundant power at most restaurants for good

01:38:39   reasons um at least not long-running redundant power data centers have tons of redundant power

01:38:44   that's kind of the whole thing like they have a huge ridiculous setup of like two different power

01:38:49   sources two two different sets of generators and backups and all these different you know switching

01:38:54   and all these different things it's it's you know it's made for that that's what they're designed for

01:38:58   so in practice the odds of an outage affecting servers i'm running in my house is far higher the odds of an

01:39:07   outage affecting a data center and especially for that outage lasting you know more than a minute or

01:39:11   two um odds are much much better at a data center that you're going to stay online then uh there is

01:39:17   the question of uh the amounts of power and bandwidth needed now i know from setting up groups of six mac

01:39:25   minis in my house as i was like setting up the the computer before i bring them to the data centers

01:39:29   my total usage for the 48 mac minis if i ran them all in my house would need about 1.6 gigabits per

01:39:40   second sustained all the time down that is more than i can get in a home connection i don't think

01:39:47   i have anything above gigabit available to me right now so 1.6 gigabits sustained would be the 48 mac

01:39:52   minis uh downstream when they're at full load also they would use about 2 000 watts then also there is

01:40:01   the the heat and the cooling and just the amount of physical space they need like yeah they're small

01:40:07   but you know my house is not that big that i have a room i could devote to like a full height rack to

01:40:14   run a bunch of mac minis in like that normally if they weren't mac minis there would also be the noise

01:40:19   i know that's not probably an issue for the mac minis but for basically anything that's not a mac

01:40:23   mini noise becomes a factor pretty quickly yeah and so you you start you know you have all this space

01:40:28   you have all the heat you have the bandwidth from the from your business fios connection that is going

01:40:33   to be used like you know if i had a business fios connection to run these on i would probably also

01:40:38   want a separate residential connection to run the rest of my personal stuff on because that it would

01:40:42   take up the whole connection so once you start adding up these these costs you're talking hundreds

01:40:49   of dollars probably if i mean for a two gigabit connection business wise fiber wise you're probably

01:40:54   talking about 500 a month um then you're talking about you know 2 000 watts all the time of power

01:41:00   consumption that's a few hundred bucks probably and by the way that's gonna you know that 2 000 watts

01:41:05   that's like you know roughly one and a half space heaters worth of heat that you know in the

01:41:11   summertime you have to cool that so you're also going to be spending more energy cooling that heat

01:41:17   at you know out of whatever room it's in you start getting into hundreds of dollars per month and you

01:41:22   know almost maybe you know almost a thousand dollars a month it's like well then you could just get the

01:41:26   data center for that um and then finally in terms of you know steven said um uh it sounds like the

01:41:32   people in the data center will not physically help me if something needs to be rebooted that's

01:41:36   actually not true uh they have a service called remote hands i think most data centers have something

01:41:40   like this where um i can pay some hourly rate uh that i can have them go like and talk them through

01:41:46   doing stuff to my rack and they will do it for me um it is i think it's like you know one or two

01:41:53   hundred dollars an hour so it's not something i would want to do willy-nilly but if i really needed

01:41:57   something i could call them and have them go do something for me so all of that that's why data

01:42:03   center hosting makes sense because you could do all these things at home some of them would be

01:42:09   significantly less fault tolerant if they were in your home and ultimately the economics of doing

01:42:15   things once you have more than a couple of these it it starts to not make sense to do it at home

01:42:19   you left off your most important point which is uh uh sort of the opposite of your it's not my fault

01:42:24   it's not my fault but it is my problem well if something goes right if you were hosting this in

01:42:29   your house and something went wrong with your like redundant power supplying or something went wrong

01:42:34   with your internet connection you have to fix that yeah because it's your run you're running the quote

01:42:39   unquote data center out of your house now if it's in the data center it's still your problem but you

01:42:44   can't you physically can't fix it you don't own the data center you can't fix that problem so at the

01:42:49   very least you have the relief of knowing well there's nothing i can do about it it sucks and it is my

01:42:54   problem and not that it really matters that much for transcription because it's not like it's a live

01:42:57   service the customers are using but like the relief of knowing i can't fix the power or cooling problem

01:43:03   to the data center they have to fix that and if they turn out to be really bad at fixing that you could go

01:43:08   to another data center in theory because in theory it's a competitive business whereas if you're hosting

01:43:12   it in your house guess what now you've got another job yeah like what if i'm like on vacation and and do

01:43:18   i have to like call one of my neighbors and say hey can you please go into my house tell them you'll pay

01:43:21   them 200 an hour to be a remote hand yeah right yeah and and go into your house and uh yeah don't

01:43:28   don't go into the other rooms just into the one room that has the mac minis in it all right thank you to

01:43:32   our sponsors this episode quince zapier and lisa and thanks to our members who support us directly

01:43:37   you can join us at atp.fm slash join one of the many perks of membership is atp overtime our weekly bonus

01:43:44   topic this week on overtime we're going to be talking about expansion for macs in a world without

01:43:50   the mac pro what does that look like what are our options here we'll talk about that in overtime you

01:43:54   can join to listen to here at atp.fm slash join thank you so much everybody we'll talk to you next week

01:44:00   now the show is over they didn't even mean to begin because it was accidental

01:44:09   oh it was accidental

01:44:12   john didn't do any research marco and casey wouldn't let him because it was accidental

01:44:19   accidental it was accidental it was accidental and you can find the show notes at atp.fm

01:44:28   and if you're into mastodon you can follow them at c-a-s-e-y-l-i-s-s so that's casey list

01:44:39   all right i bought myself a toy a couple of weeks ago um is it a mac pro no goodness no

01:45:09   who would ever buy who would ever buy a mac pro deliberately right john i would buy one at the

01:45:13   prices right uh no i bought myself um a gl inet toy so gl inet does a lot of things if you know

01:45:21   them you probably know them for their travel routers which i still have one in the tailgate tub i have

01:45:28   a different one that um that i'm using occasionally when i travel in addition to the utr the the unified

01:45:33   travel router uh depending on the situation but this is not a router at all this

01:45:39   is something different what this is is a poe powered kvm so what i realized when i was doing

01:45:49   the work on um building rebuilding the nuc that was previously my channel server in connecticut

01:45:55   and then became the the host of all my docker containers among other things and i put proxmox on it

01:46:00   what i realized when i started that process was i need a physical screen and i need a physical

01:46:06   keyboard and a physical mouse all of these things i have they i have them in the house i have a travel

01:46:12   monitor that i use you know when i go and work remotely which is typically most wednesday mornings

01:46:17   i do that because it's it's much easier to study for atp than it is to write code on you know a 13 inch

01:46:23   mac or 14 inch macbook pro and a like 10 or 12 inch accessory monitor and i have that but it it it's either

01:46:32   usbc or either mini or micro hdmi whatever the small one is i always get it backwards and i have

01:46:38   a like full size to micro cable or whatever it is but that's like a pain in the butt because then i got

01:46:43   to go get it and i usually know where it is but i got to find it in the pile of cables and then you got

01:46:47   to get it and then you got to power the display if you're not powering it over you know usbc if it's

01:46:52   not data and power in one then you got to power it as well and then i have an old actually i think for my

01:46:57   iMac pro i have my keyboard and trackpad from that and so i can get those out but then i need a couple

01:47:03   of lightning cables and i'm starting to run low on those actually because there's almost no devices

01:47:08   that i use in my life that still need them and it's just kind of annoying and burdensome granted first

01:47:13   world problem for sure but burdensome nonetheless now my friend alex from tailscale a past sponsor i think

01:47:20   future sponsor if i'm not mistaken um on the tailscale youtube channel he had pointed out that glinet

01:47:26   makes the comet poe which is uh or comet poe remote kvm control so what this is is a little box and it

01:47:34   has you know an ethernet jack a usc two usbc ports an hdmi port and a usb a port and what you do is

01:47:42   you plug in um either either usbc for power or if you have poe available power over ethernet available

01:47:50   which i do then you just plug in an ethernet cable that gives you power network connect connectivity

01:47:55   you have a usbc cable that they provide that that can go into usba or that does go to usba and that

01:48:03   provides both keyboard and mouse across one cable and then you connect hdmi from the device into this

01:48:09   box and you can optionally add uh some some sort of like external uh disk drive on the usba port

01:48:17   but what this does is it gives me network attached quote-unquote physical access to a computer

01:48:23   because as far as the computer that i'm using is concerned this is a physical monitor it's a physical

01:48:29   keyboard it's a physical mouse it's just that this fancy little box happens to be beaming it across the

01:48:34   internet and the reason i saw it on the tailscale youtube channel is because it will natively jump on

01:48:40   your tail net which is super freaking cool so what that means is i have a couple of screenshots that i

01:48:47   will share both in the chat and in the show notes and everywhere else but here's a screenshot of me

01:48:52   using the proxmox nuk uh from this glinet thing and you can see that i'm i just went to a url on my

01:49:00   tail net and i have a terminal window where i whiffed the password a couple times but that's neither here

01:49:05   nor there but anyways i have a terminal window and so that's pretty cool but it's not just for linux

01:49:12   machines if you want you can connect it to i don't know a mac and so here's another screenshot of me

01:49:18   using it with my mac mini that runs uh plex and jelly jellyfin and channels and so this is a really silly

01:49:27   little contraption that generally speaking will live in a box but on the occasions i need to connect physically

01:49:34   to something this is going to be so much nicer and so much easier than dragging out several you know

01:49:38   17 different cables um being hyperbolic but you know 17 cables a monitor a monitor power uh you know

01:49:44   that one random hd micro or mini or whatever hdmi to full size hdmi cable it's all just in this one

01:49:51   little box and then i use it from whatever computer i want i could use it from an ipad if i wanted and

01:49:56   it's great and if i wanted to install software on whatever this is connected to i can uh you know like

01:50:02   upload an iso to the kvm and have the iso masquerade as a usb key connected to the device

01:50:08   you know to the computer i've never tried this but i believe that to be the case so

01:50:11   i don't know there's not much to be said here other than that i'm happy to answer questions if you have

01:50:15   any but it's just this is one of those neat little things that did i need it absolutely not

01:50:19   and it happened to be on sale whatever that sale was a couple of weeks ago on amazon i forget what

01:50:24   specifically it was but they're always inventing something yeah they're always inventing some

01:50:27   reason have a sale and so it was a little under 100 bucks and i think this is a device that should

01:50:32   cost a little under 100 bucks and it's really neat to have and i really like it and marco if you had

01:50:38   the need to have quote-unquote physical access to only one of your 48 mac minis i highly suggest this

01:50:45   thing i would not necessarily put this on all 48 of them but if you if you had one that was like a

01:50:50   controller or something for all the others this would be a really great way to have what did you call

01:50:55   like remote hands or helping hands this would be a pretty good way to do it because again

01:50:59   as far as the physical computer is concerned this is a physical keyboard a physical mouse and a physical

01:51:04   monitor it just so happens that they're not those things and they're actually all being presented on

01:51:09   the internet that's pretty cool i i can actually see like i actually might at some point get one of

01:51:15   these two like right now the way my servers are set up they there there is no one controlling server

01:51:21   like they all are peers and they all just connect to the web service and get jobs and work on them

01:51:28   and submit them back to the web service cattle not pets right and and there are like and like right now

01:51:34   like you know what one thing i've thought about is should i at some point like put a some some other

01:51:41   kind of server in the data center in the rack with the other ones on their network to to do kind of

01:51:47   controlling things like right now uh my uh my power the ats the transfer switch my power switch thing

01:51:53   i have one that has remote control over the port so like right now if i needed to reboot the mac minis

01:51:59   like one of the mac minis i mean he's a power cycle one and i can't do it remotely via like you know ssh or

01:52:04   remote desktop i can log into one of the other ones as long as any of them are working i can log into one

01:52:11   of them and open up like via remote desktop open up like the local browser interface to the power unit

01:52:18   and reboot the entire bank of six that whatever the problem one is plugged into so i still have the remote

01:52:27   hands problem of like i you can tell me maybe you'll discover this or maybe a mac and then we'll tell you

01:52:31   but i think there's still some stuff where you need to press the physical button on the mac mini like the

01:52:37   enter recovery mode and stuff and this kvm will not hold down the power button on your mac mini

01:52:42   for me i mean that's true if i was to the point where i had to enter recovery mode i would just drive

01:52:47   to the data center and like pick it up bring it back to my house and remote hands deal with but yeah but

01:52:51   what but what am i in recovery mode for yeah i don't know i just say like it this this the when you have

01:52:56   actual like data center hardware they try to make it so that you can do everything remotely or the very

01:53:01   at least do everything from like a console in the data center but with macs i'm again maybe mac admins

01:53:06   can correct me it just seems to me that there's probably some remaining stuff like that you that

01:53:10   you need in the apple silicon age you still need to hold down the power button physically and the kvm

01:53:15   can't do that unlike the good old days where the power button used to be on the keyboard i'm not sure

01:53:18   you guys were mac users back then but it was cool i'm familiar with this as a thing but no i was not a mac

01:53:24   user then also it's worth noting marco that this this thing does have like uh the web ui has uh like

01:53:31   a little toolbox they call it and so this is where you can like paste something and then into the web

01:53:36   ui and then tell tell the web ui okay basically type this on the remote machine right so it's kind of

01:53:42   like a like a very it's a very rudimentary but functional uh clipboard sharing right uh well another

01:53:48   thing it has is a wake on land section so you can add like you know mac addresses and name them

01:53:54   and have the kvm fire awake on land packet over to whatever device you need is that useful for you i

01:54:00   don't know but it's neat that it's a thing uh and so yeah i really really like this little box it's the

01:54:07   sort of thing that i certainly didn't need i used birthday money for it so it was the perfect birthday

01:54:11   gift um and again like i don't know if i would want to watch a full screen video on it like i mean

01:54:16   using the mac on it it's pretty freaking quick well especially when it's in the same house but

01:54:21   it's pretty freaking quick but the point is i typically plan to use this for situations where

01:54:27   i just need to plug something into a physical screen and this is probably going to be the most convenient

01:54:33   way to do it because as long as i have poe or you know a usbc input from something even like a battery

01:54:39   pack as long as i have an ethernet connection which i have in several places in the house this is such an

01:54:45   easy way to get what is quote-unquote physical access to a computer and i really like it and

01:54:50   the fact that it's on my tail net makes it super nice because you know i was thinking in the scenario

01:54:56   where i'm like going out of town on vacation or something like that i can put this on either the

01:54:59   mac mini or the the proxmox box box box box box uh anyways i put this on one of those and you know

01:55:06   if i have some sort of issue that would typically require physical contact with the machine i can just

01:55:10   log into this thing and perhaps work whatever magic i wouldn't be able to otherwise so uh yeah

01:55:15   maybe i'm the only one who wants it maybe this is not appealing to anyone else and and like you don't

01:55:19   have to use this with tail scale by the way it's just nice that it's tail scale native so you don't

01:55:23   have to think about it but uh really good stuff and i'll put a link to an affiliate link to be honest

01:55:28   in the show notes and i'll put a link to alex's video which covers it in about 10 minutes but

01:55:31   i thought it was cool

01:55:32   Beep, beep, beep.