ATP

00:00:00 ◼   ► I had a terrible thought uh a few hours ago, and maybe it's it's

00:00:05 ◼   ► been kind of recurring recently.

00:00:08 ◼   ► Wait, may I guess?

00:00:09 ◼   ► It's not a boat.

00:00:10 ◼   ► No, that's a that's a very, very good guess.

00:00:13 ◼   ► Um is is the terrible thought, hey, maybe I should drag Casey to a

00:00:16 ◼   ► fish concert sometime.

00:00:17 ◼   ► Oh no, that's that's not even in question.

00:00:19 ◼   ► There's no question I definitely should do that.

00:00:21 ◼   ► Um no, the the even more terrible thought is I think I have a

00:00:26 ◼   ► problem that could be solved by home bridge or home assistant rather, which one's

00:00:31 ◼   ► the woman that the box is.

00:00:32 ◼   ► Oh yes.

00:00:33 ◼   ► Oh my time is come.

00:00:35 ◼   ► This is Casey's episode for sure.

00:00:37 ◼   ► Yeah, everything's coming up Mill House tonight.

00:00:39 ◼   ► This is great.

00:00:40 ◼   ► It's assistant, right?

00:00:40 ◼   ► That's like the more the even more complicated one.

00:00:42 ◼   ► Yeah, so all kidding aside.

00:00:44 ◼   ► So home bridge is basically let's be as thin a layer as we can.

00:00:49 ◼   ► That's simply a facade to get things that aren't natively home kit into home

00:00:53 ◼   ► kit.

00:00:53 ◼   ► Whereas home assistant is more of a let's boil the ocean, take over the

00:00:57 ◼   ► world sort of scenario.

00:00:59 ◼   ► I I hate to do like I know this is a rabbit hole I

00:01:02 ◼   ► should not fall into.

00:01:03 ◼   ► Oh, but it's great.

00:01:04 ◼   ► Everyone says that.

00:01:07 ◼   ► Well, so here's the thing, right?

00:01:08 ◼   ► Like all, you know, again, all kidding aside, like home assistant is one of

00:01:12 ◼   ► those things that you have to be able to see through the matrix, and

00:01:16 ◼   ► it is brutal to get to the point that you can see through the

00:01:20 ◼   ► matrix.

00:01:20 ◼   ► Like it is a learn, and it's gotten way better since I started using

00:01:23 ◼   ► it, I don't know, like two years ago, but it's bad.

00:01:27 ◼   ► Like getting uh getting your head wrapped around the way home assistant thinks of

00:01:30 ◼   ► the world is not fun.

00:01:32 ◼   ► Now, that being said, you have plenty of tutors all around you that would

00:01:35 ◼   ► be more than happy to help you.

00:01:36 ◼   ► Hello.

00:01:36 ◼   ► Uh, and I'm not the only one for the record, but um, it is

00:01:39 ◼   ► not fun to start down the path and start getting things squared away.

00:01:45 ◼   ► But if you can stick with it and give it a chance, then suddenly

00:01:50 ◼   ► the world is opened up to you, and almost anything you can conceive can

00:01:54 ◼   ► be done.

00:01:55 ◼   ► It's really astonishing.

00:01:56 ◼   ► What are you looking to accomplish?

00:01:58 ◼   ► Um I have so I I finally got the beach heating and air conditioning

00:02:03 ◼   ► finally all fixed.

00:02:03 ◼   ► It's all tell me more.

00:02:05 ◼   ► All new systems put in.

00:02:07 ◼   ► It's it's a Dakon Daiken.

00:02:09 ◼   ► Everything by them, it's just like a whole brand new everything.

00:02:13 ◼   ► Everything had to be replaced, but it works, which is great.

00:02:16 ◼   ► That was a very long, very expensive process that has finally ended.

00:02:21 ◼   ► Now, did you go with something that is like comparatively dumber?

00:02:24 ◼   ► Because if I recall correctly, didn't you have like a very, very, very fancy

00:02:27 ◼   ► system, and that was potentially part of the problem?

00:02:29 ◼   ► Yeah, like the like it was it was originally this Mitsubishi system that it

00:02:34 ◼   ► was very it was you know, it's like a high-end, like high efficiency heat

00:02:37 ◼   ► pump everything.

00:02:38 ◼   ► Um and it had all these all these like advanced functionality things that we

00:02:41 ◼   ► would probably never use.

00:02:43 ◼   ► Um, and you know, it was just like it was like what the what

00:02:45 ◼   ► the builder had spec'd for the house uh years ago.

00:02:49 ◼   ► And it's it it was very complicated to the point where nobody out here

00:02:53 ◼   ► on Fire Island knew how the heck to even service it.

00:02:55 ◼   ► So they did a very bad job of servicing it and eventually broke it.

00:02:59 ◼   ► Um to it to a point that even once I did finally get people

00:03:02 ◼   ► out of here who knew what they were doing with these systems, they were

00:03:05 ◼   ► like, yeah, we can't save this.

00:03:06 ◼   ► It's like too much damage has been done.

00:03:09 ◼   ► It would cost uh way more than it's worth to try to disassemble every

00:03:14 ◼   ► part of everything and clean it and all sorts of ridiculous things.

00:03:17 ◼   ► Um so anyway, I replaced it with a Daikan system because they are what

00:03:22 ◼   ► everyone out here uses, which means that all the people out here know how

00:03:26 ◼   ► to install and service them correctly.

00:03:28 ◼   ► Um I don't really honestly, care if it's better or worse.

00:03:31 ◼   ► Like that's it is it is more practical to own here.

00:03:35 ◼   ► Um, and frankly, it was way cheaper than the Mid Speech system would have

00:03:38 ◼   ► been to replace.

00:03:39 ◼   ► Um, so that that helps because out here nothing lasts, so when it does

00:03:42 ◼   ► die, it'll be cheaper to replace.

00:03:44 ◼   ► So, anyway, it it works pretty well.

00:03:47 ◼   ► The the only downside is that it, of course, has its own proprietary thermostats,

00:03:52 ◼   ► and they as we know from the service guys like you definitely do not

00:03:58 ◼   ► want to have an HVAC guy see you using a Nestor and Echo B

00:04:02 ◼   ► because they hate them and they blame them for every possible problem under the

00:04:05 ◼   ► sun.

00:04:06 ◼   ► Like, I don't know what smart thermostats did to HVAC people, but they so

00:04:12 ◼   ► irrationally hate them, it's kind of comical.

00:04:15 ◼   ► Um, so anyway, I'm using their stupid thermostats.

00:04:18 ◼   ► It's fine.

00:04:19 ◼   ► And they have like Wi-Fi connectivity with their app, and their app is, as

00:04:23 ◼   ► you would expect, total trash.

00:04:25 ◼   ► And of course, as you would expect, it integrates with nothing.

00:04:29 ◼   ► There is no support for any of the home platform integrations.

00:04:33 ◼   ► Of course.

00:04:34 ◼   ► So, anyway, uh that's that's where I started looking at hmm.

00:04:39 ◼   ► Maybe home bridge or home assistant might be a tool to use here because

00:04:44 ◼   ► every time I want to either turn off the air conditioning, turn on the

00:04:48 ◼   ► air conditioning, or just change the temperature of the house.

00:04:51 ◼   ► There's four different thermostats for all the different split zones, and so I have

00:04:56 ◼   ► to go into the app and change four different zones.

00:04:58 ◼   ► And oh, by the way, the app requires face ID.

00:05:01 ◼   ► Great, thank you for keeping me secure.

00:05:04 ◼   ► Um it's just kind of it's a very cumbersome process.

00:05:07 ◼   ► And and the app is not good, and so you know, when I do

00:05:10 ◼   ► change temperature, it sometimes it doesn't stick, sometimes it like changes, and then it

00:05:14 ◼   ► pops back to where it was.

00:05:15 ◼   ► I guess something refreshed and it clobbered itself in a race condition or something.

00:05:18 ◼   ► So it's not a good app, is what I'm saying.

00:05:21 ◼   ► So anyway, I would love just something that I could automate and like you

00:05:25 ◼   ► know, go yell at some voice assistant somewhere, set the house to 74.

00:05:29 ◼   ► And it just does it.

00:05:30 ◼   ► Right.

00:05:29 ◼   ► So that's kind of that's what got me here.

00:05:33 ◼   ► So while you were talking, I looked to see if there is an integration,

00:05:37 ◼   ► which is home assistant speak for, you know, is there is there a software

00:05:41 ◼   ► package?

00:05:42 ◼   ► It's very easy to add an integration.

00:05:44 ◼   ► That part is very straightforward.

00:05:45 ◼   ► But anyways, there's a software package for the Dyk and Dakon, whatever they're called.

00:05:48 ◼   ► And sure enough, yes, there is.

00:05:50 ◼   ► Um, and so there this would be in principle anyway, fairly easy for you

00:05:55 ◼   ► to integrate.

00:05:55 ◼   ► Now, the only thing I will say is that um shouting at something, the

00:06:00 ◼   ► way you would want to do that is by exposing this to home kit,

00:06:03 ◼   ► which is very straightforward.

00:06:04 ◼   ► But I don't want you to start down the path of home assistant and

00:06:07 ◼   ► thinking it.

00:06:07 ◼   ► I mean, home assistant does have its own voice stuff, but my understanding is

00:06:11 ◼   ► it's fine.

00:06:12 ◼   ► It's the kids would call it mid.

00:06:14 ◼   ► Um, but anyways, uh, you could expose all this through home assist excuse me

00:06:18 ◼   ► through home kit, and then you'd be good to go.

00:06:19 ◼   ► Yeah, that's all I need to do.

00:06:21 ◼   ► This is this is a solvable problem, and I don't think this would be

00:06:24 ◼   ► that particularly difficult.

00:06:26 ◼   ► I think the the question is how would you get home assistant in your

00:06:29 ◼   ► house?

00:06:29 ◼   ► And the obvious answer for that is one of the physical hardware products that

00:06:33 ◼   ► they sell.

00:06:33 ◼   ► I believe that's what Jason Snell is using now.

00:06:35 ◼   ► I think he's using a home assistant green, I want to say.

00:06:38 ◼   ► I forget the difference.

00:06:39 ◼   ► Yeah, I saw that, it's like 200 bucks.

00:06:41 ◼   ► I mean, I also I have a few Mac minis, if you haven't heard,

00:06:44 ◼   ► I could so if you wanted to go that route, and I'm not snarking

00:06:48 ◼   ► with you at all.

00:06:49 ◼   ► If you want to go that route, I think that is very doable, but

00:06:51 ◼   ► then you're gonna want to go down the path of like Docker.

00:06:53 ◼   ► Now you're adding a whole new layer of complexity that you probably don't want

00:06:56 ◼   ► to futz with.

00:06:57 ◼   ► And honestly, I would just if especially if I'm spending your money, just spend

00:07:00 ◼   ► the $200 on home assistant green or whatever, again, whatever the right one is.

00:07:03 ◼   ► I I don't I don't know their hardware well.

00:07:06 ◼   ► I just know that it tends to be very well regarded and work really

00:07:09 ◼   ► well.

00:07:10 ◼   ► But I think there's like a green and then there's some other color, different

00:07:13 ◼   ► version, and again, I I can't speak to the specifics about these.

00:07:16 ◼   ► Um, but it should be fairly straightforward to figure it out.

00:07:18 ◼   ► All right, thanks.

00:07:14 ◼   ► Maybe uh maybe I'll try that out.

00:07:20 ◼   ► Because yeah, I I do I do want something that is relatively low FUTS.

00:07:25 ◼   ► Um and yeah, that that is kind of appealing.

00:07:28 ◼   ► So we'll see.

00:07:28 ◼   ► Although, you know, you I say that now, so talking about setting up a

00:07:31 ◼   ► home assistant thing.

00:07:32 ◼   ► So you know, it's all relative, but it's all relative.

00:07:35 ◼   ► But I I think also like, you know, now like I I assume with

00:07:38 ◼   ► like modern AI stuff that should make it easier if I run into configuration

00:07:42 ◼   ► uh questions and you know, so uh maybe this is a decent time to

00:07:46 ◼   ► try it out.

00:07:47 ◼   ► I don't know.

00:07:48 ◼   ► Yeah, I mean I think I think this is a very worthwhile adventure for

00:07:52 ◼   ► you.

00:07:52 ◼   ► It'll make for some good content, selfishly speaking.

00:07:54 ◼   ► So I'm in on board in that regard.

00:07:56 ◼   ► Um, but no, I mean a home assistant green or some other equivalent hardware

00:08:00 ◼   ► dingus, uh, you basically plug it into Ethernet and then you open up a

00:08:04 ◼   ► web browser and follow the steps from there.

00:08:06 ◼   ► And you you are more than capable of doing this, but I would be

00:08:09 ◼   ► happy to hold your hand if you needed the help.

00:08:11 ◼   ► So yeah, if you go down that path, let me know.

00:08:13 ◼   ► And one of the nice things is I mean, you don't have to have

00:08:15 ◼   ► home assistant like infiltrate your entire home.

00:08:18 ◼   ► You know, you could just add the Diken integration and ignore basically any of

00:08:24 ◼   ► the things that it auto-discovers and then expose the Diken thermostats to home kit,

00:08:28 ◼   ► and then you'd be good to go.

00:08:29 ◼   ► Um, but you know, you could also take it take it to the next

00:08:32 ◼   ► step or the next level where you say, okay, if I'm adjusting one thermostat,

00:08:36 ◼   ► just automatically adjust all the others to match, you know, or or something along

00:08:39 ◼   ► those lines.

00:08:39 ◼   ► Um you can g do some really bananas automations if you so desire.

00:08:43 ◼   ► You if you if you know when the people are in the house, if

00:08:45 ◼   ► you know when the family's in the house, you could say, you know, if

00:08:47 ◼   ► they leave the house, change the the set points to such and such.

00:08:50 ◼   ► And when any of us return, change it back, you know, or you can

00:08:53 ◼   ► do some of that stuff that Nest used to do kind of by hand,

00:08:56 ◼   ► if that's appealing to you.

00:08:57 ◼   ► You know, in within like, you know, a week of having I'm gonna be

00:09:00 ◼   ► I'm gonna set up some automation where, like, when I'm on the ferry, turn

00:08:59 ◼   ► on the air conditioning so that that way it's like when I get here,

00:09:06 ◼   ► like you know I'm gonna do stuff like that.

00:09:08 ◼   ► I mean, this is this is a solvable problem for sure.

00:09:10 ◼   ► Um, you can do some of this absolute nut job stuff.

00:09:13 ◼   ► And that's why home assistant, for all of its drawbacks, it's very fiddly.

00:09:18 ◼   ► Well, it's fiddly to set up.

00:09:20 ◼   ► Once it's set up, it tends to be pretty bulletproof, but you have to

00:09:22 ◼   ► get there first, right?

00:09:23 ◼   ► Uh it's kind of like Linux on the desktop, am I right?

00:09:25 ◼   ► Uh but anyway, it's fiddly to set up.

00:09:28 ◼   ► It's it's hard to wrap your mind around at first, but again, they've gotten

00:09:31 ◼   ► it much better than it used to be.

00:09:32 ◼   ► Uh, but once you can see the matrix, if you will.

00:09:36 ◼   ► Oh my gosh, so many problems are suddenly solved by this stupid software, and

00:09:40 ◼   ► it's incredible.

00:09:41 ◼   ► So uh yeah, uh, let me know if you go down this path.

00:09:43 ◼   ► I think it'll make for some very fun and very uh enjoyable content.

00:09:46 ◼   ► Oh, it was nice knowing, yeah.

00:09:49 ◼   ► All right, let's do some follow-up and let's start with the open AI donut

00:09:52 ◼   ► hardware.

00:09:53 ◼   ► Elijah Yap writes, this theory might fit the description.

00:09:56 ◼   ► It's an eye.

00:09:57 ◼   ► The moving parts could be the eyelids protecting a center screen, and eyes convey

00:10:00 ◼   ► a lot of possible emotions.

00:10:01 ◼   ► And Elijah provided a lot of emotions via emoji, which uh gave me a

00:10:05 ◼   ► good laugh.

00:10:06 ◼   ► For a more vivid depiction, Google's Google Chrome's Japan marketing team has a whole

00:10:11 ◼   ► campaign of the mascot doing cute things.

00:10:13 ◼   ► That's interesting.

00:10:15 ◼   ► I really hope it's not, because I just immediately picture like the eye of

00:10:19 ◼   ► Sauron from Lord of the Rings.

00:10:20 ◼   ► Like, there's already a lot of like sort of creepy creepiness factor to AI.

00:10:25 ◼   ► So you're gonna make a little Cyclops donut with an eye in the middle.

00:10:28 ◼   ► It's supposed to be endearing and make you feel good about it.

00:10:31 ◼   ► I'll make you feel like you're constantly being watched.

00:10:33 ◼   ► The eye sees all.

00:10:35 ◼   ► I mean, that's fair.

00:10:36 ◼   ► I went to uh Flight of the Navigator personally, but that's just me.

00:10:39 ◼   ► Uh Chris Lawrence writes: In overtime, John was speculating that the donut-shaped open AI

00:10:43 ◼   ► device was red blood cell-shaped.

00:10:45 ◼   ► The term for this shape for the shape of a red blood cell was

00:10:44 ◼   ► drilled on into us in medical school.

00:10:44 ◼   ► A biconcave disc.

00:10:52 ◼   ► I can totally see this being used in a Johnny Ive press release.

00:10:55 ◼   ► Yeah, that's exactly the shape I was thinking of.

00:10:56 ◼   ► Look how it would be fun to like grip that in the middle.

00:10:58 ◼   ► Like I've got some glasses from like Crate and Barrel that have they're like

00:11:01 ◼   ► just glass cylinders, but they have little dents in the in them, like four

00:11:05 ◼   ► dents around them for you to put your fingers in, and it's very pleasing

00:11:07 ◼   ► to grab the glasses that way, you get a more secure grip.

00:11:10 ◼   ► So if it doesn't have a hole, it could be a biconcave disk, and

00:11:13 ◼   ► that would be cool.

00:11:14 ◼   ► But you know, that kind of uh information is just not important enough for

00:11:18 ◼   ► Gruber to tell us.

00:11:19 ◼   ► Not Gruber, sorry.

00:11:20 ◼   ► Just not important enough for Mark Erman to tell us.

00:11:24 ◼   ► We'll get to Gruber later.

00:11:26 ◼   ► Indeed.

00:11:27 ◼   ► All right.

00:11:27 ◼   ► With regard to backups, an anonymous contributor writes during episode 740, you talked about

00:11:32 ◼   ► the Mac backups and the difficulty of getting some non-technical people to plug in

00:11:36 ◼   ► a backup drive on a regular basis, if ever.

00:11:38 ◼   ► Steve Jobs presented a solution to this issue at Macworld 2008 with the introduction

00:11:42 ◼   ► of Time Capsule, which allowed for wireless backups.

00:11:45 ◼   ► Reason number a million why it was kind of not a great idea for

00:11:48 ◼   ► Apple to depart the uh Wi-Fi base station business.

00:11:52 ◼   ► I mean, uh, it was basically like a very simple no options Apple flavored

00:11:58 ◼   ► NAS, basically.

00:11:59 ◼   ► It was network attached storage, but coming in through the side door, because it's

00:12:03 ◼   ► like, uh, you know, we sell Wi-Fi routers and this one's got a hard

00:12:06 ◼   ► drive in it, isn't that neat?

00:12:07 ◼   ► Oh, and by the way, your Mac can see it as a time machine

00:12:09 ◼   ► destination.

00:12:10 ◼   ► It was like getting people to get network attached storage without telling them that

00:12:14 ◼   ► it was network attached storage, and it had that Apple quality of look, if

00:12:17 ◼   ► you buy this thing and you plug it in, all your Apple stuff know

00:12:21 ◼   ► about it, and it just does, you know, it just becomes an option for

00:12:24 ◼   ► you.

00:12:24 ◼   ► You're you know, I think your Mac would even prompt you and say, Hey,

00:12:26 ◼   ► I see this thing that's on the network.

00:12:28 ◼   ► Do you want me to back up this Mac to that?

00:12:29 ◼   ► And you'd be like, okay, that's great for laptops.

00:12:31 ◼   ► You don't have to remember to plug anything in.

00:12:29 ◼   ► They do it while they're idle.

00:12:34 ◼   ► Like, boy, wouldn't that be great, Apple to have that kind of solution?

00:12:37 ◼   ► Or, you know, do iCloud backup for Macs.

00:12:39 ◼   ► But any of those things would be nice, but instead we get none of

00:12:41 ◼   ► them.

00:12:42 ◼   ► Womp womp.

00:12:44 ◼   ► All right, with regard to multiple Apple IDs with the same phone number.

00:12:47 ◼   ► Uh, Adam One writes, For many years I would set up separate Apple accounts

00:12:50 ◼   ► for my smaller customers.

00:12:51 ◼   ► During setup, I would attach the same customer designated phone number to each account.

00:12:55 ◼   ► About two years ago, Apple seemingly tightened security, and I found that I could

00:12:59 ◼   ► no longer reliably use this method.

00:13:00 ◼   ► In one case, I attempted to apply this method to just three Apple accounts

00:13:03 ◼   ► and was blocked.

00:13:04 ◼   ► The solution required me to call Apple and speak with their security team to

00:13:07 ◼   ► unblock the attempted create account creation using the same phone number.

00:13:10 ◼   ► The gentleman I spoke with was very courteous and lifted the block once I

00:13:13 ◼   ► explained the circumstances.

00:13:14 ◼   ► He then explained that I triggered some automatic security safeguards.

00:13:17 ◼   ► He also gently suggested I should be setting up small companies with Apple business

00:13:20 ◼   ► accounts, which is a strategy I've since adopted.

00:13:23 ◼   ► Before I let the gentleman go, I asked him whether it was even still

00:13:25 ◼   ► possible to use one phone number on more than one account.

00:13:27 ◼   ► He said yes, but a number of unstated factors went into the automated system's

00:13:31 ◼   ► decision-making process.

00:13:32 ◼   ► I asked if it was possible to make three to five accounts under the

00:13:35 ◼   ► best case circumstances.

00:13:36 ◼   ► He answered the three to five was, quote, a good rule of thumb.

00:13:39 ◼   ► I've since tested that hypothesis and found it generally to be true.

00:13:42 ◼   ► In many cases, if I make an Apple ID directly on the device rather

00:13:45 ◼   ► than at the accounts.

00:13:52 ◼   ► One fun fact, I have one customer set up accounts years ago and put

00:13:55 ◼   ► 20 Apple accounts onto one phone number and 15 on another.

00:13:58 ◼   ► Those accounts still work fine today with those two phone numbers.

00:14:01 ◼   ► Man, those were the good old days.

00:14:03 ◼   ► Yeah, Apple tends not to give you details about its various security systems.

00:14:08 ◼   ► And uh the thing that makes this makes me think of is my issue

00:14:12 ◼   ► with getting my Apple ID locked.

00:14:14 ◼   ► I I've complained about that on the show before.

00:14:16 ◼   ► It was a really bad at one point, and then it got less bad.

00:14:14 ◼   ► I don't know why.

00:14:20 ◼   ► Uh recently it's been getting bad again.

00:14:23 ◼   ► Um I just assume everyone who has this happen to them, it's not just

00:14:27 ◼   ► me, assumes like, oh, someone's trying to get into my Apple ID.

00:14:30 ◼   ► You know what I mean?

00:14:31 ◼   ► Like, and like putting in bad passwords, and then you're getting it locked out

00:14:33 ◼   ► of your account.

00:14:34 ◼   ► And when this happens, basically Apple says, your Apple ID is locked, and you

00:14:37 ◼   ► have a button that says unlock account, and it makes you go through a

00:14:40 ◼   ► bunch of stuff and enter in your phone number and it sends you an

00:14:43 ◼   ► SMS, and you gotta do two factor and you gotta enter your password.

00:14:46 ◼   ► And uh back in the day, it also used to try to herd you

00:14:49 ◼   ► towards re changing your password, which I would never do.

00:14:52 ◼   ► You didn't have to do, you could unlock the client with account without changing

00:14:54 ◼   ► your password, but it would send you there and try to make you think

00:14:57 ◼   ► you have to re change your password.

00:14:59 ◼   ► Anyway, I'm kind of sick of doing it.

00:15:01 ◼   ► And then at a certain point, um, it started saying uh I started getting

00:15:05 ◼   ► 503 errors, HTTP 503 responses.

00:15:07 ◼   ► I forget what that status code stands for.

00:15:10 ◼   ► Someone can look it up while I'm talking.

00:15:11 ◼   ► But um basically, like your the HTT request of the unlock your apple ID

00:15:15 ◼   ► website, which is Apple's website to be clear, it's like something on Apple.com.

00:15:19 ◼   ► It would throw 503 errors, and you get this little uh Apple HTML dialog

00:15:24 ◼   ► box saying sorry, couldn't complete that request.

00:15:27 ◼   ► And it's like a seven step process to unlock your Apple ID, and every

00:15:30 ◼   ► HTTP request has to go through.

00:15:32 ◼   ► So you could be on a page and you'd watch, like in the in

00:15:34 ◼   ► the dev tools, you'd watch it firing off all these, you know, uh Ajax

00:15:39 ◼   ► requests essentially, uh fetch requests, whatever, behind the scenes, and you'd see them going

00:15:43 ◼   ► through and it'd be like 200, 200, 200, up 503, and you get the

00:15:46 ◼   ► dialogue.

00:15:46 ◼   ► So you gotta you've gotta have all of the behind the scenes uh HTTP

00:15:50 ◼   ► requests work before you can get to the next page.

00:15:53 ◼   ► And my experience of of it was like that it was throttling, like that

00:15:57 ◼   ► I could have a cert because I would I would just leave the window

00:15:59 ◼   ► there for five minutes and then come back to it, and then I could

00:16:01 ◼   ► get to the next page, but then to get to the page after that,

00:16:04 ◼   ► it would it wouldn't work again.

00:15:59 ◼   ► So I just leave it for five minutes.

00:16:06 ◼   ► You can't leave it too long because the session will time out, and then

00:16:08 ◼   ► you gotta start all over again.

00:16:10 ◼   ► And I thought this was ridiculous and I hated it.

00:16:12 ◼   ► And then I discovered, I'm like, oh, I found the solution.

00:16:15 ◼   ► Turn off uh my Wi-Fi, my Ethernet, and tether to my phone because the

00:16:19 ◼   ► phone is on a different network, and that would work.

00:16:21 ◼   ► I'm like, that's it.

00:16:21 ◼   ► That must be there, must be like, you know, maybe there's something in my

00:16:24 ◼   ► house that's constantly trying to log into my Apple ID, or like just you

00:16:27 ◼   ► know, it opened always works on my phone.

00:16:29 ◼   ► But that lasted like three days.

00:16:31 ◼   ► And then I started getting the same thing on my phone, or it'd be

00:16:33 ◼   ► like, I'd tether through my phone and I would get the 503 response.

00:16:35 ◼   ► So my theory on this, based on just hearing other people complain about it,

00:16:39 ◼   ► is there's some unwritten rules about activity they see associated with your Apple ID,

00:16:44 ◼   ► and they just decide to lock your Apple ID, and they won't let you

00:16:48 ◼   ► unlock it because of a second set of unwritten rules that says, hey, too

00:16:51 ◼   ► many requests coming from this IP address about Apple IDs.

00:16:54 ◼   ► Is it a device in my house that's constantly trying to log into my

00:16:57 ◼   ► Apple ID?

00:16:57 ◼   ► Is it someone trying to log into every Apple ID in the world somewhere

00:17:00 ◼   ► on the same ISP as me?

00:17:02 ◼   ► Who knows?

00:17:03 ◼   ► Apple won't tell you, uh, but it's a crappy part of the security apparatus

00:17:07 ◼   ► that kind of necessarily has to surround Apple IDs, and I don't like it.

00:17:12 ◼   ► I'm sorry.

00:17:13 ◼   ► Let's talk Ramageddon because when will we stop talking about Ramageddon?

00:17:16 ◼   ► In like three years.

00:17:17 ◼   ► Uh Apple has secured lower display prices amid RAM shortage.

00:17:20 ◼   ► Uh Joe Rosignal at Mac Rumors writes Apple secured lower prices for OLED displays

00:17:24 ◼   ► for the upcoming iPhone 18 Pro models, according to industry sources cited by South

00:17:30 ◼   ► Korea's deal site.

00:17:31 ◼   ► For example, the report said that Apple will pay LG and Samsung around $68

00:17:35 ◼   ► per OLED display destined for the iPhone 18 Pro Max, down significantly from the

00:17:39 ◼   ► $110 to $120 dollars that it paid those suppliers for each of the iPhone

00:17:43 ◼   ► 17 Pro Max displays.

00:17:44 ◼   ► Apple's said to have made an aggressive push for lower display display prices due

00:17:44 ◼   ► to the ongoing memory chip shortage.

00:17:50 ◼   ► Well, apparently they've got still got enough weight to throw around for some vendors.

00:17:54 ◼   ► I guess the vendors that are not flush with cash from the huge RAM

00:17:58 ◼   ► prices, but uh uh I would do wonder how much that's going to offset

00:18:02 ◼   ► the uh what they have to pay for the other components.

00:18:04 ◼   ► We'll see.

00:18:04 ◼   ► I mean the rumors still say that the uh 18 Pro and Pro Max,

00:18:09 ◼   ► which I remember are the only that and the the foldable whatever, are the

00:18:13 ◼   ► ones that are coming out in September and the plain 18.

00:18:15 ◼   ► The rumors say uh it's not coming out in September.

00:18:17 ◼   ► So anyway, for the high-end phones that are coming out in September, I guess

00:18:21 ◼   ► I forget if the uh Air 2 is in that group.

00:18:24 ◼   ► Uh the rumor is that they're gonna be more expensive.

00:18:27 ◼   ► Like the phones are gonna be more expensive.

00:18:28 ◼   ► So maybe this is just getting uh a cheaper price on displays is just

00:18:32 ◼   ► enough to only have to raise the price $200 instead of $250.

00:18:37 ◼   ► It's gonna be rough.

00:18:38 ◼   ► Additionally, Apple raises traded values for most iPhone, iPad Mac, and Apple Watch models.

00:18:43 ◼   ► Marcus Mendez writes, Apple updated its trade in values table today, increasing offers for

00:18:48 ◼   ► most of its products, with some devices seeing jumps of nearly 30%.

00:18:51 ◼   ► The company also added several Samsung Galaxy Google Pixel and OnePlus models to the

00:18:54 ◼   ► list.

00:18:55 ◼   ► So John did some homework and some mathematics and came up with some examples.

00:18:59 ◼   ► The iPhone 16 Pro is now $630, up $11%.

00:19:03 ◼   ► Apple Watch Series 10, $165, up 9%.

00:19:06 ◼   ► The iPad Pro, I'm not sure what vintage, John, but I guess it doesn't

00:19:09 ◼   ► really matter.

00:19:09 ◼   ► $720 up 4%.

00:19:11 ◼   ► And MacBook Air, $580 up 10%.

00:19:15 ◼   ► Yeah, that we were talking about this before, like people were complaining.

00:19:18 ◼   ► Oh, they raised the prices on all their stuff, but have they raised trade

00:19:20 ◼   ► in values?

00:19:21 ◼   ► No, well, now they have.

00:19:22 ◼   ► But I look at the percentages, and remember when we read the percentages of

00:19:25 ◼   ► how much the prices of this stuff was increasing?

00:19:27 ◼   ► I know it was not Apples to Apples because the watches weren't included in

00:19:29 ◼   ► that price increase or whatever.

00:19:32 ◼   ► But uh as you might expect, the trade-ins are increasing by a lower percentage,

00:19:36 ◼   ► not just a lower amount, because obviously trade-ins, you get less money than a

00:19:39 ◼   ► new device, but a lower percentage than a lot of their sibling devices increase.

00:19:44 ◼   ► And and you know, we're reading like iPhone 16 Pro trading values, not necessarily

00:19:48 ◼   ► iPhone 17 Pro.

00:19:48 ◼   ► But anyway, you see the 9 to 5 Mac story for the full table.

00:19:51 ◼   ► But uh the news is basically it's good that Apple is finally increasing the

00:19:55 ◼   ► trade in value because that's what the market is dictating for these devices.

00:19:59 ◼   ► But it's bad for all of us that the trade in values are increasing

00:20:02 ◼   ► by less than the new device prices are increasing.

00:20:07 ◼   ► We are sponsored this week by Tailscale.

00:20:09 ◼   ► So if you're an ATP listener, it is likely that you have at least

00:20:12 ◼   ► a couple of devices that you may want to be able to talk to.

00:20:15 ◼   ► For perhaps your Mac and a Synology, maybe you have a device at a

00:20:19 ◼   ► friend's or a parent's house.

00:20:21 ◼   ► Maybe you have a Raspberry Pi at a parent's house to help with troubleshooting

00:20:24 ◼   ► or something like that.

00:20:25 ◼   ► Maybe you have 48 Mac minis in a co-location facility.

00:20:28 ◼   ► You never know what could happen.

00:20:29 ◼   ► Well, imagine you could talk to any of these other devices, no matter what

00:20:33 ◼   ► network you're on or what network they're on.

00:20:36 ◼   ► If you use a special IP address provided by Tailscale or hostname for that

00:20:40 ◼   ► matter, as long as both these devices are on your tail net, you will

00:20:44 ◼   ► be able to talk to them.

00:20:45 ◼   ► Tailnet is amazing.

00:20:46 ◼   ► It busts through firewalls, it busts through network address translation, it does amazing things.

00:20:52 ◼   ► Tailscale has compensated us to talk to you about their software, but they cannot

00:20:57 ◼   ► pay me to say this is possibly my favorite software I've used in years

00:21:00 ◼   ► and years and years and years.

00:21:02 ◼   ► It is so good.

00:21:02 ◼   ► So what Tailscale does is, like I said, it creates a mesh network between

00:21:05 ◼   ► all your devices.

00:21:06 ◼   ► This is different, mind you, than Hub and Spoke.

00:21:09 ◼   ► And despite the fact that it uses WireGuard, which is a VPN technology, you

00:21:12 ◼   ► shouldn't think of it as a VPN.

00:21:14 ◼   ► It's really its own thing.

00:21:14 ◼   ► And what makes it so magical is that no matter what network you're on,

00:21:18 ◼   ► no matter what continent you're on and what what network or continent your target

00:21:22 ◼   ► device is on, you will be able to talk between your devices on your

00:21:25 ◼   ► tailnet.

00:21:26 ◼   ► It is truly incredible.

00:21:28 ◼   ► And not having to care where you are when it comes to accessing your

00:21:32 ◼   ► other machines and devices is amazing.

00:21:35 ◼   ► It's so freeing.

00:21:36 ◼   ► So here's the thing.

00:21:37 ◼   ► Tailscale's personal plan will be free forever.

00:21:40 ◼   ► They've regularly committed to that.

00:21:42 ◼   ► And it's free forever for up to six users and unlimited devices.

00:21:46 ◼   ► They have been making it more and more permissive over time.

00:21:49 ◼   ► True story.

00:21:50 ◼   ► There's no credit card required.

00:21:51 ◼   ► They'll never even hint to asking for one.

00:21:54 ◼   ► They don't even care.

00:21:55 ◼   ► So you can go to Tailscale.com slash accidental tech to get started.

00:21:59 ◼   ► And additionally, if you'd like to bring Tailscale to work, you can use the

00:22:01 ◼   ► code AccidentalTech for three free months of any paid plan.

00:22:05 ◼   ► Thousands of companies like Microsoft NVIDIA and the New York Times have already made

00:22:08 ◼   ► the switch.

00:22:08 ◼   ► So go to Tailscale.com slash accidental tech, T A I L S C A

00:22:13 ◼   ► L E dot com slash A C C I D E N T A

00:22:16 ◼   ► L T E C H to get started.

00:22:17 ◼   ► Thank you to Tailscale for sponsoring the show.

00:22:23 ◼   ► Any real news going on this week, Marco?

00:22:26 ◼   ► Um not really this week, but last week I did something pretty big.

00:22:29 ◼   ► All right, Captain Pedantic.

00:22:31 ◼   ► Fair enough.

00:22:32 ◼   ► Congratulations on shipping.

00:22:33 ◼   ► This is very exciting.

00:22:34 ◼   ► Thank you very much.

00:22:34 ◼   ► Yeah, I shipped my new app, Unforgetful.

00:22:37 ◼   ► I'm finally saying the name aloud on this podcast for the first time.

00:22:42 ◼   ► This is the app I've talked obliquely about it for a while on this

00:22:45 ◼   ► show, because I started writing it back in February.

00:22:47 ◼   ► Um, this is my reminder app that basically takes over the job of reminder

00:22:52 ◼   ► notifications and does a kind of a different philosophy on them than Apple's Built

00:22:57 ◼   ► in Reminders app that better fits my browser.

00:23:00 ◼   ► The the core of the problems that I was trying to solve is related,

00:23:06 ◼   ► I think, to my undiagnosed but fairly obvious uh adult ADHD.

00:23:12 ◼   ► When a reminder is out of sight, it's out of mind.

00:23:15 ◼   ► If I don't get notified for something, it's not going to happen.

00:23:20 ◼   ► And for a while, for you know, I wrote a I wrote a blog

00:23:23 ◼   ► post announcing the app, and I it was it's fairly personal.

00:23:26 ◼   ► Um, and and I kind of go through like my uh a brief a

00:23:29 ◼   ► brief history of you know growing up being a very forgetful person and not

00:23:34 ◼   ► really having you know any kind of useful diagnosis because I think if you

00:23:38 ◼   ► forgot all your homework when you were, you know, not hyperactive in the 90s,

00:23:42 ◼   ► you were just lazy.

00:23:43 ◼   ► And there's lots of awful psychological baggage that result from that that you know

00:23:49 ◼   ► that has affected lots of us, myself included, uh, from that generation.

00:23:53 ◼   ► But I realized that tools can help me.

00:23:56 ◼   ► And throughout my adult life, what has helped me by far more than anything

00:24:01 ◼   ► else has been calendar alerts.

00:24:03 ◼   ► If there's a calendar event I have to do, I set alerts on it

00:24:07 ◼   ► and I get notified.

00:24:08 ◼   ► And I get notified a lot, and I do them.

00:24:10 ◼   ► And that's the only way I'm ever gonna I ever do anything is with

00:24:13 ◼   ► calendar alerts.

00:24:15 ◼   ► And I wanted something similarly effective for to-dos, like tasks, reminders, whatever the term

00:24:21 ◼   ► is that you want to use, all the same thing.

00:24:23 ◼   ► There was a huge advantage that I got with reminders once Siri got good

00:24:27 ◼   ► enough that I could anywhere I was dictate a new reminder.

00:24:31 ◼   ► And people in my life, oh, they'll see me all constantly.

00:24:35 ◼   ► People around me will see me like, take out my phone.

00:24:37 ◼   ► Remind me in 15 minutes, take out the trash.

00:24:38 ◼   ► Like, you know, stuff like that.

00:24:39 ◼   ► And and if I, and even for things that are really minor that most

00:24:43 ◼   ► people would think, why would you need a reminder to do that?

00:24:44 ◼   ► Why don't you just remember?

00:24:48 ◼   ► I don't remember.

00:24:49 ◼   ► I need the reminder.

00:24:51 ◼   ► And this has served me very well as as an adult for the most

00:24:55 ◼   ► part, but there's been a few exceptions.

00:24:58 ◼   ► Most notably, if I don't get reminded at a particular time, I never open

00:25:04 ◼   ► the reminders app.

00:25:05 ◼   ► And part of that is because I hate the reminders app.

00:25:08 ◼   ► It's it is not at all what I want it to be.

00:25:11 ◼   ► A lot of people love it.

00:25:12 ◼   ► So I'm not going to say it sucks or anything.

00:25:13 ◼   ► Like it seems like it works great for a lot of people.

00:25:15 ◼   ► Does not work great for me.

00:25:17 ◼   ► I don't get the organization of it at all.

00:25:19 ◼   ► I never know which of the screens I want to be on.

00:25:21 ◼   ► They're all full of different subsets that are not useful to me.

00:25:24 ◼   ► Lots of things are overdue.

00:25:26 ◼   ► Lots of things are from the past.

00:25:28 ◼   ► Lots of things are from the distant future, and it's hard for me to

00:25:31 ◼   ► figure out like what screen I need to be on to see whatever I

00:25:33 ◼   ► need to do and whatever's active right now.

00:25:35 ◼   ► The other problems with reminders are, again, my my main UI or my or

00:25:40 ◼   ► my main interaction with it is voice for input and notifications for output.

00:25:47 ◼   ► The voice part is okay.

00:25:49 ◼   ► Um, it's a little unreliable in former versions of iOS.

00:25:54 ◼   ► In iOS 27, the one that's currently in beta, it takes a significant step

00:26:00 ◼   ► backwards.

00:26:01 ◼   ► I think dictating reminders to Siri in iOS 27 is awful.

00:26:07 ◼   ► Now, this has changed a little bit throughout the betas, so maybe it won't

00:26:10 ◼   ► ship awful in the fall, but that's only a few weeks away.

00:26:14 ◼   ► Um, and it's still not great.

00:26:16 ◼   ► It is starting to ask me fewer asinine clarifying questions in the later betas,

00:26:20 ◼   ► but it's still really slow now.

00:26:23 ◼   ► It's way slower than it was in previous OSs, and it's doing all these

00:26:29 ◼   ► smart LLM things, and that's wonderful for lots of other reasons, but for creating

00:26:29 ◼   ► reminders, it's a pretty big step backwards.

00:26:35 ◼   ► I didn't know that was going to happen when I started making my reminder

00:26:38 ◼   ► app, which is now now, as you know, is named Unforgetful.

00:26:41 ◼   ► Um I didn't know that at the time.

00:26:43 ◼   ► I this I started it in February.

00:26:44 ◼   ► But over the course of the summer, I'm I've realized like, oh, I need

00:26:47 ◼   ► to ship this app now, because I think a lot of people in the

00:26:50 ◼   ► fall, as they get used to this new version, they're gonna realize, oh, dictating

00:26:54 ◼   ► reminders in Siri is is pretty rough now.

00:26:57 ◼   ► Um anyway, that's the side side diversion.

00:27:00 ◼   ► But how does how does Unforgetful help with that?

00:27:02 ◼   ► Aren't you still dictating reminders with Siri to put things in Dunforgetful?

00:27:05 ◼   ► So it helps in a few ways.

00:27:07 ◼   ► So let me that was that was a diversion.

00:27:09 ◼   ► I'll get back to that in a minute.

00:27:10 ◼   ► But um the short version is that's why I have a dictation feature in

00:27:13 ◼   ► Unforgettable.

00:27:14 ◼   ► That feature didn't exist before like last month.

00:27:17 ◼   ► But I realized, oh man, these this beta is so painful for for dictating

00:27:21 ◼   ► reminders with Siri.

00:27:23 ◼   ► I should also probably have a fast dictation thing.

00:27:25 ◼   ► So what I did was I I built a dictation engine.

00:27:28 ◼   ► Um, and as of a few days ago, the 102 update also enables some

00:27:35 ◼   ► natural language processing on that input that's very, very fast.

00:27:38 ◼   ► So you can say you you can you know hit the microphone and you

00:27:40 ◼   ► can say tomorrow 9 a.m.

00:27:42 ◼   ► buy milk, and it will actually schedule it for tomorrow at 9 a.m.

00:27:45 ◼   ► and it will parse that out of the title as you'd expect.

00:27:47 ◼   ► Also, with that, um I optimized the design of the widgets and lock screen

00:27:53 ◼   ► controls and everything else to provide lots of easy, fast ways into dictation mode.

00:27:58 ◼   ► I mean, you know, there's widgets everywhere, there's control center controls, you can long

00:28:01 ◼   ► press the app icon, all those things will give you options for jumping right

00:28:06 ◼   ► into the microphone uh recording mode.

00:28:09 ◼   ► I even made it so that when you fire one of those off, it

00:28:12 ◼   ► does not even animate the sheet presentation for the dictation mode.

00:28:17 ◼   ► It just pops it in.

00:28:14 ◼   ► So you save 0.4 seconds there or whatever it is.

00:28:21 ◼   ► Like it I want it to be so fast.

00:28:24 ◼   ► And what I have it mapped to is I put the microphone button on

00:28:28 ◼   ► my lock screen in the bottom right, because I made a lock screen widget

00:28:31 ◼   ► also.

00:28:32 ◼   ► So you can add it to wherever wherever the system allows widgets, you can

00:28:36 ◼   ► put it on the action button, like whatever, whatever you want to do.

00:28:39 ◼   ► There's Mac controls that I think work.

00:28:41 ◼   ► It's really weird.

00:28:42 ◼   ► Mac controls, it's really weird, like the new control center API thing for the

00:28:45 ◼   ► Mac.

00:28:45 ◼   ► It's very, very strange.

00:28:47 ◼   ► But uh, I think it works.

00:28:49 ◼   ► Um, but for me, my fast capture mode is the microphone button on my

00:28:54 ◼   ► lock screen on my phone.

00:28:56 ◼   ► Uh, and that jumps right into it and lets me record.

00:28:58 ◼   ► It's so much faster than Siri.

00:29:00 ◼   ► But anyway, um, going back to kind of the purpose of the app, for

00:29:04 ◼   ► me it's all about fast capture and ubiquitous fast captures, and I can still

00:29:08 ◼   ► use Siri if it's if I can't easily reach the phone or whatever.

00:29:11 ◼   ► And then also on the notification side, it is very important to me that

00:29:17 ◼   ► no task ever get into a state that it does not ever notify me

00:29:21 ◼   ► again.

00:29:22 ◼   ► Because that happens too easily with reminders.

00:29:25 ◼   ► For me, it happens in two big ways.

00:29:26 ◼   ► Number one, if I forget to set a reminder date, if I just say

00:29:32 ◼   ► to Siri somewhere, remind me to buy milk, it will never remind me because

00:29:37 ◼   ► I didn't say it's like Simon didn't say, I didn't say remind me tomorrow

00:29:41 ◼   ► to buy milk.

00:29:41 ◼   ► I didn't say remind me in 15 minutes to buy milk.

00:29:43 ◼   ► I just did I just said remind me.

00:29:45 ◼   ► So it'll be like, all right, well, I added the reminders.

00:29:47 ◼   ► And so that's on you.

00:29:48 ◼   ► Good luck.

00:29:49 ◼   ► And that kind of like I will never look at it.

00:29:51 ◼   ► I will never find it again.

00:29:52 ◼   ► The other the other way that you can lose things in reminders if you

00:29:55 ◼   ► are me is say you remind me in 15 minutes, and it notifies you

00:29:58 ◼   ► in 15 minutes.

00:30:00 ◼   ► And if you don't snooze that notification, if you do anything else, if you

00:30:04 ◼   ► ignore it, if you accidentally swipe it away and dismiss it in any other

00:30:08 ◼   ► way, if you clear notifications, whatever.

00:30:10 ◼   ► If any if that notification, if anything else happens to it besides you hitting

00:30:14 ◼   ► snooze, it will never notify you again.

00:30:18 ◼   ► So you have one chance.

00:30:20 ◼   ► And this is not a good recipe for success if your brain works like

00:30:24 ◼   ► my brain.

00:30:25 ◼   ► The original concept of unforgetful back in February was mainly to just solve those

00:30:30 ◼   ► two problems.

00:30:31 ◼   ► Just notifications that will constantly remind me over and over again until something is

00:30:37 ◼   ► actually marked complete.

00:30:38 ◼   ► Now there's an app called Do DUE that does this.

00:30:41 ◼   ► We've talked about it before.

00:30:42 ◼   ► Lots of people like it.

00:30:43 ◼   ► Totally fine app.

00:30:44 ◼   ► It's not really for me.

00:30:45 ◼   ► Like I don't really love the UI and some of the mechanics of it,

00:30:48 ◼   ► but it's that kind of concept for notification.

00:30:50 ◼   ► So I so like if any if you want, you know, to to avoid

00:30:53 ◼   ► getting a bunch of emails that this app already exists.

00:30:55 ◼   ► I'm I'm very aware.

00:30:56 ◼   ► Thank you.

00:30:57 ◼   ► And it's perfectly fine app.

00:30:58 ◼   ► My life has for years now been managed by well, since forever it's been

00:31:04 ◼   ► managed by my calendar.

00:31:05 ◼   ► And anything that isn't a calendar item is in the do DUE app.

00:31:08 ◼   ► Like this is how I run my life.

00:31:10 ◼   ► And and I think that, you know, if if unforgetful was out years ago,

00:31:14 ◼   ► then I probably would have put all of those things into that instead.

00:31:17 ◼   ► But I've got so many freaking things in do now to move them over

00:31:20 ◼   ► would be quite a burden.

00:31:22 ◼   ► But um, but no, I I e if if unforgetful is for you, that's

00:31:27 ◼   ► perfect.

00:31:27 ◼   ► If do is for you, that's fine too.

00:31:29 ◼   ► But they're both very analogous in certain ways and very different in other ways.

00:31:33 ◼   ► Yeah, and and what I've realized, so I I decided fairly early on I

00:31:39 ◼   ► wanted to actually release this app to the public.

00:31:41 ◼   ► And uh it it has been a very interesting journey, I think, for me,

00:31:44 ◼   ► in part because it's a very narrowly focused app.

00:31:50 ◼   ► This is an app that you either know immediately it's for you, or you

00:31:54 ◼   ► know immediately it's not for you.

00:31:56 ◼   ► And this has been an interesting um, you know, it's been interesting to like

00:32:01 ◼   ► looking through the feedback and and trying to figure out how I'm how how

00:32:04 ◼   ► I should deal with that feedback.

00:32:06 ◼   ► Because for me, the app is it's very powerful if I keep that philosophy

00:32:13 ◼   ► of it will always remind you, no matter what, of every task in it.

00:32:18 ◼   ► And it doesn't, you know, it can it can space it out by a

00:32:20 ◼   ► few hours or tomorrow or whatever, but like it will, it will never stop

00:32:23 ◼   ► reminding you of a task until you mark it done.

00:32:26 ◼   ► Critically, it reads and writes your data from reminders, from the system reminders database.

00:32:31 ◼   ► It isn't just an import, it isn't a one-way thing.

00:32:33 ◼   ► It reads it and it writes it back to that same database.

00:32:36 ◼   ► So there is no import or export.

00:32:38 ◼   ► You can use Unforgetful with reminders, or you can use it with other apps

00:32:41 ◼   ► that use the reminder database, or you can use it for a while and

00:32:44 ◼   ► never look at reminders for ever again, or you can at the end of

00:32:47 ◼   ► that decide, you know what, I want to go back to reminders, just delete

00:32:50 ◼   ► unforgetful, and it's and reminders is still right there, and all your information is

00:32:52 ◼   ► still right there.

00:32:53 ◼   ► So it's very flexible in that.

00:32:55 ◼   ► One of the downsides of that though is that when somebody tries out Unforgetful

00:32:59 ◼   ► for the first time, it shows everything in their reminders database.

00:33:03 ◼   ► And you can select different sub lists if you want to like only research

00:33:05 ◼   ► it to like, you know, one list or whatever, you don't want your grocery

00:33:08 ◼   ► list in there, that's fine.

00:33:09 ◼   ► Um, but one of the most common feedback uh points I get is people

00:33:14 ◼   ► saying, oh my god, it filled it up with all these tasks that I

00:33:17 ◼   ► had scheduled for the distant future, but with no due dates.

00:33:20 ◼   ► I don't want those in there.

00:33:21 ◼   ► Now, I I even ran a poll yesterday on Mastodon for about an hour.

00:33:27 ◼   ► I know I deleted it early, but I ran a poll asking basically like,

00:33:31 ◼   ► hey, do you what do you want Unforgetful to do with tasks that have

00:33:36 ◼   ► no due date?

00:33:38 ◼   ► Because to me, I always interpreted that to mean I'm never going to do

00:33:41 ◼   ► that, as I mentioned earlier, because like I'm never gonna be notified of this,

00:33:43 ◼   ► so therefore it's a failure, it'll never happen, it's invisible.

00:33:46 ◼   ► That's not something I want to allow.

00:33:48 ◼   ► But about half of respondents to the poll said they want basically like a

00:33:54 ◼   ► someday section in the app.

00:33:56 ◼   ► And so I mocked it up, put it in the app, and I ran

00:34:00 ◼   ► it, you know, a couple of test builds and and I played with it

00:34:03 ◼   ► myself, and I'm like, I don't think I should add this.

00:34:08 ◼   ► Like, because to me, a someday section means I'm never going to do it.

00:34:15 ◼   ► And a someday task just becomes a bucket to dump stuff in that I'm

00:34:19 ◼   ► never gonna do.

00:34:20 ◼   ► And it in my philosophy is if I'm never gonna do it, I should

00:34:23 ◼   ► just delete it.

00:34:24 ◼   ► And no shame attached to that.

00:34:26 ◼   ► Like, just delete it.

00:34:27 ◼   ► This is a list of things that I intend to do, not a list

00:34:30 ◼   ► of things that I'm probably never gonna do.

00:34:32 ◼   ► You mentioned what people were calling these things when they talked to you about

00:34:35 ◼   ► them.

00:34:36 ◼   ► Oh, I have when I put a reminder in the reminders, but I don't

00:34:39 ◼   ► have a due date because it's stuff for the far future.

00:34:41 ◼   ► When I heard you say that, I'm like, well, if it's for for the

00:34:44 ◼   ► far future, it should have a date on it for the far future.

00:34:48 ◼   ► Right.

00:34:48 ◼   ► Perhaps not coincidentally.

00:34:50 ◼   ► I have a reminders list called Far Future that has dates in it that

00:34:54 ◼   ► has reminders in it that are in the far future, and they all have

00:34:57 ◼   ► dates, but they're in the far future.

00:35:00 ◼   ► And I the reason they're in that sub list is because I don't want

00:35:03 ◼   ► to see them in my like regular reminders list because I know they're like

00:35:06 ◼   ► two, three, four years from now, but they all have dates on them.

00:35:09 ◼   ► So I mean, I'm not I'm not gonna convince anyone to change how they

00:35:11 ◼   ► use reminders.

00:35:12 ◼   ► Everyone uses the apps the way they want to use it, but uh it's

00:35:14 ◼   ► just uh I was surprised by the 50% response in that poll as well.

00:35:18 ◼   ► Because I feel like look, if you really want this to be in the

00:35:20 ◼   ► far future, set it for and oh, I don't know the exact date.

00:35:23 ◼   ► Fine, set it for uh, you know, on the the the closest you can

00:35:26 ◼   ► get uh within a year, two years, three years, five years, just make a

00:35:30 ◼   ► guess and throw it in there because anything is better than no date at

00:35:34 ◼   ► all.

00:35:34 ◼   ► And I I'd never even thought about the idea of reminders without dates.

00:35:38 ◼   ► If I ever get a reminder without a date, it's because of a failure

00:35:40 ◼   ► of dictation and I'd immediately fix it.

00:35:42 ◼   ► Right, exactly.

00:35:43 ◼   ► Like I'd never intend that.

00:35:45 ◼   ► And that's again, that's part of why I just like so unforgetful, it treats

00:35:49 ◼   ► undated reminders like they're due today.

00:35:52 ◼   ► On the next hour boundary, it's gonna remind you, and then you can decide

00:35:55 ◼   ► what to do with it.

00:35:55 ◼   ► You can you can say remind me tomorrow, you can change the date, you

00:35:58 ◼   ► can snooze it, you can change it, you can whatever you want to do.

00:36:01 ◼   ► Anyway, so when people were asking for this, I I thought, like, well, my

00:36:04 ◼   ► instinct was like with Overcast, with Overcast, my goal is a broad audience.

00:36:09 ◼   ► I want I want Overcast to be the podcast app for everybody.

00:36:13 ◼   ► Most people, I think their needs for podcasts should be solved with Overcast.

00:36:17 ◼   ► Obviously, that's a little bit different now with like video and stuff, but like

00:36:20 ◼   ► the idea generally with Overcast was like I want this to be broad market.

00:36:24 ◼   ► So my instinct when I got this feature request of like undated tasks, what

00:36:29 ◼   ► should happen?

00:36:29 ◼   ► Everyone wants a someday section.

00:36:31 ◼   ► My instinct was to build it, so that's why I did, and I tried

00:36:34 ◼   ► it.

00:36:34 ◼   ► And as I was like trying to write the wording for the setting and

00:36:38 ◼   ► the setting screen, I was like, I hate this.

00:36:43 ◼   ► It felt like to use a wonderful programmerism, it felt like I was giving

00:36:47 ◼   ► people a foot gun.

00:36:48 ◼   ► If you're using Unforgetful, if this app is what you need, then if I

00:36:53 ◼   ► allow you to have an entire category of tasks that you never get reminded

00:36:58 ◼   ► about, this is not the app for you.

00:36:58 ◼   ► Like if that's what you want, this is not the app for you.

00:37:03 ◼   ► And then look for our sister app, Forgetful.

00:37:06 ◼   ► Right.

00:37:07 ◼   ► Like, and and I've been very careful when designing Unforgetful.

00:37:11 ◼   ► Like that to make it impossible to fail at your task.

00:37:14 ◼   ► That's the whole goal.

00:37:15 ◼   ► If you have a reminder, Unforgetful will do whatever it can to make sure

00:37:20 ◼   ► that you don't forget about it.

00:37:22 ◼   ► That's the whole point of the app.

00:37:24 ◼   ► That's why it's called that.

00:37:26 ◼   ► And by the way, naming it, oh my god, I'll get to that.

00:37:28 ◼   ► But so with this app, I'm I'm being very conscious.

00:37:31 ◼   ► Like, one of the unfortunate um limitations of the reminder API event kit, uh,

00:37:36 ◼   ► there's a lot of fields that reminders supports that I don't have access to.

00:37:40 ◼   ► I can't even tell if reminder has them or not.

00:37:42 ◼   ► So one of them is subtask relationships.

00:37:45 ◼   ► Reminders, I think of a few years ago, reminders added the ability to make

00:37:49 ◼   ► subtasks under a task.

00:37:51 ◼   ► There's this is a common feature of apps like Things and Omnifocus, you know,

00:37:54 ◼   ► all like you know, the big the big task management apps.

00:37:56 ◼   ► And the idea is oh, you can make a project, you know, make like

00:37:59 ◼   ► a top-level task.

00:37:59 ◼   ► That's the project.

00:38:00 ◼   ► And then you have the sub the subcheck marks of like, all right, we'll

00:38:02 ◼   ► do this, then do this, and then then the then the project is complete.

00:38:05 ◼   ► That's great.

00:38:06 ◼   ► There's there's a place for that.

00:38:07 ◼   ► But I decided early on, like, first of all, since I couldn't see those

00:38:10 ◼   ► relationships, well, I can't support them anyway.

00:38:13 ◼   ► But also, that's fine.

00:38:16 ◼   ► This app is not the app to manage that type of project.

00:38:20 ◼   ► If you like this is the app to manage, call the doctor's office tomorrow.

00:38:24 ◼   ► Simpler things.

00:38:26 ◼   ► This is not a project management app.

00:38:28 ◼   ► This is not a whole system.

00:38:30 ◼   ► It's it's not even really ideal for like collaborating with people.

00:38:34 ◼   ► I might change that at some point, but like it's you know, reminders has

00:38:37 ◼   ► this concept of like shared lists, and I admit I haven't actually done that

00:38:39 ◼   ► much testing on that because that's not really what it's for.

00:38:43 ◼   ► And there's all these different features.

00:38:44 ◼   ► There's tags and organization features that I just not only do I not have

00:38:49 ◼   ► access to, thanks to the API, but I'm not going to build those features.

00:38:53 ◼   ► Because what is tagging and you know, folders and everything?

00:38:57 ◼   ► What is that used for?

00:38:58 ◼   ► It's used to hide things selectively.

00:39:01 ◼   ► I don't want to hide things selectively.

00:39:03 ◼   ► I never want to hide things selectively.

00:39:04 ◼   ► That's the whole point of the app.

00:39:05 ◼   ► And so I am taking this different philosophy with this app of like, again,

00:39:10 ◼   ► this is either for you or it's not for you.

00:39:12 ◼   ► It's very hard for me to remember that when I'm reading all the feedback

00:39:17 ◼   ► and you know, some, you know, some of it's positive, some of it's not.

00:39:20 ◼   ► But what makes it very encouraging is like on this point specifically about about

00:39:24 ◼   ► like the undated tasks, whether they should be like in a someday bucket or

00:39:27 ◼   ► remind you right now.

00:39:29 ◼   ► I've I heard from a few people already, I think three or four people

00:39:32 ◼   ► already, who have said initially they were mad that all their someday tasks all

00:39:37 ◼   ► landed up there immediately and they were getting reminded.

00:39:39 ◼   ► But now they've set due dates, like John was just saying, like, they just

00:39:42 ◼   ► set dates or got rid of them or did them.

00:39:45 ◼   ► And now it's fine.

00:39:46 ◼   ► And I'm like, you know what?

00:39:47 ◼   ► That actually tells me this is a success.

00:39:49 ◼   ► It is a hundred percent the right move for me never to add that

00:39:52 ◼   ► feature.

00:39:53 ◼   ► Because what I want is that outcome.

00:39:55 ◼   ► I want you to either be reminded or complete them or set a date

00:40:00 ◼   ► in the future.

00:40:01 ◼   ► And I don't care if that date's five years from now.

00:40:03 ◼   ► Fine, whatever.

00:40:04 ◼   ► Set it as far in the future as you need to set it as

00:40:06 ◼   ► you know, to match your intention.

00:40:08 ◼   ► But if your intention is do this someday, well, how about you get reminded

00:40:12 ◼   ► about it once a month till you do it?

00:40:14 ◼   ► Because otherwise it's not gonna happen.

00:40:16 ◼   ► I just looked at my far future list, by the way, to uh see

00:40:20 ◼   ► what was in there, and I realized I have to check off one of

00:40:22 ◼   ► them because we already talked about it on the show.

00:40:24 ◼   ► It was the um Do AirPods Max 2 exist uh scheduled for early September.

00:40:29 ◼   ► We already did that.

00:40:30 ◼   ► Was that for a bet that we had?

00:40:29 ◼   ► Was that one of our bet things?

00:40:29 ◼   ► Yeah, we had it.

00:40:32 ◼   ► I was like, I think we talked about it in uh when they came

00:40:35 ◼   ► out in like um I just looked it up like episode 683.

00:40:38 ◼   ► Uh it was it was a bet, and I won.

00:40:40 ◼   ► Yay.

00:40:41 ◼   ► Anyway, I will check that off.

00:40:42 ◼   ► Now there are fewer items on my far future list.

00:40:45 ◼   ► So let me tell you about naming this app.

00:40:49 ◼   ► This was I consider myself fairly good at naming things.

00:40:53 ◼   ► Do you?

00:40:55 ◼   ► Oh, oh wow.

00:40:57 ◼   ► I was gonna say this name is a very Marco name.

00:41:00 ◼   ► I would say that Overcast is the aberration in your experience naming things because

00:41:05 ◼   ► Overcast is a straight up the middle good name with no weirdness.

00:41:08 ◼   ► This is not so trying to name a productivity app, a to-do app.

00:41:15 ◼   ► As it turns out, this has been a pretty common need since the dawn

00:41:18 ◼   ► of time.

00:41:19 ◼   ► So there are a lot of task management and project management and to do

00:41:25 ◼   ► and collaboration type, like organization type productivity apps out there in the world.

00:41:32 ◼   ► And many of the many of them have existed, many of them still exist.

00:41:36 ◼   ► There's many live trademarks in this in this world.

00:41:40 ◼   ► So my naming process is, which I am pretty good at, John, but my

00:41:45 ◼   ► naming I disagree, but okay.

00:41:48 ◼   ► My naming process is, you know, I'll brainstorm some stuff, I'll look into like,

00:41:53 ◼   ► you know, root words and parts of speech, different styles.

00:41:57 ◼   ► And I will say I tried having LLMs generate me name ideas, and they

00:42:01 ◼   ► are just horrendous at this task.

00:42:03 ◼   ► They are awful at naming things.

00:42:05 ◼   ► There were some good ideas that LMs threw at me, but like they kind

00:42:08 ◼   ► of got me thinking in different directions.

00:42:09 ◼   ► But you should have done like you you know that um, what is that?

00:42:12 ◼   ► Uh McSweeney article about uh email addresses that are hard to read over the

00:42:15 ◼   ► phone.

00:42:15 ◼   ► Have you seen that one?

00:42:16 ◼   ► Yeah, yeah.

00:42:14 ◼   ► Right.

00:42:17 ◼   ► So your app should have been called capital D lowercase O, capital D E

00:42:21 ◼   ► E, Capital Y O U, Capital E E E.

00:42:25 ◼   ► So it would be do D-U-E.

00:42:27 ◼   ► Wow.

00:42:28 ◼   ► Because that's what everybody says when they talk about that damn app, because that

00:42:31 ◼   ► app also has a problematic name.

00:42:33 ◼   ► Yeah, right.

00:42:34 ◼   ► You can't just say do.

00:42:36 ◼   ► Anyway, yeah.

00:42:37 ◼   ► So often I would come up with what I thought was a pretty good

00:42:41 ◼   ► name, and then I'd go do searches.

00:42:44 ◼   ► I'd do a trademark search, I'd do an app store search, and they just

00:42:47 ◼   ► were all not only taken, but like really taken.

00:42:51 ◼   ► Like taken in ways that like this is still like an active live company

00:42:54 ◼   ► using this name.

00:42:55 ◼   ► Like there's no like I really can't use this name.

00:42:58 ◼   ► Everything was taken.

00:43:00 ◼   ► And then I would often, you know, I'd I'd go in a different direction.

00:43:03 ◼   ► Okay, well, if I can't do something descriptive, how what if I do something

00:43:07 ◼   ► that's not even related?

00:43:08 ◼   ► Like I I wrote it on the way to Kona, uh to you know,

00:43:11 ◼   ► Kona, Hawaii.

00:43:12 ◼   ► What if I use Kona as the name?

00:43:14 ◼   ► I love that name.

00:43:15 ◼   ► Nope, that's some productivity app.

00:43:17 ◼   ► Um were you doing the naming after the icon?

00:43:20 ◼   ► Because you had a bit could have called it like Lighthouse Reminders or something.

00:43:22 ◼   ► No, the icon came after the name.

00:43:24 ◼   ► I went through so many names.

00:43:28 ◼   ► And so often I would land on something that was like, oh, that's not

00:43:31 ◼   ► bad.

00:43:31 ◼   ► Let me look at oh, nope, can't do that.

00:43:33 ◼   ► Or I would try I would think of some kind of metaphor, like, you

00:43:37 ◼   ► know, oh what what what about memory?

00:43:40 ◼   ► Uh what about oh um elephants?

00:43:43 ◼   ► They never forget.

00:43:44 ◼   ► Oh, I just invented Evernote.

00:43:45 ◼   ► Oh, evernote.

00:43:46 ◼   ► Like everything's been taken.

00:43:47 ◼   ► Every metaphor, everything has been taken.

00:43:50 ◼   ► Safe Harbor?

00:43:51 ◼   ► Lots of lighthouse-related ones.

00:43:53 ◼   ► Yeah.

00:43:53 ◼   ► Safe harbor for ADHD people who can't remember things.

00:43:56 ◼   ► Yeah, not bad.

00:43:57 ◼   ► It took it was this was the hardest naming challenge I have ever had.

00:44:01 ◼   ► And I I landed on Unforgetful.

00:44:03 ◼   ► I don't think it's my best name ever, to be honest, but I think

00:43:59 ◼   ► it's a I think it's a decent name.

00:44:07 ◼   ► It has the advantage, I think, correct me if I'm wrong, I believe it's

00:44:10 ◼   ► not actually a word, which has the advantage of it's not going to be

00:44:13 ◼   ► taken.

00:44:13 ◼   ► You know what I mean?

00:44:14 ◼   ► Like it's it's unique in that way.

00:44:16 ◼   ► It is, you know, it is your own branding and not just like I

00:44:18 ◼   ► took a word from the dictionary and called my app that.

00:44:20 ◼   ► So there is a a minor issue that I hope will no longer I

00:44:24 ◼   ► hope I hope it won't be an issue for long because it'll just have

00:44:27 ◼   ► an established name after a while.

00:44:29 ◼   ► But there's an issue that the app store search will try to auto-correct you.

00:44:34 ◼   ► If you type in unforgetful, it'll autocorrect you to unforgettable.

00:44:39 ◼   ► Ah crap.

00:44:41 ◼   ► That is actually pretty bad.

00:44:44 ◼   ► But I I'm hope like if that is always the case.

00:44:48 ◼   ► I'm assuming that was taken by the way, was unforgettable taken.

00:44:51 ◼   ► Um there were some closer trademark problems with that.

00:44:54 ◼   ► Um and I kind of I liked Unforgetful better because a word, apparently.

00:45:00 ◼   ► Yeah, it's it's like it's a word, but it's not really a word that

00:45:02 ◼   ► anybody would use.

00:45:03 ◼   ► But it it is technically a word.

00:45:04 ◼   ► But like I was focusing in on the like ADHD angle pretty early.

00:45:10 ◼   ► And one of the key design principles of the app is it never makes

00:45:15 ◼   ► you feel stress or shame.

00:45:18 ◼   ► No feelings of overwhelm.

00:45:19 ◼   ► That's why I didn't call it overdue, in addition to being, you know, probably

00:45:23 ◼   ► disrespectful to the do app.

00:45:25 ◼   ► Um, I didn't want to call it overdue because that's stressful.

00:45:28 ◼   ► I didn't want any feelings of of stress, shame, overwhelm.

00:45:33 ◼   ► So I wanted it to be lighter and encouraging, you know, fresh starts.

00:45:37 ◼   ► Oh, things that are related to like sunrises, because every day is a new

00:45:40 ◼   ► day, you know.

00:45:40 ◼   ► You know, stuff I I looked at so many metaphors.

00:45:43 ◼   ► Um the waves uh, you know, coming in.

00:45:46 ◼   ► Uh I thought like one of the one of the probably the best name

00:45:50 ◼   ► that the LLMs came up with was Nudge, but it wasn't quite good enough

00:45:54 ◼   ► for me.

00:45:55 ◼   ► And I think there, I think there actually is an app called that anyway,

00:45:57 ◼   ► so I couldn't use it anyway.

00:45:58 ◼   ► But there was like like beacon or pin, like, okay, you know, not bad,

00:46:02 ◼   ► not I wanted to call it waves, can't do that.

00:46:03 ◼   ► It's taken.

00:46:04 ◼   ► I wanted like I just I wanted to call it reprise, even though it's

00:46:06 ◼   ► my PHP framework, but it's like I tried it trying to throw in like

00:46:09 ◼   ► fish references, musical references.

00:46:12 ◼   ► Every single word was taken, trademarked, like everything somewhere in the task management or

00:46:19 ◼   ► productivity space that was too close for comfort.

00:46:21 ◼   ► So it was just incredibly difficult to name this app.

00:46:25 ◼   ► Um, but I went with I went with unforgettable because you know, I I

00:46:28 ◼   ► had, you know, of my of my kind of finalists, it was the one

00:46:32 ◼   ► I liked the most, and it had like the least contention in the app

00:46:36 ◼   ► store.

00:46:36 ◼   ► It had the least contention for search results.

00:46:38 ◼   ► Like, if I can get Apple to stop autocorrecting it to a different word,

00:46:42 ◼   ► it'll be great because it it will be very easy to find in the

00:46:44 ◼   ► app store and it will be very high range.

00:46:46 ◼   ► Just put just put unforgettable in your keywords.

00:46:49 ◼   ► Oh, that's interesting.

00:46:50 ◼   ► Maybe I'll do that.

00:46:51 ◼   ► Yeah, that's a good idea.

00:46:53 ◼   ► Thank you.

00:46:53 ◼   ► Okay.

00:46:54 ◼   ► Um as for the actual journey of programming the app, um, you know, this

00:46:59 ◼   ► was originally this was mostly vibe coded.

00:47:02 ◼   ► Like what I wrote on the plane was mostly me learning about vibe coding.

00:47:06 ◼   ► I fairly quickly realized I didn't like AI code structure very much, and that

00:47:13 ◼   ► it had a lot of limitations.

00:47:15 ◼   ► And so I ended up most of the app's code I have written it

00:47:19 ◼   ► myself.

00:47:19 ◼   ► Um I still, you know, there is AI generated code still in there for

00:47:23 ◼   ► various boring framework y type parts of the code, like you know, some of

00:47:26 ◼   ► the Swift UI views, like that's there's not that much for me to do

00:47:30 ◼   ► that's distinctive there.

00:47:31 ◼   ► But like the data layer, I wrote almost all of it myself.

00:47:35 ◼   ► More importantly, everything dealing with dates and times and calendars and days, I wrote

00:47:42 ◼   ► almost all of that myself.

00:47:43 ◼   ► Yeah, because that's the easy stuff.

00:47:45 ◼   ► Oh my god.

00:47:46 ◼   ► Dealing with dates is so and times and everything.

00:47:50 ◼   ► It's comically tricky.

00:47:52 ◼   ► Yep.

00:47:52 ◼   ► There are so many pitfalls.

00:47:55 ◼   ► And the fortunate, like Apple provides very nice APIs.

00:47:59 ◼   ► The calendar API, the date components, like all these different parts of it, it

00:48:03 ◼   ► provides very good APIs.

00:48:04 ◼   ► Um what I found was overall kind of like the the best balance for

00:48:10 ◼   ► me is I wrote the time parsing and and time handling code, and then

00:48:15 ◼   ► I had AI check it for bugs and write a bunch of test cases

00:48:18 ◼   ► for me.

00:48:19 ◼   ► I mean, I I can't tell you how many tests, I think there's like

00:48:22 ◼   ► 500 tests in the application testing.

00:48:25 ◼   ► All sorts of different like different edge cases, and and part of why there's

00:48:29 ◼   ► so many.

00:48:30 ◼   ► So you might think, if you are writing the reminders API, um recurring task

00:48:37 ◼   ► support.

00:48:37 ◼   ► How do you do that?

00:48:39 ◼   ► And the answer is it's actually very, very simple.

00:48:42 ◼   ► Um all you have to do is mark a task as completed in event

00:48:48 ◼   ► kit, and reminders automatically generates the next one in its recurrence series.

00:48:53 ◼   ► And that's awesome.

00:48:54 ◼   ► Except, what if you want to tell the user when the next time will

00:48:59 ◼   ► be in the UI before it happens?

00:49:01 ◼   ► Like if you want to, if if they're entering the recurrence rule and you

00:49:04 ◼   ► want to say this will repeat on September 1st and October 1st, what if

00:49:08 ◼   ► you want to show them that?

00:49:09 ◼   ► That's actually pretty helpful.

00:49:10 ◼   ► There's no way to do that.

00:49:12 ◼   ► That's fine.

00:49:12 ◼   ► Huh, okay.

00:49:13 ◼   ► What if, what if you want to actually change some of these meanings?

00:49:17 ◼   ► Like if you want to, like one of the features of Unforgettable is if

00:49:21 ◼   ► you have suppose you have a task that recurs every day and you are

00:49:26 ◼   ► two weeks behind and you haven't done it yet.

00:49:28 ◼   ► If you go back and you check complete, you're gonna have to check complete

00:49:31 ◼   ► 13 more times to get yourself to the present day.

00:49:35 ◼   ► And I thought that is awful.

00:49:38 ◼   ► I've done it before.

00:49:39 ◼   ► I've been there, it's horrendous.

00:49:40 ◼   ► It just tells me how much of a failure I've been and it makes

00:49:42 ◼   ► me feel bad and it's tedious and error prone.

00:49:45 ◼   ► So everything about it is bad.

00:49:46 ◼   ► So I'm like, I want unforgettable.

00:49:48 ◼   ► When you complete an old overdue recurring task where you've missed intervals in the

00:49:53 ◼   ► intervening time, I want it to just skip all those and just get to

00:49:56 ◼   ► today.

00:49:57 ◼   ► Well, then that means I have to change the logic of what it means

00:50:00 ◼   ► to complete a recurring task.

00:50:02 ◼   ► So between wanting to display the future dates in the UI and that, I

00:50:06 ◼   ► pretty quickly realized I think I need to write my own recurring task parser

00:50:12 ◼   ► thing that like extrapolates the next dates forward for a recurring task.

00:50:16 ◼   ► So I did.

00:50:17 ◼   ► I wrote every single one of those cases.

00:50:19 ◼   ► And if you look in the reminders app and you go to custom repeat

00:50:22 ◼   ► thing, every single one of those I wrote the logic for that with the

00:50:27 ◼   ► calendar APIs and doing it all like you know manually and myself and everything.

00:50:31 ◼   ► And then I had AI go find find the bugs, which I actually patting

00:50:36 ◼   ► myself on the shoulder right now.

00:50:37 ◼   ► I actually did okay.

00:50:38 ◼   ► I didn't have that many bugs in the version I wrote, but I had

00:50:41 ◼   ► more than zero and it fixed them.

00:50:42 ◼   ► And then it made 500 test cases to make sure I got them all.

00:50:46 ◼   ► Um, that was a really fun challenge, but it was it was a lot.

00:50:50 ◼   ► Um, because yeah, dealing with dates and times, very, very tricky stuff.

00:50:54 ◼   ► And just when you're when you're making any kind of app like this, there's

00:50:57 ◼   ► so many questions of like, okay, what should happen in this edge case?

00:50:59 ◼   ► You might think, oh, remind me every day about these tasks or whatever.

00:51:07 ◼   ► Okay, well, you know, what if this time doesn't exist in this day?

00:51:10 ◼   ► Or what if, you know, what if somebody says remind me on the 31st

00:51:13 ◼   ► of the month and it's an it's a month that doesn't have 30 days

00:51:15 ◼   ► or doesn't have 31 days.

00:51:16 ◼   ► Um, or uh just all sorts of things, what you know, what if the

00:51:19 ◼   ► person flies across the country, changes time zones between this interval, what should happen?

00:51:23 ◼   ► You know, like there's so many just decisions about how things should behave in

00:51:29 ◼   ► an app that deals with timed events.

00:51:32 ◼   ► And so I I hope what I have done here is make a system

00:51:37 ◼   ► that makes sense, that is like internally consistent, coherent, predictable.

00:51:43 ◼   ► I think I have it's it makes sense to me.

00:51:46 ◼   ► I hope it makes sense to other people.

00:51:48 ◼   ► Um, but overall, like dealing with times has been by far the most challenging

00:51:55 ◼   ► part of this app.

00:51:56 ◼   ► I don't doubt it.

00:51:57 ◼   ► Uh time arithmetic is brutal.

00:52:01 ◼   ► And I actually just had a very small bug um with something in call

00:52:05 ◼   ► sheet that's honestly the story is not that interesting, but it was basic time

00:52:08 ◼   ► arithmetic gets you every time or gets me every time.

00:52:11 ◼   ► Um that is absolutely brutal.

00:52:13 ◼   ► So now that it's out, and and forgive me, I'm not trying to like

00:52:16 ◼   ► cut to the end or anything like that, but what is the what is

00:52:19 ◼   ► the response been?

00:52:20 ◼   ► Because it seems extremely positive from what I've seen.

00:52:22 ◼   ► And it's funny because when I first saw the app, and we uh John

00:52:26 ◼   ► and I only knew about it, I don't know, like a few weeks ago,

00:52:29 ◼   ► month ago maybe.

00:52:30 ◼   ► Um when I first saw the app, I thought this is great.

00:52:34 ◼   ► I don't know that this is something that I personally need, and I worried

00:52:36 ◼   ► that this may be something that not that many people need.

00:52:39 ◼   ► Now, since this is not, you know, your your primary income, I I wasn't

00:52:44 ◼   ► maybe worried was too dramatic a word, but I I figured, oh, there'll be

00:52:44 ◼   ► a handful of people that like this, and that'll be that.

00:52:49 ◼   ► But based on just what I've seen, it seems like there are a lot

00:52:53 ◼   ► of people who are looking for exactly this.

00:52:56 ◼   ► So from my perspective, it looks like the response was great.

00:52:59 ◼   ► So far, it's been I you know, I had kind of like like a

00:53:04 ◼   ► low bound and a high bound of what I expected, and it's basically right

00:53:09 ◼   ► in the middle.

00:53:10 ◼   ► Oh, good.

00:53:10 ◼   ► Okay.

00:53:11 ◼   ► I'm happy with it.

00:53:11 ◼   ► It's not blowing me away, it's not crucially disappointing.

00:53:14 ◼   ► It's right kind of right in the middle of what I what I kind

00:53:18 ◼   ► of expected slashed hoped for.

00:53:20 ◼   ► Um, so I'd say so far it's doing well.

00:53:22 ◼   ► I think the only major um question mark is so the app it has

00:53:27 ◼   ► a hard paywall with a $20 a year subscription and with a one-month free

00:53:32 ◼   ► trial.

00:53:33 ◼   ► And I wanted to do a whole month free trial so people could actually

00:53:36 ◼   ► like use it, like move into it, like live with it, try it and

00:53:40 ◼   ► see how it works for you before you pay me anything.

00:53:43 ◼   ► Because I want you to be able to try it.

00:53:45 ◼   ► And then of course, with reminders, integration, like you can try it and all

00:53:49 ◼   ► your stuff is right there.

00:53:50 ◼   ► And then if it's not for you, you can leave and all your stuff

00:53:52 ◼   ► is still right there in reminder.

00:53:54 ◼   ► So you can leave whenever you want to also.

00:53:55 ◼   ► So I figure like I'll make it really easy for people to try it

00:53:59 ◼   ► with no pressure, and then after a month, if they're still using it, then

00:54:03 ◼   ► I can get paid.

00:54:03 ◼   ► How long is the one month free trial?

00:54:06 ◼   ► What?

00:54:06 ◼   ► How long is the one month free trial?

00:54:08 ◼   ► What do you mean?

00:54:09 ◼   ► It's a month.

00:54:09 ◼   ► Date math.

00:54:10 ◼   ► Date math.

00:54:11 ◼   ► There it is.

00:54:11 ◼   ► That's up to Apple.

00:54:12 ◼   ► I guess that's whatever they do in App Store.

00:54:15 ◼   ► That's my question.

00:54:15 ◼   ► Oh, what does Apple do?

00:54:17 ◼   ► Like, so if you, you know, because if you did this yourself, and well,

00:54:20 ◼   ► people will complain no matter what.

00:54:22 ◼   ► They would say, I bought it in a month where there's not that many

00:54:26 ◼   ► days, and I bought it on the 31st, but then my subscription ended on

00:54:29 ◼   ► the 28th of the next month.

00:54:31 ◼   ► Not fair.

00:54:32 ◼   ► Honestly, I have no, I don't actually know what what how Apple handles that.

00:54:29 ◼   ► Neither do I.

00:54:37 ◼   ► That's what I'm saying.

00:54:37 ◼   ► Date math.

00:54:38 ◼   ► It's terrible.

00:54:40 ◼   ► Because in that situation, like I've I do this in my app is why

00:54:43 ◼   ► I thought of it.

00:54:44 ◼   ► I just always give them 31 days, period.

00:54:46 ◼   ► That's why that's how I do the date math.

00:54:47 ◼   ► And so you can get a one-month thing that lasts you two calendar months

00:54:51 ◼   ► because you span over February, you add up, you know what I mean?

00:54:54 ◼   ► 31 days, no complaints.

00:54:55 ◼   ► But that's just because I'm rolling it myself, which is has its own problems

00:54:58 ◼   ► because Apple doesn't support the kind of payments I do in my app, but

00:55:02 ◼   ► dates.

00:55:02 ◼   ► Terrible.

00:55:03 ◼   ► Yeah.

00:55:04 ◼   ► Um, I actually um I I think my usage of Cloud Kit is a

00:55:09 ◼   ► little bit interesting.

00:55:10 ◼   ► And I wanted to, I wanted to mention that.

00:55:12 ◼   ► So I don't directly need CloudKit, you know, at first glance because I'm reading

00:55:17 ◼   ► and writing from the reminders database.

00:55:19 ◼   ► And your reminder's database is probably stored in iCloud for you know on your

00:55:22 ◼   ► iCloud account, but that's not of my concern.

00:55:23 ◼   ► I just read it locally.

00:55:25 ◼   ► But I started realizing I needed CloudKit for a few additional things.

00:55:29 ◼   ► So one is like there's a few cases where I want to store metadata

00:55:35 ◼   ► about each reminder to provide some feature that I want to offer.

00:55:39 ◼   ► So for instance, the snooze times are dynamic.

00:55:43 ◼   ► Uh they there's like a few different brackets of like you know, short, medium,

00:55:46 ◼   ► and long snooze intervals.

00:55:48 ◼   ► The one it offers you depends on what you pick last.

00:55:50 ◼   ► So if you picked a very short interval as your last snooze, like the

00:55:53 ◼   ► 15-minute one, it'll give you the set of short ones.

00:55:55 ◼   ► If you picked, you know, a day as to to snooze it last time,

00:55:59 ◼   ► it'll give you the long ones.

00:56:00 ◼   ► And so I wanted to store for each reminder the last used snooze interval.

00:56:05 ◼   ► And I also, I had this idea for location-triggered reminders, that the biggest problem

00:56:12 ◼   ► I have with location-triggered reminders is that the radius of a geofense that triggers

00:56:18 ◼   ► them is a hundred meters.

00:56:14 ◼   ► It's a pretty big radius.

00:56:21 ◼   ► And this is you can try to make them smaller, but they'll be less

00:56:23 ◼   ► reliable.

00:56:24 ◼   ► So Apple recommends in like the programming guide for core location and everything, they

00:56:27 ◼   ► recommend 100 meters as the smallest geofence radius.

00:56:34 ◼   ► When you're just kind of like doing background pulling of geofencing and not doing

00:56:38 ◼   ► like active turn-by-turn directions, the GPS radio is in a very low power state.

00:56:42 ◼   ► It's not pulling constantly.

00:56:43 ◼   ► And so if you have like a smaller radius that you're trying to trigger

00:56:46 ◼   ► based on, you might actually miss it as you move through it.

00:56:49 ◼   ► So a hundred meters is the kind of the default geofense radius.

00:56:52 ◼   ► And whenever you tell reminders, hey, remind me when I get home, do this.

00:56:56 ◼   ► Um it'll remind you when you're within 100 meters of your house.

00:56:59 ◼   ► That can be when you're like half a block away.

00:57:02 ◼   ► Like you like you might, you know, you might be near your house, but

00:57:06 ◼   ► if you say remind me when I get home to, you know, put the

00:57:08 ◼   ► milk in the fridge, I mean, hopefully you don't need that reminder, maybe a

00:57:13 ◼   ► bad example.

00:57:14 ◼   ► But like, you know, if you say that, it's gonna remind you when you're

00:57:18 ◼   ► like three houses down in the car or maybe five houses down on foot.

00:57:23 ◼   ► And by the time you get into the house where the fridge is, you

00:57:27 ◼   ► might have forgotten.

00:57:28 ◼   ► I know I do.

00:57:30 ◼   ► So for me, geofense alerts for from reminders have always been not particularly useful

00:57:35 ◼   ► because it's like, well, I don't want to be reminded when I'm like driving

00:57:39 ◼   ► up near the house.

00:57:41 ◼   ► I want to be reminded when I'm in the house when I can actually

00:57:44 ◼   ► do something about this task.

00:57:46 ◼   ► Because I don't want to forget in the intervening time.

00:57:49 ◼   ► And so I I wanted to add delayed triggers for notifications.

00:57:53 ◼   ► I wanted to add, all right, make it so I can set myself to

00:57:55 ◼   ► remind me five minutes or ten minutes or thirty minutes after I get to

00:57:59 ◼   ► a location, remind me of this thing then.

00:58:02 ◼   ► That way I'm like, once I'm like in the house, jackets off, stuff's put

00:58:06 ◼   ► down, I'm settled, then I can actually do something.

00:58:09 ◼   ► So I had to I had to store a field on reminders.

00:58:13 ◼   ► Um and I also wanted to be able to do things, you know, just

00:58:16 ◼   ► like you know, for future features, like if I want to add URL support

00:58:18 ◼   ► or whatever, I can't read reminders URLs.

00:58:20 ◼   ► Thanks, event kit.

00:58:21 ◼   ► Please, I man, if there's one thing, I would love to be able to

00:58:24 ◼   ► read URLs.

00:58:25 ◼   ► Um, but the ability to add fields, you can't really do it in event

00:58:28 ◼   ► kit.

00:58:28 ◼   ► There is no way in event kit to add arbitrary metadata to reminders.

00:58:32 ◼   ► Now, other apps have come up with various hacks for this.

00:58:35 ◼   ► Sometimes they'll try to cram it into the notes field, they'll put like an

00:58:38 ◼   ► identifier in the notes field, or they'll just actually literally write the metadata in

00:58:42 ◼   ► right into the notes field.

00:58:43 ◼   ► And that's fine.

00:58:44 ◼   ► I I wanted to avoid it if I could, or if it if there

00:58:47 ◼   ► was like a better solution.

00:58:48 ◼   ► And then totally separately from that, I ran into an issue during testing one

00:58:53 ◼   ► day.

00:58:54 ◼   ► For the notifications to work best, ideally you turn off reminders notifications and you

00:59:00 ◼   ► leave unforgettable notifications on and it replaces reminders notifications.

00:59:05 ◼   ► But unforgettable on the phone is limited by iOS's background refresh intervals.

00:59:10 ◼   ► If you don't allow it to background refresh, or if you tell Sear in

00:59:14 ◼   ► your phone, remind me in three minutes to do something, and unforgetful doesn't have

00:59:20 ◼   ► a chance to background refresh itself in that three minute span, it's not going

00:59:24 ◼   ► to be able to notify you because it doesn't know about that reminder yet.

00:59:28 ◼   ► Now, I have a Mac version too.

00:59:29 ◼   ► It's a Mac and iOS app.

00:59:31 ◼   ► The Mac version can constantly monitor.

00:59:33 ◼   ► So it as soon as the Mac gets a new reminder through iCloud or

00:59:37 ◼   ► C or whatever, the Mac the Mac app instantly knows here it is, and

00:59:42 ◼   ► it create it can create a notification for it.

00:59:44 ◼   ► So I'm like, what if I use iCloud background refresh triggering?

00:59:49 ◼   ► When the Mac app sees a new reminder, if it's running on a Mac

00:59:53 ◼   ► somewhere and you're out somewhere with your phone, it can send a CloudKit silent

00:59:58 ◼   ► push to the phone.

01:00:00 ◼   ► And I think, oh my God, that's how to do it.

01:00:03 ◼   ► So I started using CloudKit for that.

01:00:05 ◼   ► That when the so that and this is why when the Mac app, the

01:00:09 ◼   ► Mac app gives you gives it one of those annoying confirmation messages when you

01:00:13 ◼   ► try to close it, like Chrome.

01:00:15 ◼   ► Because I actually don't recommend that you close the Mac app ever.

01:00:19 ◼   ► Because if the Mac app is running, it can do this.

01:00:22 ◼   ► It can pick up notification, it can pick up new reminders immediately, and then

01:00:25 ◼   ► send a Cloud Kit push to wake up the phone app so the phone

01:00:29 ◼   ► picks them up way faster than if it has to wait for a background

01:00:33 ◼   ► refresh.

01:00:33 ◼   ► Can you make the Mac app just run in the background?

01:00:35 ◼   ► Like make it a little background job, faceless background job.

01:00:38 ◼   ► I initially I had it as one of those like doc or can or

01:00:42 ◼   ► um menu bar or both toggles.

01:00:45 ◼   ► I think I'm gonna put that back.

01:00:46 ◼   ► But I think there's a way you can get it totally in the background,

01:00:48 ◼   ► and the reason I think so is because Golden Gate puts icons in the

01:00:51 ◼   ► menu bar for apps that are doing stuff in the background.

01:00:54 ◼   ► I don't think I want to deal with that, but I I definitely would

01:00:58 ◼   ► like to I'd probably in one of the nearby nearby future versions, I'm probably

01:01:01 ◼   ► gonna add an option to put it just in the menu bar because a

01:01:04 ◼   ► lot of people have requested that.

01:01:05 ◼   ► Um anyway, so that's why I recommend if you run run the Mac app

01:01:10 ◼   ► all the time if you can, because that will keep your phone app up

01:01:12 ◼   ► to date faster.

01:01:14 ◼   ► Um anyway, so I had to I I dealt with CloudKit for that, and

01:01:17 ◼   ► then I'm like, you know, this this was kind of like midway through the

01:01:20 ◼   ► summer, as I'm trying to figure out what to do with these different features

01:01:23 ◼   ► that I want to have metadata for, like the delayed notification or delayed location

01:01:27 ◼   ► triggers and things like that, and the dynamic snooze.

01:01:30 ◼   ► I was like, hmm.

01:01:31 ◼   ► Well, now that I'm using CloudKit and Fable was out, and I'm like, Fable's

01:01:36 ◼   ► really good at coding.

01:01:37 ◼   ► What if Fable writes me a bridge between my Blackbird database and CloudKit?

01:01:45 ◼   ► I've already I've been planning to do this forever.

01:01:47 ◼   ► I called it Skybridge.

01:01:49 ◼   ► And like I made a prototype back in the day.

01:01:51 ◼   ► I never finished it.

01:01:52 ◼   ► So I had Fable look at my prototype, and using the style that I

01:01:57 ◼   ► use and the Blackbird mechanics that I use, just finish it.

01:02:01 ◼   ► Like take this prototype and finish it and test it and make it good

01:02:05 ◼   ► and make it work and fix the bugs.

01:02:08 ◼   ► And it did.

01:02:09 ◼   ► So and this is you can actually see what it did.

01:02:11 ◼   ► There's um, if you look at my my Blackbird library's open source, and there's

01:02:14 ◼   ► a branch named Fable, Skybridge is in there, and you can use it if

01:02:18 ◼   ► you want.

01:02:19 ◼   ► It's just automatically syncs a Blackbird table to CloudKit.

01:02:23 ◼   ► And it's awesome.

01:02:26 ◼   ► I'm so I'm so happy about this.

01:02:29 ◼   ► Like I couldn't when it did that, I'm like, oh my god, this would

01:02:32 ◼   ► have taken me a week to do this myself, and my version would have

01:02:37 ◼   ► had more bugs.

01:02:38 ◼   ► And it would have taken me another month after that to find them all

01:02:41 ◼   ► and fix them all, and I never would have had tests.

01:02:44 ◼   ► I'm so happy to have this.

01:02:46 ◼   ► And since having Skybridge, and since I already did all the work on Blackbird,

01:02:50 ◼   ► it's a fully functional Blackbird app with no servers, just with CloudKit as the

01:02:54 ◼   ► as the sync back end.

01:02:55 ◼   ► So now I can add arbitrary metadata to any reminder whenever I want.

01:03:00 ◼   ► And I can if I want to add another column, it's no big deal.

01:03:03 ◼   ► I'm I just do it and I go to this, you know, the CloudKit

01:03:06 ◼   ► dashboard and publish the changes and all that stuff.

01:03:08 ◼   ► But like, it's fine.

01:03:09 ◼   ► What a game changer that has been.

01:03:12 ◼   ► So I'm I'm now able to do good features like that.

01:03:14 ◼   ► Anyway, so that's I I thought the Cloud Kit story was a little bit

01:03:17 ◼   ► interesting on that, especially the background push.

01:03:18 ◼   ► I think that I think that really helps keep the app up to date.

01:03:22 ◼   ► Um, but anyway, so to actually answer your question, Casey from a thousand years

01:03:27 ◼   ► ago, the res how the reception has been.

01:03:30 ◼   ► Again, kind of what I was saying earlier about about with Overcast uh versus

01:03:33 ◼   ► this, like I'm having to exercise a lot of discipline, not adding every feature

01:03:39 ◼   ► everybody asks for.

01:03:40 ◼   ► Because if somebody's asking me for things like you know, folders and organization, I

01:03:45 ◼   ► think it's I can confidently say if you're looking for this app to do

01:03:50 ◼   ► that, it's not really the app for you, or this is not the app

01:03:53 ◼   ► for that job.

01:03:55 ◼   ► Um, and if you use reminders that way.

01:03:58 ◼   ► Like I was listening to um our friend Steven Robles over on Mac Power

01:04:01 ◼   ► Users.

01:04:02 ◼   ► He recently made a Kanban app.

01:04:05 ◼   ► Uh he vibe coded the whole app.

01:04:06 ◼   ► It was great.

01:04:06 ◼   ► It's a good episode from Mac Power Users.

01:04:08 ◼   ► I think it was last week, and we'll link to it.

01:04:10 ◼   ► He was describing how he used reminders before he made his own app, and

01:04:14 ◼   ► he was a super power user of reminders, using things like subtasks and tags

01:04:18 ◼   ► and all these different and like I I've never used it that way.

01:04:21 ◼   ► I've never used anything that way.

01:04:22 ◼   ► I've I've tried using Omnifocus back in the day, I tried using things back

01:04:26 ◼   ► in the day.

01:04:27 ◼   ► I would stick with them for a little while, and then I would it

01:04:30 ◼   ► would fall on the floor and I'd forget about it.

01:04:32 ◼   ► And I I think any everybody with ADHD type tendencies, I think this is

01:04:35 ◼   ► a very familiar story.

01:04:37 ◼   ► You know, this is this is the the a category of apps that gets

01:04:41 ◼   ► a lot of like tire kicking.

01:04:43 ◼   ► Like, hey, uh yeah, I I can be the kind of person who organizes

01:04:46 ◼   ► my stuff this way, and you do it for like a little while, you

01:04:50 ◼   ► know, you do it for maybe you can stick with it for a few

01:04:52 ◼   ► days or a few weeks, or maybe if you're lucky, a couple of months,

01:04:57 ◼   ► but then it then you don't you stop checking it.

01:04:59 ◼   ► Like this is why I've never gotten into like paper notebooks.

01:05:03 ◼   ► Even like my my you know, my journey into remarkable tablets, I would just

01:05:07 ◼   ► eventually stop looking at them.

01:05:08 ◼   ► I would stop checking them.

01:05:09 ◼   ► And that's that's been true of everything I've ever done that can't notify me

01:05:13 ◼   ► constantly.

01:05:14 ◼   ► So I can pretty confidently state what I want this app to be and

01:05:20 ◼   ► what it is and what it isn't.

01:05:22 ◼   ► It is not an organizational system.

01:05:25 ◼   ► It is an automated nagging machine to get you to do stuff.

01:05:29 ◼   ► If that's what you need, it's very effective.

01:05:33 ◼   ► That is not what a lot of people need, though.

01:05:35 ◼   ► There, I mean, there are a lot of people out there who need it,

01:05:37 ◼   ► but like I think if you pull the average people out there in the

01:05:40 ◼   ► world, it's not gonna be the majority who need this.

01:05:44 ◼   ► If I take a hundred people off the street and I describe this app

01:05:47 ◼   ► to them, half of them are gonna look at me like I'm crazy.

01:05:51 ◼   ► I think about 40 of them are gonna be like, eh, I don't really

01:05:56 ◼   ► need it.

01:05:56 ◼   ► And then maybe 10 of them are gonna be like, oh my god, yes,

01:05:59 ◼   ► I need that.

01:06:00 ◼   ► That's kind of what I'm looking at.

01:06:02 ◼   ► And I don't really know how to reach those people.

01:06:05 ◼   ► I'm learning.

01:06:06 ◼   ► I'm gonna do a lot of experiments with different types of ads and different,

01:06:10 ◼   ► you know, trying to find different audiences, targeting different ads in different ways.

01:06:15 ◼   ► I really have no idea how to market this app.

01:06:19 ◼   ► And I don't think it's ever gonna be a mass appeal thing because it

01:06:24 ◼   ► doesn't have mass appeal.

01:06:26 ◼   ► To be a mass appeal task management app, you have to have a lot

01:06:30 ◼   ► of features and enable a lot of different ways of working.

01:06:34 ◼   ► Like that's you know, things and omnifocus.

01:06:36 ◼   ► Like these are these apps are great for that because they're very customizable.

01:06:40 ◼   ► They have very, very deep features and broad feature sets.

01:06:45 ◼   ► So what those kinds of apps, whatever your system is, they allow you to

01:06:51 ◼   ► mold the app to execute your system.

01:06:55 ◼   ► That's not this.

01:06:57 ◼   ► This is the app for people who systems don't work for.

01:07:03 ◼   ► And I'm very confident keeping it that way.

01:07:06 ◼   ► Because if for some reason you out there, if your brain works like mine,

01:07:10 ◼   ► if you can forget something five minutes after someone told you, this app might

01:07:16 ◼   ► be for you.

01:07:16 ◼   ► And there's enough of you out there.

01:07:18 ◼   ► There's enough of us.

01:07:19 ◼   ► I do I'm confident it can.

01:07:21 ◼   ► Yeah, I'm confident that this can be a reasonable business.

01:07:25 ◼   ► It's ne I don't think it's ever gonna be anywhere near the size of

01:07:27 ◼   ► Overcast in terms of either user base or income.

01:07:30 ◼   ► I I I would be shocked if it ever came near it.

01:07:33 ◼   ► Um I think I'd be happy with a tenth of it.

01:07:36 ◼   ► But I know that there's other people like me out there.

01:07:40 ◼   ► And if you are out there and if you want a better reminders app

01:07:44 ◼   ► for yourself, try unforgetful.

01:07:46 ◼   ► See see how it works for you.

01:07:47 ◼   ► Give it the month-free trial.

01:07:49 ◼   ► If you don't like it, don't worry, delete it, go back to reminders, no

01:07:52 ◼   ► big deal.

01:07:52 ◼   ► Um, but but give it a shot.

01:07:54 ◼   ► Um, because I think there's a I think there's a lot more of us

01:07:56 ◼   ► out there than than we realize.

01:07:58 ◼   ► But I don't know how to I don't know how to reach us.

01:08:00 ◼   ► So help me out.

01:08:02 ◼   ► Did you get unforgetful in the app store?

01:08:04 ◼   ► Like no colon suffix.

01:08:06 ◼   ► Um no, I actually I didn't even try.

01:08:09 ◼   ► You always gotta try.

01:08:10 ◼   ► No, I you don't because the way the app store search works, uh this

01:08:14 ◼   ► is you know, look, I had on I had overcast no colon forever.

01:08:18 ◼   ► Um, but I stopped using it because app store search is brutal.

01:08:24 ◼   ► You have to jam your title and your subtitle and your keywords full of

01:08:29 ◼   ► every keyword you can think of.

01:08:29 ◼   ► And where you put them matters, what you know, which of those fields they're

01:08:29 ◼   ► in, they all it matters.

01:08:36 ◼   ► If there's a phrase and both words appear in the title or subtitle, that

01:08:40 ◼   ► matters more than if they're split across the two fields.

01:08:42 ◼   ► Like there's all these different things with app store search optimization.

01:08:45 ◼   ► I wish the app store search wasn't so dumb, but it is, and this

01:08:50 ◼   ► is this is the environment I operate in.

01:08:53 ◼   ► So if you want your app store app to succeed, you have to do

01:08:58 ◼   ► things like shove a bunch of keywords in the title in the subtitle.

01:09:01 ◼   ► You just have to.

01:09:02 ◼   ► So I didn't even the idea of not doing that, it it never was

01:09:07 ◼   ► a possibility.

01:09:07 ◼   ► So what is your subtitle?

01:09:09 ◼   ► Never lose a task.

01:09:10 ◼   ► Oh, that's what's a good one.

01:09:12 ◼   ► That's after the colon.

01:09:13 ◼   ► When I was searching for um uh apps called Nag, of which there are

01:09:17 ◼   ► many, um, one of them has a good subtitle, which is Nag, and they

01:09:21 ◼   ► use a hyphen instead of colon, which I think is a mistake.

01:09:23 ◼   ► But anyway, Nag, Relentless Reminders.

01:09:25 ◼   ► That sounds more like do, actually.

01:09:27 ◼   ► That's actually not that's that's a decent title.

01:09:29 ◼   ► I like that.

01:09:29 ◼   ► Because yours, you didn't mention this, which I think is uh everyone should read

01:09:32 ◼   ► Margot's post about this because it's it's really good and explains even more stuff,

01:09:35 ◼   ► if you can imagine.

01:09:36 ◼   ► Um and one of the the the key things that that I didn't realize

01:09:38 ◼   ► until I read your post about it, but I'd been test flighting the app,

01:09:41 ◼   ► but I just thought it was like doing weird stuff because the reminders thinks

01:09:44 ◼   ► syncs stuff.

01:09:45 ◼   ► But you know, this this app's tagline being relentless reminders.

01:09:48 ◼   ► Uh unforgetful is intentionally not relentless.

01:09:51 ◼   ► Like if you have a bunch of notifications that would normally show up like

01:09:55 ◼   ► all at the same time, you space them out so as not to throw

01:09:58 ◼   ► a pile of crap on top of people, which I it would never have

01:10:01 ◼   ► occurred to me that you were doing that intentionally.

01:10:03 ◼   ► And it's it's why the um it's why never lose a task is a

01:10:07 ◼   ► much better subtitle for your app than relentless reminders because you're trying not to

01:10:10 ◼   ► be relentless.

01:10:11 ◼   ► I mean, you're relentless in a gentle way, I guess.

01:10:13 ◼   ► Well, yeah, because again, like so that that again that goes back to some

01:10:16 ◼   ► key design things.

01:10:17 ◼   ► No shame, try to minimize any any way to feel overwhelmed.

01:10:21 ◼   ► And so what you know, obviously, look, I get I'm not great at keeping

01:10:25 ◼   ► up with my to-dos.

01:10:26 ◼   ► Obviously, what happens frequently is sometimes I'll get into I'll get to a state

01:10:31 ◼   ► where like for a week, I will every day be deferring to tomorrow like

01:10:36 ◼   ► seven to ten different tasks.

01:10:39 ◼   ► You know, and my default tomorrow day, my default tomorrow time, 9 a.m., what

01:10:43 ◼   ► would happen then is at 9 a.m., I would get seven to ten notifications.

01:10:48 ◼   ► Um now, on what happens on modern Apple platforms is they stack and they

01:10:53 ◼   ► get summarized.

01:10:54 ◼   ► And so if that happens every day, and if every single day you get

01:10:59 ◼   ► the same seven notifications at 9 a.m.

01:11:03 ◼   ► that you've been deferring for a week, you're gonna only see the top one.

01:11:09 ◼   ► And maybe you're gonna see the weird summary of it, you know, instead, but

01:11:13 ◼   ► that's it.

01:11:14 ◼   ► You're not gonna see the ones that are buried in the middle, which means

01:11:18 ◼   ► they're never gonna get done.

01:11:19 ◼   ► And this has multiple issues.

01:11:20 ◼   ► Number one, you feel hopeless.

01:11:22 ◼   ► Oh my god, it's another system I've lost track of, it's too big, I

01:11:24 ◼   ► can't, I I've lost control, I I'm a failure, and I must just bury

01:11:29 ◼   ► it and never look at it again.

01:11:30 ◼   ► So that's obviously a failure mode that I'm trying to avoid.

01:11:34 ◼   ► So that's where the spacing came out.

01:11:36 ◼   ► If you have seven things that are gonna all remind you at 9 a.m.,

01:11:40 ◼   ► I'm gonna space them out by your interval that you set.

01:11:42 ◼   ► You can customize it.

01:11:42 ◼   ► You can and you can set it to zero if you don't want them

01:11:44 ◼   ► spaced out.

01:11:45 ◼   ► But you can say, like, you know, just remind me, space them out every

01:11:47 ◼   ► 15 minutes.

01:11:48 ◼   ► That way, every 15 minutes, as you're working, you get one notification.

01:11:53 ◼   ► And you can look at it and you can say, Oh, you know what?

01:11:55 ◼   ► Fine, I will finally call the doctor.

01:11:57 ◼   ► But also to avoid the stacking problem.

01:11:59 ◼   ► Uh and this is the again.

01:12:02 ◼   ► This is like if you think this sounds insane, you don't your brain doesn't

01:12:07 ◼   ► work the way ours do.

01:12:10 ◼   ► It also matters what order you're reminded in.

01:12:13 ◼   ► So I built a feature that randomizes the order.

01:12:18 ◼   ► And it uses a seated, predictable random number generator, and it's seated based on

01:12:23 ◼   ► the date.

01:12:24 ◼   ► So that way your Mac and your phone randomize them in the same way

01:12:28 ◼   ► every day.

01:12:29 ◼   ► So no matter where, no matter what device you're on, you're gonna get a

01:12:33 ◼   ► randomized order of those tasks that stack up every day.

01:12:35 ◼   ► The ones that are scheduled for the same time.

01:12:37 ◼   ► They change every day in a random order.

01:12:40 ◼   ► It's the same in all your devices.

01:12:41 ◼   ► So that way, when you get that stack notifications, nothing ever gets lost in

01:12:47 ◼   ► the middle forever.

01:12:48 ◼   ► Because every day it's in a different order.

01:12:50 ◼   ► It's gonna be on top eventually.

01:12:53 ◼   ► So you will see it.

01:12:54 ◼   ► And like, this is the kind of problem that a functioning human being who

01:12:58 ◼   ► whose brain does not work the way mine does, like, why are you letting

01:13:01 ◼   ► them stack that I've never had things stack up that I'm just deferring for

01:13:05 ◼   ► a week straight?

01:13:06 ◼   ► Why don't you just do them instead of deferring them for a week?

01:13:08 ◼   ► Yeah, but that's not how my brain works.

01:13:10 ◼   ► So the way I the way I work, such as it is, is a

01:13:14 ◼   ► lot of snoozing, a lot of deferrals, a lot of things eventually happening in

01:13:17 ◼   ► batches or being given up on.

01:13:19 ◼   ► That's the type of feature that I build in here.

01:13:21 ◼   ► That's what this app is for.

01:13:22 ◼   ► And if you're again for the 90% of you out there who are like,

01:13:26 ◼   ► that's insane.

01:13:26 ◼   ► This is probably not for you.

01:13:28 ◼   ► But for the 10% of you out there who are like, oh, I would

01:13:30 ◼   ► actually really use that.

01:13:32 ◼   ► Try it.

01:13:33 ◼   ► I assure you, this this might work for you.

01:13:37 ◼   ► Oh, this is very exciting.

01:13:38 ◼   ► Uh, congratulations again.

01:13:40 ◼   ► I am pleased to see another entry in your I don't know, Arsenal isn't

01:13:45 ◼   ► the word I'm looking for, but in your repertoire, whatever you want to call

01:13:47 ◼   ► it.

01:13:48 ◼   ► Um, this is very, very exciting news.

01:13:44 ◼   ► And you should be very proud and very happy.

01:13:51 ◼   ► And I hope you're both of those things.

01:13:52 ◼   ► Thank you.

01:13:53 ◼   ► Yeah, I'm glad you finally come over to my side and started making a

01:13:55 ◼   ► weird little apps that most people don't like.

01:13:58 ◼   ► Oh, well, if you want a weird little app that most people don't like,

01:14:01 ◼   ► let me tell you the story about Fast Text.

01:14:04 ◼   ► Yeah, honestly, like I find it very um very interesting.

01:14:08 ◼   ► Like I as I've as I've talked about recently, um, you know, my my

01:14:12 ◼   ► job in Overcast has evolved to really be more of a caretaker of Overcast.

01:14:18 ◼   ► I don't I can't really change that much about it without angering a lot

01:14:22 ◼   ► of people.

01:14:23 ◼   ► What people want from me with Overcast is not tons of change.

01:14:27 ◼   ► They want just keep the app the same-ish, keep it working, keep it good,

01:14:32 ◼   ► fix any bugs, and do my pet feature.

01:14:35 ◼   ► Like that's what everyone, but nobody wants me to really change Overcast.

01:14:38 ◼   ► And so Unforgettable was a chance for me to do something very different.

01:14:43 ◼   ► This is like it's from scratch, it had no existing user base, so I

01:14:47 ◼   ► could do whatever I wanted, especially with a very small focus user base like

01:14:51 ◼   ► it's going to have.

01:14:52 ◼   ► So I it was a very different creative itch.

01:14:55 ◼   ► It was a very different problem space.

01:14:56 ◼   ► It's my first productivity app I've ever released.

01:14:59 ◼   ► I've never done anything productive.

01:15:01 ◼   ► In fact, I make apps that distract people from their productivity usually.

01:15:04 ◼   ► Oh, how quickly we forget about Quitter.

01:15:06 ◼   ► Uh, I guess that's kind of productive.

01:15:08 ◼   ► Um mostly giving you a hard time.

01:15:10 ◼   ► And the the chapter thing, what's it called?

01:15:12 ◼   ► Uh, forecast.

01:15:13 ◼   ► Yeah, but the forecast is a very like limited release.

01:15:15 ◼   ► The productivity tool.

01:15:16 ◼   ► If you had to pick the category in the app store, you'd pick productivity.

01:15:19 ◼   ► Maybe media.

01:15:20 ◼   ► Anyway, um, so yeah, this is like my basically my first productivity tool.

01:15:23 ◼   ► My first series one, my first one that I'm actually charging money for, my

01:15:25 ◼   ► first one that's really a business.

01:15:27 ◼   ► It's also my first um Mac App Store app.

01:15:30 ◼   ► And I, in general, I don't make a lot of Mac apps at all.

01:15:29 ◼   ► So it's one of my first Mac apps.

01:15:35 ◼   ► So it there's a lot of, you know, kind of firsts for me in

01:15:39 ◼   ► this.

01:15:40 ◼   ► And as an app developer, that's always exciting.

01:15:42 ◼   ► That's always very interesting.

01:15:43 ◼   ► Yeah.

01:15:44 ◼   ► Um, so I I've really enjoyed it quite a lot, and I continue to

01:15:48 ◼   ► enjoy it.

01:15:48 ◼   ► And frankly, if no one else ever uses it, I'll be disappointed, but I'm

01:15:53 ◼   ► still gonna use it.

01:15:55 ◼   ► Like, that's like forecast.

01:15:56 ◼   ► I use forecast every week.

01:15:57 ◼   ► That's how I edit the show.

01:15:59 ◼   ► Even though Forecast is not a major app and I haven't actually issued an

01:16:02 ◼   ► update in many years, I still use it constantly.

01:16:06 ◼   ► No matter how many people use this app, I'm going to use it myself

01:16:10 ◼   ► all the time.

01:16:11 ◼   ► And therefore, I am just I'm very happy with it.

01:16:14 ◼   ► And I I hope other people find it.

01:16:16 ◼   ► I hope it's good for them too.

01:16:18 ◼   ► By all means, please try it.

01:16:20 ◼   ► But if not, hey, you know, it's it's at least going to be good

01:16:23 ◼   ► for me.

01:16:25 ◼   ► Alright, so there's been a lot of uh news, hand wringing, I don't know

01:16:29 ◼   ► what you want to call it.

01:16:29 ◼   ► There's been a lot of stuff going on with regard to AI watermarking.

01:16:34 ◼   ► Now, uh we're gonna be talking a lot, perhaps primarily about Anthropic's Claude, and

01:16:40 ◼   ► it's worth noting that they are certainly a former sponsor, I believe a future

01:16:44 ◼   ► sponsor as well.

01:16:45 ◼   ► Um, so we I just want to make that plain up front.

01:16:48 ◼   ► But with that said, let's talk about what they're up to.

01:16:50 ◼   ► So Anthropic wrote that Anthropic has signed the EU AI Acts Article 502 Code

01:16:56 ◼   ► of Practice on Transparency of AI generated content.

01:16:59 ◼   ► Here's what this means for Claude.

01:17:01 ◼   ► New models will mark AI generated content from day one.

01:17:04 ◼   ► Claude models launched in the EU on or after August 2nd will support machine

01:17:08 ◼   ► readable marking at launch.

01:17:09 ◼   ► Generated text will carry embedded watermarks and generated files, will include digitally signed provenance

01:17:13 ◼   ► metadata where supported.

01:17:15 ◼   ► We'll help you detect Claude's marks.

01:17:17 ◼   ► We'll support users and other third parties to detect Claude's marks as the clo

01:17:22 ◼   ► as the code requires.

01:17:23 ◼   ► And we'll share details in forthcoming documentation.

01:17:26 ◼   ► Existing models are in progress.

01:17:28 ◼   ► The law includes a transition period for anthropic models launched before August 2nd.

01:17:32 ◼   ► And we're working to add marking support for those models as well.

01:17:36 ◼   ► There's a very long press release that goes on and on, but uh we're

01:17:39 ◼   ► not gonna read all of it.

01:17:40 ◼   ► But the initial question would be why are they doing this?

01:17:43 ◼   ► And the easy answer is well, they just said why they're doing it.

01:17:45 ◼   ► It's one of those EU regulations.

01:17:47 ◼   ► Apparently, you has this, you know, this AI act that says you have to

01:17:51 ◼   ► uh, you know, uh identity put some marking on your AI output so it

01:17:55 ◼   ► can be identified as AI.

01:17:57 ◼   ► Um, and there's lots of systems that do that.

01:17:59 ◼   ► There's uh there's a couple of standards that do that that uh a big

01:18:02 ◼   ► one that Adobe supports that I can never remember the acronym for.

01:18:05 ◼   ► Um, but those usually in the context of like, oh, we want AI generated

01:18:08 ◼   ► images, we want some way so that we can detect that they're AI generated.

01:18:12 ◼   ► And there's like you know, a metadata sidecar that you can stick on the

01:18:15 ◼   ► image that has some cryptographic info that says this was AI generated, blah blah

01:18:18 ◼   ► blah.

01:18:19 ◼   ► And obviously, all the systems, well, you know, they can just strip off the

01:18:22 ◼   ► metadata sidecar, just like you strip off location data from a photo.

01:18:24 ◼   ► And it's like, yeah, but it's like this is like an advisory system for

01:18:27 ◼   ► people who want to participate in the process and maybe it fulfills some regulations

01:18:31 ◼   ► and blah blah blah.

01:18:32 ◼   ► Then the next level is like, oh, and we have my our AI image

01:18:35 ◼   ► generator can use like steganography, I think that's what it's called, where you put

01:18:38 ◼   ► like secret hidden data in the pixels that humans can't see, but that AI

01:18:42 ◼   ► can see, and that's harder to strip out, but of course you can still

01:18:45 ◼   ► strip it out because you know you can change the pixels.

01:18:47 ◼   ► Uh, but at least that's clever, it's not like a metadata sidecar thing.

01:18:51 ◼   ► I mean, it's not really sidecar, it's the metadata is in the file, but

01:18:54 ◼   ► anyway, you know what I mean.

01:18:55 ◼   ► Like it's they would change the pixels as a way to do it, but

01:18:57 ◼   ► then you get to things like AI generated text, and you're like, how the

01:19:00 ◼   ► hell are you gonna watermark text?

01:18:59 ◼   ► Are you going to because text, unlike an image?

01:19:03 ◼   ► Usually text files don't have a place where you can shove metadata that's not

01:19:07 ◼   ► visible.

01:19:07 ◼   ► The whole point of a plain text file, it's a sequence of you know

01:19:10 ◼   ► Unicode code points or whatever, you know, uh encoding scheme it is, and that's

01:19:14 ◼   ► it.

01:19:14 ◼   ► There's no other like images have all sorts of stuff in them that doesn't

01:19:18 ◼   ► get displayed when you display the image, but text files don't.

01:19:20 ◼   ► That's what makes them text files.

01:19:22 ◼   ► So, how are you going to watermark uh text files?

01:19:25 ◼   ► I I forgot to look this up, but like watermark them is derived from

01:19:29 ◼   ► I'm assuming when they would like take a piece of paper and put a

01:19:32 ◼   ► mark on it.

01:19:32 ◼   ► I don't know if they actually ever use water, but like if you held

01:19:35 ◼   ► up to the light, you could see you you see it on like uh

01:19:38 ◼   ► paper currency in our country sometimes.

01:19:39 ◼   ► Anyway, that's the deration of it.

01:19:41 ◼   ► But like again, with a plain text file, what are you gonna do with

01:19:45 ◼   ► a plain text so I can identify it as uh AI generated, except unless

01:19:49 ◼   ► maybe at the very end it says buy clawed, right?

01:19:52 ◼   ► Um, so that's that's the problem they're faced with.

01:19:55 ◼   ► Uh but like, but why are they doing it?

01:19:57 ◼   ► Like, why you know again with the image stuff, it's like, oh, we want

01:19:59 ◼   ► to do this to be nice and so people know when things are AI

01:20:01 ◼   ► generated and blah blah blah, and yeah, it's not a security thing, we're not

01:20:04 ◼   ► fooling anybody because obviously anyone can just remove the watermark, but it's like we're

01:20:07 ◼   ► trying to do this to be nice.

01:20:09 ◼   ► There is one other angle that I hear everybody talking about with this, and

01:20:12 ◼   ► I have no idea if this is in Anthropic's mind at all.

01:20:15 ◼   ► I don't know if it, you know, it's just a happy accident because the

01:20:18 ◼   ► EU regulation, I think doesn't say anything about this, but the angle is um

01:20:24 ◼   ► first the premise that when AI models are training on stuff, that they don't

01:20:29 ◼   ► do well if they are trained on like their own output, like uh or

01:20:34 ◼   ► the output of other models, right?

01:20:36 ◼   ► So they want to train on stuff that was not AI generated.

01:20:39 ◼   ► If you train an AI on AI generated output, it just kind of goes

01:20:42 ◼   ► into a degenerative loop and uh things get worse and worse.

01:20:46 ◼   ► So I've heard.

01:20:46 ◼   ► I don't know the foundation of that, but that's what I've heard.

01:20:49 ◼   ► And so the idea here is why would Claude, why would Anthropic do this,

01:20:53 ◼   ► besides obviously complying with the EU regulation?

01:20:56 ◼   ► Like Claude is doing it, uh Anthropics doing it worldwide, which is just a

01:20:59 ◼   ► simplification, I think, which is what Apple should do in response to EU regulations.

01:21:02 ◼   ► But anyway, they're saying, Oh, they're doing it, so they won't accidentally train their

01:21:06 ◼   ► next model on stuff that was generated by any of their current models because

01:21:10 ◼   ► they're when their models go out there, they'll be like, oh, yum yum training

01:21:13 ◼   ► data, munch munch, munch, steal everybody's data.

01:21:15 ◼   ► Um, wait, our watermark, our watermark detector has detected that this data was generated

01:21:21 ◼   ► by us, so we won't consume that because we think it will make our

01:21:24 ◼   ► model worse.

01:21:26 ◼   ► That sounds like a conspiracy theory, but I I mean, I think it's plausible

01:21:31 ◼   ► again, assuming the whole idea that if models train on AI generated output, they

01:21:35 ◼   ► get worse, which also sounds plausible to me.

01:21:37 ◼   ► So keep that in the back of your mind.

01:21:38 ◼   ► I mean, we're not gonna again, we're not gonna read the whole rest of

01:21:40 ◼   ► this press release, but we will eventually read something else from anthropic about this

01:21:43 ◼   ► topic about you know why is anthropic doing this besides the fact that they

01:21:49 ◼   ► have to comply with the regulations, and how does this apply to other companies?

01:21:53 ◼   ► Anthropic's not the only one doing this, obviously.

01:21:55 ◼   ► Every this EU law applies to everybody, and Google has something similar and so

01:21:58 ◼   ► on and so forth.

01:21:59 ◼   ► So um, keep that in the back of your mind as we continue here.

01:22:02 ◼   ► But you know, the getting back to my earlier question is like, okay, so

01:22:04 ◼   ► I can see how you can watermark an image, whether with metadata or with

01:22:09 ◼   ► like clever changing of pixels, but how the hell do you watermark text?

01:22:12 ◼   ► And that's the question everyone had in their mind.

01:22:14 ◼   ► This story has mostly taken place already between two episodes of the show.

01:22:17 ◼   ► So we luckily have the benefit of time passing.

01:22:20 ◼   ► But in the beginning, everyone was like, the the press release didn't say however

01:22:24 ◼   ► they were gonna do it.

01:22:24 ◼   ► It's like, how how are they gonna do this?

01:22:26 ◼   ► How are you going to watermark a plain text file?

01:22:29 ◼   ► All right, I'm gonna do a lot of reading, so boys feel free to

01:22:32 ◼   ► interrupt me.

01:22:32 ◼   ► All right, how Claude's text watermark works.

01:22:29 ◼   ► This is from Anthropic.

01:22:36 ◼   ► In this article, we share answers to some of the questions we've received about

01:22:38 ◼   ► how our chosen watermark marking method works.

01:22:41 ◼   ► We use a method of watermarking that does not have any practical impact on

01:22:45 ◼   ► the quality or content of Claude's outputs.

01:22:47 ◼   ► The difference between watermarked and unwatermarked text will not be distinguishable to readers.

01:22:52 ◼   ► Nothing is added to the text, and there are no hidden characters.

01:22:55 ◼   ► Watermarking doesn't require extra tokens and will not be more expensive.

01:22:58 ◼   ► Watermarking carries no identifying information and can't be traced to a specific person, organization,

01:23:03 ◼   ► or chat.

01:23:04 ◼   ► Watermarking won't be specific to Claude.

01:23:06 ◼   ► As of August 2nd, the EU requires AI providers serving its market to mark

01:23:09 ◼   ► AI-generated content.

01:23:11 ◼   ► Other major model developers have signed the same code of practice and will be

01:23:14 ◼   ► implementing their own watermarks.

01:23:16 ◼   ► The watermark only applies to words Claude chooses.

01:23:18 ◼   ► When Claude proofreads text written by a person, what it gives back has generally

01:23:22 ◼   ► only been lightly edited.

01:23:23 ◼   ► Depending on the length of the text and how heavily Claude is edited it,

01:23:26 ◼   ► those changes might not be enough to make Claude's involvement detectable.

01:23:29 ◼   ► The more Claude writes, the more decisions it has to make, and the more

01:23:32 ◼   ► space there is for a watermark.

01:23:35 ◼   ► A watermark can only determine that Claude was likely involved with the content at

01:23:38 ◼   ► some point.

01:23:39 ◼   ► It cannot distinguish Claude wrote this from Claude heavily edited this.

01:23:43 ◼   ► So uh one thing that annoys me about anthropic is why did they wait

01:23:45 ◼   ► whatever number of days to release this?

01:23:47 ◼   ► Because everyone was flipping out about how it was gonna work.

01:23:49 ◼   ► This is these bullet points are very succinct and explain a bunch of stuff,

01:23:52 ◼   ► and then they go on to say, you know, basically if if Claude didn't

01:23:56 ◼   ► put the word there, it doesn't, it's not part of the watermark, which makes

01:23:59 ◼   ► sense, but it's worth saying, right?

01:24:01 ◼   ► And as I said, the more word Claude puts there, the more it's watermarked.

01:24:05 ◼   ► But like obviously, if you wrote those words and Claude didn't, they don't contribute

01:24:08 ◼   ► to the watermark.

01:24:09 ◼   ► So you need more words.

01:24:09 ◼   ► And they don't come out and say this, but I would imagine the watermarked

01:24:13 ◼   ► attack.

01:24:13 ◼   ► I think they do say it later that we don't have in this passage

01:24:15 ◼   ► here.

01:24:16 ◼   ► But the watermark detection is not thumbs up, thumbs down, yes, AI, no AI.

01:24:14 ◼   ► It's a confidence score.

01:24:22 ◼   ► It's like, uh, probably, maybe.

01:24:24 ◼   ► Do you get a percentage?

01:24:26 ◼   ► Because you can't, it's not, you know, it's not a slam dunk case.

01:24:30 ◼   ► Again, if Claude only changed three words, can you detect that at all?

01:24:33 ◼   ► What is the threshold?

01:24:34 ◼   ► Like they they talk about this in uh uh, I think on another podcast

01:24:37 ◼   ► talked about it being like 200 words being the uh the thing that you're

01:24:40 ◼   ► supposed to be able to detect according to the EU regulation.

01:24:43 ◼   ► But uh it's a little wishy-washy, but yeah, that's this is their explanation.

01:24:46 ◼   ► And all the stuff they crossed off no weird characters, no hidden text, it's

01:24:50 ◼   ► not gonna cost you more tokens, it doesn't identify you.

01:24:53 ◼   ► That's all very reassuring and makes sense.

01:24:55 ◼   ► But like, but how does it actually work?

01:24:58 ◼   ► How it works.

01:24:59 ◼   ► Large language models like Claude work by generating one word at a time.

01:25:02 ◼   ► Each time the model decides on the next word, it chooses amongst a list

01:25:05 ◼   ► of potential candidates, ultimately selecting the most sensible or likely based on the preceding

01:25:10 ◼   ► text.

01:25:10 ◼   ► Take the sentence, the weather today was cold, and the next word is very

01:25:14 ◼   ► unlikely to be sugary, but it is very quite likely even to be overcast

01:25:19 ◼   ► ding uh or gray.

01:25:21 ◼   ► Under most circumstances, it doesn't matter much to the reader which of the these

01:25:25 ◼   ► latter two words the model ultimately chooses.

01:25:27 ◼   ► The meaning of the sentence is largely the same either way.

01:25:30 ◼   ► In cases like this, the choice is settled by a random number.

01:25:33 ◼   ► Now let me pause you here because this is my favorite the most fun

01:25:36 ◼   ► thing.

01:25:37 ◼   ► You know, I love when people have to be confronted with how these things

01:25:39 ◼   ► work, because it is very difficult to hold in your mind how they work

01:25:43 ◼   ► because you know it's like pet rocks or you know, pet pets.

01:25:46 ◼   ► Like they're the mental model people have of these things is so different than

01:25:50 ◼   ► how they actually work that even if you explain it to them and they

01:25:52 ◼   ► understand it, they will discard that immediately and go back to believing they work,

01:25:55 ◼   ► however, you know, that's like a little uh magic elf in a machine or

01:25:58 ◼   ► something, right?

01:25:59 ◼   ► And so by there's anthropic, and they have to explain in like very simplified

01:26:04 ◼   ► terms um how their watermarking works, which requires them to discuss how their product

01:26:08 ◼   ► works.

01:26:09 ◼   ► They haven't gotten to the watermark, and they're just saying, here's what an LLM

01:26:11 ◼   ► does.

01:26:12 ◼   ► And what they're saying is when our you know it generates one word at

01:26:15 ◼   ► a time based on the probability, based on the words that came before it,

01:26:18 ◼   ► right?

01:26:19 ◼   ► And when we get to the point where uh you know, we have to

01:26:22 ◼   ► pick from some words that may be like roughly equally likely, we pick randomly.

01:26:27 ◼   ► And we're gonna be reading that and say, Well, you pick randomly, like this

01:26:33 ◼   ► is not the watermark, this is just how LMs work.

01:26:35 ◼   ► So I can imagine if people read this that they would be very upset

01:26:37 ◼   ► by wait a second, it's the same either way, so we just pick randomly.

01:26:41 ◼   ► That's your product.

01:26:42 ◼   ► This is upsetting.

01:26:43 ◼   ► But anyway, continue.

01:26:45 ◼   ► Watermarking uses low-stakes choices like these, which occur many times over a piece of

01:26:49 ◼   ► generated text to leave a pattern in Claude's responses.

01:26:53 ◼   ► That pattern is undetectable to the reader, but it's detectable to anyone who has

01:26:56 ◼   ► a key that encodes it.

01:26:57 ◼   ► When watermarking is used, choices are still made at random, but the source of

01:27:01 ◼   ► the randomness is different.

01:27:03 ◼   ► Instead of using an arbitrary random number generator to pick the next word, watermarking

01:27:06 ◼   ► uses the key and a few words that come before to settle what the

01:27:10 ◼   ► word what word the model should pick.

01:27:12 ◼   ► That is the words that Claude picks are still random, but now one can

01:27:16 ◼   ► check the sequence of words and see if it's consistent with the choices Claude

01:27:19 ◼   ► would make if it was using the key.

01:27:21 ◼   ► If it is, one can assign a probability that the text was generated by

01:27:24 ◼   ► Claude.

01:27:25 ◼   ► Amazing synergy by Marco, who just talked about the fact that he is doing

01:27:29 ◼   ► random number generation, but he is uh, you know, he's using pseudom random pseudo-random

01:27:33 ◼   ► number generator.

01:27:33 ◼   ► We will link to the Wikipedia page for pseudo-random number generation.

01:27:36 ◼   ► But the upshot is it's a function that you can call that gives you

01:27:39 ◼   ► a random number.

01:27:40 ◼   ► Uh, but like where do random numbers come from?

01:27:42 ◼   ► This is a very complicated topic.

01:27:43 ◼   ► Um, for if you're not doing it for something for crypto cryptographic or whatever,

01:27:46 ◼   ► you use a simple pseudo-random number generator where you put it give it a

01:27:50 ◼   ► seed value, and for any given seed value, the random number sequence from the

01:27:55 ◼   ► random number generator every time you call it will be exactly the same.

01:27:57 ◼   ► So if you seed it with one, two, three, you're gonna get the same,

01:28:00 ◼   ► you know, call it random number, random number, they're always going to be exactly

01:28:03 ◼   ► the same.

01:28:04 ◼   ► It's always gonna be the same, you know.

01:28:05 ◼   ► Every time you call it, it's a sequence of numbers.

01:28:07 ◼   ► And you're like, that doesn't sound random at all.

01:28:09 ◼   ► You say if I seed it with one, two, three and I call rand,

01:28:11 ◼   ► rand, rand, rand, it's gonna be the same answers from each one of those.

01:28:14 ◼   ► Not that each one is the same, but the first one will be 10,

01:28:16 ◼   ► the next one will be 20, the next one will be 15.

01:28:18 ◼   ► You know, like it'll always be 10, 20, 15, 10, 20.

01:28:21 ◼   ► If you seed it with one, two, three a year from now and call

01:28:24 ◼   ► it'll be 10 20, 15, 10, 20.

01:28:26 ◼   ► It's always the same sequence.

01:28:27 ◼   ► That's how pseudo-random number generators work.

01:28:30 ◼   ► And there is random choice within all LLMs, and we'll get to why in

01:28:34 ◼   ► a little bit by linking to all those videos where he's linked to.

01:28:36 ◼   ► And what they're saying is we always choose at random, but now we are

01:28:41 ◼   ► we have a random number generator that we know how it will we we

01:28:46 ◼   ► know how it will spit out the numbers.

01:28:47 ◼   ► And you know, like Marco doesn't know what the random numbers will be, but

01:28:51 ◼   ► he knows if he seeds both of them the same way, he will get

01:28:54 ◼   ► the same sequence of numbers.

01:28:56 ◼   ► We don't know what again, you don't know what those numbers will be ahead

01:28:58 ◼   ► of time, but you know, like that they're gonna be the same sequence.

01:29:01 ◼   ► So their random number generator is obviously more sophisticated than this, but they essentially

01:29:05 ◼   ► know because we decide how the random numbers are generated, we know whether you

01:29:10 ◼   ► know the our generator, our watermarking generator was used to generate these numbers versus

01:29:15 ◼   ► you know some other random number.

01:29:16 ◼   ► And and no one is more random than the other, right?

01:29:19 ◼   ► Because any any in anything's where you're a student random number generator.

01:29:23 ◼   ► What's a good seed value?

01:29:24 ◼   ► Well, a good seed value is one that's not going to repeat and one

01:29:28 ◼   ► is essentially unique, so a lot of the times the current time in like

01:29:31 ◼   ► microseconds since you know 1970 or whatever is part of the formula for the

01:29:29 ◼   ► seed.

01:29:36 ◼   ► But what you never want to do is use the same seed twice because

01:29:39 ◼   ► then you'll get a sequence of numbers that you've seen repeat before.

01:29:42 ◼   ► But each individual, the whole point of pseudo-random number generators is what you hope

01:29:45 ◼   ► is for every time you ask for one, you get one that is unpredictably

01:29:49 ◼   ► different than the previous one.

01:29:51 ◼   ► Obviously not unpredictable because it's a deterministic algorithm.

01:29:54 ◼   ► But like from an outside observer, if you don't know what the seed was,

01:29:57 ◼   ► you won't know what the next number is going to be.

01:29:59 ◼   ► So again, it should be exactly equally random to no matter what the seed

01:30:06 ◼   ► is, no matter what random number generator you have, as long as they both

01:30:08 ◼   ► have the same amount of entropy or whatever.

01:30:10 ◼   ► This is a very complicated mathematical topic that I don't know the details of.

01:30:13 ◼   ► But the whole point is one set of randomness is not necessarily any more

01:30:19 ◼   ► or less random than the other.

01:30:20 ◼   ► It's a thing that you can choose when you make the algorithms.

01:30:23 ◼   ► And Anthropic is saying it's random, just like it was before, but the random

01:30:28 ◼   ► number generator is one that we control in a way that we can predict,

01:30:32 ◼   ► but it's still random.

01:30:33 ◼   ► It's just as random as anything.

01:30:34 ◼   ► It's just as random as as Marco's random number, uh random order of his

01:30:38 ◼   ► notifications.

01:30:39 ◼   ► It's random enough that a user considers it random, but it behind the scenes

01:30:42 ◼   ► it's not random at all.

01:30:43 ◼   ► It's entirely determined by the seed.

01:30:44 ◼   ► And fun fact, that's why you can um copy someone's Minecraft world by just

01:30:51 ◼   ► copying the seed.

01:30:52 ◼   ► That's the literally like the Minecraft seed for the Minecraft world, that's what that

01:30:56 ◼   ► is.

01:30:56 ◼   ► It is a random number that that is serves as the seed to the

01:31:01 ◼   ► random number generator that generates the entire world.

01:31:04 ◼   ► And so that's why two different computers can you can just say just type

01:31:08 ◼   ► in this number, and you will generate exactly the same world that my computer

01:31:12 ◼   ► with the same version of the game generated, because that's literally what it's doing.

01:31:15 ◼   ► Yeah, because it's a like there is a sequence of code in the game

01:31:14 ◼   ► that's going to run.

01:31:19 ◼   ► And at various points, that code is going to say, give me the next

01:31:22 ◼   ► random number.

01:31:23 ◼   ► And as long as the seed is the same as someone else's seed, and

01:31:26 ◼   ► as long as the code is the same as theirs, it's going to get

01:31:28 ◼   ► the same answers as you got.

01:31:29 ◼   ► Like that's the seed of a pseudo-random number generator determines every number that will

01:31:33 ◼   ► come before it exact uh come after it exactly.

01:31:36 ◼   ► The whole trick is, you know, choosing a, you know, a seed that it's

01:31:39 ◼   ► not going to repeat and that people don't know.

01:31:41 ◼   ► And that's another upsetting thing about about computers.

01:31:43 ◼   ► People just assume computers can easily pick a random number.

01:31:45 ◼   ► It goes really deep.

01:31:46 ◼   ► Like there's a reason computers have like quote unquote real random number generators uh