Daniel (00:09) Dave, it's nice to see you again. Good morning. David Gary Wood (00:12) Good evening. Daniel (00:13) Yeah. It's the other way around today. Like it's my morning and your evening because we gonna got shuffled up in the schedule. David Gary Wood (00:23) We did, we got creative with the scheduling. the schedule. I'm good. Yeah, you're not faring too bad at this end of the day, so that's okay. Yeah, how was your? Daniel (00:27) schedule anyway how you doing Fantastic. I'm still a bit tired. I was I have a new hobby. I have a new hobby like I am now I'm not yet, but I'm almost member of a choir David Gary Wood (00:43) Mm. Daniel (00:49) ⁓ So I don't have the official membership yet because that needs a few more attendances and then they will actually test my singing ability. But so far it's been super fun and I'm hoping I'll get accepted as a tenor or a bass. And I think my chances are pretty good because they have like one other tenor and two other basses. David Gary Wood (01:10) You Daniel (01:10) And like 40, 40 Sopranos and Alts. David Gary Wood (01:13) ⁓ Well that's cool that's really cool so you've got a new activity. I haven't heard your bass singing Daniel actually. Daniel (01:14) So, yeah. They don't have enough men. Yeah. So yeah, that is super fun. I can try to get very low. But I think I'm more comfortable in the tenor range, which is slightly above. David Gary Wood (01:30) Hahaha Daniel (01:33) I have this buddy who is like a really good bass singer and he gets like crazy low. He gets like ludicrously low. I'm going to have to like, he has a very small band called Roadshot. I'm going to send you like after this, I'm going to send you some examples. think they're on the platforms. Yeah, cool. I have, or we had a fun idea last time, which is we should introduce each other. David Gary Wood (01:47) Yeah, yeah, yeah, send me. That'll be cool. Daniel (01:59) We should introduce each other by writing an introduction from the me perspective and then letting the other people, the person who is introduced reading it themselves. So I wrote a thing that starts with, hello, I am Dave. And you haven't read it yet. You haven't seen it yet. And we're to paste it into our shared notes and then you will have to read it to introduce yourself. Are you ready to do that? David Gary Wood (01:59) Mm-hmm. Mm-hmm. I have not. I am indeed. ⁓ what? Okay, I'm just gonna I'm gonna read it and then we will see where this goes. Okay Hello, I am Dave. I have a huge fetish for blue beards and goatees, which is fortunate because have you seen my co-host? yummy when it comes to software development, I have two wolves inside of me Daniel (02:29) ⁓ Don't think, just read. David Gary Wood (02:47) The professional wolf who wants my coworkers and employees to have as good a time as possible at work and who will use as efficient processes as possible to make sure the work gets done cleanly, quickly and with care. The other wolf inside of me is the tinkerer wolf who comes out when I want to rewrite my side projects into the 17th different software design pattern. Easy for me to say instead of just releasing the damn thing. ⁓ thank you, Daniel. Recently, all my tinkering energy is going into writing software using AI, which I'm trying to do as inefficiently as possible, mostly so that my co-host Daniel can pronounce Claude code wrong again. At home, I am a fantastic dad and partner, try to be mindful, try to be a mindful citizen of a colonial nation and just generally be a decent person. Daniel (03:37) That's pretty much you. What did I miss? David Gary Wood (03:39) Dang. I was worried. I was worried it was going to be really roasty. Mine sounds mean now. It's also a lot shorter. But no, you did very well. yeah, no, no, you did. You did. I can't take the credit. You wrote the words. Daniel (03:49) You did. David Gary Wood (03:53) I definitely wouldn't have introduced myself in that way. But now maybe I might. Inside me there are two wolves. Yes. Daniel (03:59) Sure. Put it in your LinkedIn. David Gary Wood (04:02) And yes, I know I need to release a damn app soon. We can talk about that in a bit. Daniel (04:02) Tinkerer wolf. David Gary Wood (04:06) Yeah, mine is comical and well, no, this might take just a little longer because I've got two versions of it. Thanks to and if it's wrong, don't blame me. You'll see why in a second. Are you ready? Daniel (04:14) Okay, I'm excited. Okay, I'm ready. David Gary Wood (04:23) Don't think, just read. Daniel (04:24) ⁓ Okay, I'm just starting reading. I'm... This is German, okay. David Gary Wood (04:30) No, no, no, read it in German and then I'll post the English. It's an attempt at German. Daniel (04:33) Hello, my name is Daniel, an extraordinary slum toucher. My beard is deep blue, inherited from the slum fields of southern Bavaria. I enjoy coffee, computer and podcasts without pants. You Yeah, ⁓ you hit me pretty well. I'm gonna read the translation now. So this is pretty good, the way. This is basically correct. So English translation as follows. Hello, my name is Daniel. Cluster Toucher extraordinaire. Like this is different than the... Because the German translation says I'm a Smurf Toucher extraordinaire. David Gary Wood (04:58) Fine. Okay, yes, it's really got that wrong then. But anyway. Daniel (05:15) I'm a Smurf toucher, but this is a cluster toucher. So yeah, I'm actually more comfortable with touching clusters. My beard is dyed with the deepest blue tint harvested from the Smurf fields of Southern Bavaria. I enjoy coffee, computers, and podcasting with no pants on, which... Let's just leave the camera as it is. David Gary Wood (05:34) You Daniel (05:35) Like, the other camera is only for my OnlyFans. Well, this is fantastic. David Gary Wood (05:38) We weren't gonna boost that on this show, but yeah. We are introduced. Does that mean, does that marker show start like a week? We're going to do the usual intro still or, ⁓ yeah, go on. Daniel (05:40) So yeah, we both introduce. We're gonna do the usual intro because we introduce ourselves but not the show so this show is called waiting for review a show about the majestic indie developer lifestyle. Join to a scintillating host to hear about a tiny slice of their thrilling lives. I am Daniel the tireless man on earth and surrounding celestial bodies and I'm here with Dave who is the greatest AI whisperer of our time. Join us while waiting for review. David Gary Wood (06:16) We are now into the show, fully into the show. But yes, anyway, Smurf toucher or cluster toucher, like. Daniel (06:25) Yeah, like that thing that you used to translate this apparently thought, what are the cluster? I'm just going to take the smurfs from the next sentence. David Gary Wood (06:31) Yeah, yeah. That's Google Translate. So, yeah. ⁓ Well. Daniel (06:40) Would you like to talk about something slightly less silly, but still a bit silly? David Gary Wood (06:44) I'm not sure it's silly, but let's see, go for it. Daniel (06:48) I have been accepted, and I think I told you this last time, but I have more news now, more details. I've been accepted to server-side SwiftConf as a speaker, which is cool because I can talk about Swift on the server, which ⁓ is something that I've been doing for years now, but I haven't had many conversational partners to talk about it. So I'm really looking forward to going to London. David Gary Wood (06:59) Hey, awesome. Landon mate. Daniel (07:14) London, London, and talking about Swift on the server, which is going to be fun. And also, I just discovered that they started announcing speakers now, and actually, two people who I really wanted to meet for the longest time in person are going to be there. So I'm going to reach out to them and be like, you're at server-side-swift.conf.conf? don't know, server-side-swift.conf? David Gary Wood (07:16) London? Daniel (07:39) I'm too, let's have a drink or something. And yeah, it's really cool. I just booked my flight and hotel. The hotel is ludicrously expensive and the flight is ludicrously cheap. Even though I didn't go to the worst airport, I'm going to the second worst airport of And yeah, well, everyone who I tell that I'm flying to London, David Gary Wood (07:42) That'd be really, really cool. Are good airports of London? Daniel (08:02) is saying stuff like, don't take, I forgot the names of the airports again, but like, don't take the one in the North because it takes like two hours to get into the city. And it's like, just like the price of the cheap flight is kind of negated by the price for the train ticket you need to pay. And so I didn't take that one, but I do, ⁓ I'm flying into the, I think that's Heathrow, I'm flying into the one in the South of the city. David Gary Wood (08:24) Yeah. Yep. Daniel (08:26) Which is also in one hour train ride, just so this is just one hour so And also like I used to have to go to Munich Airport here in Germany Where I need to go I need to write it to our two hours on a train to get to actually get to the airport and That is now negated because I can just take the subway for four stations like I can like I can go from here to the airport in like 15 minutes if I if I'm cutting it close So that is a nice a nice change David Gary Wood (08:48) Hmm That's cool. Daniel (08:52) And I did think about, can I take a train because it's more environmentally friendly and stuff. And I always wanted to go through the tunnel tunnel. But a train costs like four times as much and it takes 10 hours. And I think, no, that's like above my limits. David Gary Wood (09:04) No. No. It'll be cool. And I think it's great that you're going to be speaking and stuff as well. Are they going to be recording the sessions or anything? Daniel (09:13) Yeah. I don't know. I don't know. I think they have recordings of the previous year's sessions on their website. So there's a high chance. So if it gets recorded, I will of course post a link. I also haven't finished my presentation yet because like usually I have like one thing that I want to talk about, but I don't really have one thing. Like this is basically just lessons learned. Like this is what I learned after doing this for five years almost. David Gary Wood (09:39) That is one thing, but with a collection of things inside it. That's cool. Yeah. Daniel (09:42) Right, so it's more like an anthology. And so I'm thinking that the thread that goes through all of these things is probably something like the history of telemetry deck, because then it's a bit more of a cohesive story, you know? But yeah, usually I have one thing that I'm looking at from different sides, and this is like, oh yeah, here's how to improve your compile times, and here's how I do tests, and here's how I do caching and SDKs and whatever. So yeah, it's a bit... more mixed off mixed stuff but maybe that's gonna be cool I don't know David Gary Wood (10:12) I quite like that idea though. I certainly like the idea of you having the telemetry deck sort of narrative running through it because it really is the thing that's grounded every bit of development and learning that you've been doing with it all, right? It's been for... Daniel (10:25) Yeah, it's also kind of the reason why I started with Swift on the server in the first place because this whole thing started out as a hobby, right? I wanted to try out, like I am really, really... hang on. This is the... I don't think you can hear this. No, you can probably hear this. This is the test alarm. This is a test of the emergency system. David Gary Wood (10:42) Yeah, yeah, I can hear that. Wimp. Daniel (10:45) It will stop in a second. I can't seem to make it stop. And it's very annoying. We can think about cutting this out. now it's on my watch as well. ⁓ David Gary Wood (10:49) Yeah. We will edit this out. Yeah, I was going to say you got cellular watch, right? Is that why you're... No, okay. It really wants you to know. Daniel (10:59) I don't know, I have a non-cellular watch. But it is... I have an app on my phone that also shows these emergency warnings. Come on! Can I... I'm gonna try and stop it. David Gary Wood (11:07) Mm. Daniel (11:13) I think we're back. So I did like, I was like, okay, let's just not have this emergency alarm, like completely destroy our session. And then I completely destroyed our session. So I grabbed my phone, which is also my camera and it discon... To try silence the alarm and then it disconnected the... David Gary Wood (11:13) Hey! We are indeed. Daniel (11:33) It completely disconnected the phone as a camera and then the recording software kind of crashed. But we're back. We seem to be recording again, so it's all good. David Gary Wood (11:37) Yeah. Was the alarm, yeah, we asked a recording. Was the alarm anything serious? Do you need to go anywhere or is it just a test? Daniel (11:46) No, it's not serious at all. Like this is what was announced a few weeks ago, I think, that they would try to test out the alarm system at 11 a.m. So that's what this was. And I... David Gary Wood (11:55) Yep. Well that's a decent time. Daniel (11:58) Yeah, I wasn't really surprised. I was just expecting it to stop after a while. And apparently it did expect me to actually acknowledge the alarm on the phone, which I wasn't prepared for. so this is why this recording emergency happened. David Gary Wood (12:11) It is. Hey dude, 11 o'clock is a decent time. I say that because a month or two ago, New Zealand did a test of their emergency alarm. And they did one at a fairly reasonable time. It was like, I think maybe 6 PM in the evening or something like that. And then it went off again the next day, like I think it was 6 AM, which for a test, nobody in the country was impressed with. Daniel (12:36) Yeah, I can imagine. wow, yeah. That's harsh. David Gary Wood (12:37) Yeah. So I was like, yeah, all right, I'm up. I'm awake. Daniel (12:42) ⁓ yeah, yeah, the UK had a test, I think, ⁓ Sunday afternoon. David Gary Wood (12:48) Yeah. Daniel (12:48) which I know about because it was during the Formula One race. And people were like glued to their screens, really like very concentrated. It all exciting. Suddenly like something next to them. David Gary Wood (12:52) dear. Well, test over. All is good. Got you back here. Where were we? Okay. Well, we will link server-side Swift in the show notes. And yes. Daniel (13:03) Test over. I also, I think I'm finished with my story for server-side Swift. So how about, of course. How about you tell me what inefficient way of using cloud code you found. You David Gary Wood (13:21) ⁓ Sorry to the YouTube viewers there. That was a hand signal Daniel across the video. Daniel (13:33) Yeah, you were victorious is what she showed me. David Gary Wood (13:35) I was indeed. Yes. What inefficient way of coding am I doing at the moment? yeah, so I think on our last show, we talked about how the vibes were off, right? I was out of vibe coding mode. Is that about right? We did. I can check down the show notes. It's right there. Yes, I'd felt fallen down. Daniel (13:51) I so. David Gary Wood (13:55) bit of a rabbit hole and well since then I've been on a little bit of a journey so I started strong with this experiment and bear in mind this is like a month where are we now this has got to be month starting month three with claw code so I'm two and a bit months into it ⁓ And before that it was chat GPT, but the experiment was use some of these AI tools for development and see how applicable they are to things that I do. And just to reintroduce what on earth I've been up to with Claude. It's been a case of like, have a play with it, see how it works and then go hard out with a brand new project. to see how well it did. And all was fine and groovy for like month one and a half into this. I spent about three, four weeks in the hole tinkering away, which you are fair to call me out on Daniel. It's been a long time, like not making any progress with it. But what happened is that it was like watching like watching a cake rise and then the middle drop out of it. as a project, right? Because the proof of concept went really well. I was flying with it. It was like, yeah, OK, cool. I'm giving it all this like, I gave it all these rules to start off with. I told it what I wanted, how I wanted the project to be, and various bits of, you know, I don't want you to use this type of approach. I want you to use this type of approach, like quite explicit. Daniel (15:07) Mm-hmm. David Gary Wood (15:30) And so it looked like my proof of concept was coming together quite nicely. And then it started to degrade. And by degrade, what I experienced was every time I went to add something new into it, it just wasn't working out. It was getting the most easy things or relatively easy, wrong. And then it was also taking routes I wasn't. happy with didn't want. And so you sort of what, what I now understand is that I guess I'm kind of, I kind of reached the limit of like how much it can hold and it's context to get the job done. And what you end up happening is, is that's all your rules. That's any relevant documents, this down the other pulls it all together and your request. And then bits get shaved off and dropped off and assumptions get made. And you see this sort of regression to I guess it's regression to the mean, right? Things kind of get averaged out. It may not literally be that, but it's that sort of smoothing of the requirement that ends up with those artifacts, those edges. Daniel (16:28) So you've been basically speed running what the American corporate, like big tech corporations are now also finding out. David Gary Wood (16:36) Yep, yep. And so there are various tools to try and like, know, things that people have made that allegedly in one way or another will improve the experience, right? And I won't go into every single one of them, over the last few weeks, I've probably tried most of the biggies or the ones that at least turn up first in search results. And I haven't had any real luck in sort of staring it out. And I'm like, All right. Pretty damn close to just canning this, starting a new project, coding it myself and just pulling in the bits I'm happy with. Cause there are bits of the code that I did with this that I am pretty happy with what I say it did, but it's, know, it's in combination with me guiding it. like it made a pretty good job of, a core image pipeline in the end, not because it was good to out of the box, but because I down loads of rules about this and then. guided it to where I wanted it to be. So I'll take that for sure and pull that through to a new project. All right, last week, if I'm that close, I may as well just try one more idea. And so that idea was, I am not happy with the architecture of this app. More than anything, what I was seeing, Daniel, is like it was doing weird things with dependency injection. It was starting to do things where it was like holding onto certain. items of states in one view and then injecting it all the way across loads of other ones in like stupid ways. And it was just like, what, like, do people really code like this? What on earth is it trying to do? What does it, where has it learned this from? This is bullshit, you know? so I laid down a new project and I created an example app and an example architecture. learn, you know, taking all the things that I didn't like it doing and taking all the things that I wanted the project to have and then going, well, okay, what does ideal state sort of look like? and that I then, I then took that and actually run it back through chat GPT. Daniel (18:31) You're going multi AI. David Gary Wood (18:33) And and I had a critique and call out weaknesses. And that was actually really cool because it threw up things that, you know, were sensible. And then I was able to make tweaks and improvements. So I did that, made those tweaks and improvements and then. In between all of this messing around, I found a new approach for working with Claw Code, which I am now about, I want to say three quarters of the way through progressing with. But yeah, shall I tell you about that? Because that's kind of got somewhere a little bit. Daniel (19:08) Sure, yeah, you sent me a message like a few days ago and I was kind of intrigued. So yeah, fire away, please do. David Gary Wood (19:15) OK, so the message I sent you, I told you about this thing called Agent OS. And I'll link it up in the show notes. But the main principle of it is that it installs a whole bunch of instructions and tasks, commands, rather, that Claw can run. And then the way you work with it. is you say, okay, I'm going to create a project. So you run the create project command and then it will then prompt you and ask you to describe the thing you're making. You know, so I did that with my new project. I gave a description. I took a whole bunch of the description I had for the original app and other bits that I got and pulled it all together and just sort of went, there you are. This is the thing I'm trying to make. And then it queries you a few times and it asks you different things. But within this, the thing I did that was different, as I said, we're making this project, but I also have this example app and architecture here. And I copied that into the roots of the project folder. And then when I want you to analyze that because that's how we're going to build this thing. The next stage with working with it is that you then go and ask it to create, you ask it to create tasks and it will go through everything you've given it, lay out a load of like development tasks. And so you're kind of going like in that very typical sort of, you know, product story ticket to dev tickets kind of flow. And again, it'll ask qualifying questions and the same the other, it asked a whole bunch of things around the the architecture. ⁓ Go on. Daniel (20:44) Quick question. It's called an OS, but is it like actually an OS? Like is it actually an operating system? ⁓ okay. Because I was wrapping myself. Why would you? Like is it Linux based? Like why would you reboot into? Okay. David Gary Wood (20:49) It's not an OS, it's just the name. Yeah. Yeah, yeah, yeah, yeah. No, no, it's just a name. Yeah. So yeah, you get your tasks together. You have this back and forth with it and then you tell it to execute through the tasks. Daniel (21:07) And how does it work under that? Does it use your existing, whatever you're using, like Claude or JetTPT or whatever? OK, it's all in Claude. All right. David Gary Wood (21:13) Yeah, it's all in Claude, right? So it installs its instructions and everything else into Claude. So you're just using Claude, but you're using these specific commands. So that was the start of it. And I kind of got it to a stage of like, well, a couple of iterations working through that process to then have like a base app down that was kind of starting to look like a bit of the structure of what I was building before, but mocked up. Um, I was like, right, fine. I can either keep going like this. Go, you know, and I can go one of probably two or three routes, right? I can keep using this setup to try and get back to where I was trying to get to before. I could leave this here now and just take the wheel. Always an option. Or I can feed the project that is completely deflated on itself and bring that in. and see how good a guy would do of migrating that code. And so I obviously did a new get branch and I am now 75 % of the way through a migration, which has been happening in stages over the last couple of days because I'm not always sat there poking it. Yeah, so it is. Definitely producing better results compared to just using Claude on its own. I've been looking through the code produce. Daniel (22:31) I mean, it's better. put it like what's what sounds like days of work into like building up this infrastructure. David Gary Wood (22:36) ⁓ installing those tools themselves is like really quick. And then working through it is just follow the instructions. But yeah. Yeah, yeah. The work was really laying down the example projects and then, you know, actually thinking through, okay, because part of the problem as well is that I haven't necessarily been explicit with it as well. So then you also get, when I was doing the initial project that has now collapsed on itself. Daniel (22:40) No, but like, they're them like, with all the instructions and everything, right? David Gary Wood (23:01) I really was just trying to see how far I could push it. It was very much like, well, I actually want the vibe code or experience because people keep going on about this stuff. And I want to see how good it really is because I actually want to get a realistic view of that. know, can I want not one shot, but, know, use this to create an app without much input. and What I'm now finding is that, yeah, with the right inputs, you can get something that is a lot better out of it and with the right tooling. But it's certainly been a lot of effort to really get to this stage of actually understanding it well enough to get it to do it. Has it been quicker? Daniel (23:41) And you could do it also because like you were a you are actually a programmer you can understand the output. So it must be like orders of magnitude harder for the just wipe coders. David Gary Wood (23:52) I would say so. it's also something to be aware of as well as like, you know, just because you are good at one particular type of programming or one domain area or whatever, it's going to be easy to get tricked by these sort of tools in terms of, all right, it's really good when I was working with it on that thing that I knew lots about. Yeah, that's because you knew well enough to when to cut it off and how to direct it. You can't generalize that to every other thing you may develop in the same way. You've got to kind of be, I think, at least objective and look for other ground source truth with what you're doing for it to actually be good. Like if I did the setup and then went, make me my Android version of this app, I could not tell you to the same degree as I can with iOS whether it's given me rubbish, you know? ⁓ Daniel (24:39) Right. Right. Right. Fun stat for you, by the way. So over the last 24 months or so, the number of new projects on Git has not significantly increased. Also, the number of new websites or domain registrations has not so sufficiently increased, which would be an indicator that, oh yeah, Vibe coding just works because there's lots of new projects, right? David Gary Wood (24:58) Mm-hmm. Daniel (25:03) But what has increased is the only thing that I could find that has increased is number of lines per Git PR globally. That's the only stat that has changed since like live coding started to become a thing. And it kind of tracks with how I use AI to code these days, which is just like the small menial tasks. And it sounds dismissive, but actually it's like super helpful because David Gary Wood (25:11) ⁓ yes, yes. Mm-hmm. ⁓ Yes. Daniel (25:29) yeah, I need a test for this new controller. I need a new database model. it's like, is actually hard, like it's actually work to like, about like what the field should be and then type it out and then also add a migration or whatever. It's just like busy work. And it is actually incredibly good for my motivation to just be like, Hey, do this busy work. And then I can immediately start with the creative stuff. So I'm actually enjoying it, but I'm using it way less than I thought it would be in the beginning. I'm just like, maybe if I'm very generous, like 10 % of my output is I generated. And I think that's probably a good thing. Because like, it takes off the menial task that I kind of have to do. David Gary Wood (25:52) Yes. I don't think it's a bad thing. think you need to, yeah. Daniel (26:14) Also like the other day I used it to like convert all my tests from XC test to Swift test, which is the new Swift testing format, which is way faster to run. So that's a, that's an improvement. ⁓ And also there I ran into the thing with the, the, the, the what window, the context, the context window where I basically had to make it like, I basically like told it to make a to-do list for itself for each file individually. then don't, don't look at all the tests at once. Like look at one test. David Gary Wood (26:24) That's cool. Yeah, that's really cool. Yes. Daniel (26:42) at a time. So yeah, so that is, well, what's working these days. And I'm interested to see how you are, how far you're going, you're coming with the agent operating system, that's for sure. David Gary Wood (26:54) Well, I kind of feel like if I haven't gotten to a stage of having a test flight release of the app in the next week or two, then I really do need to cut the experimental off. like the tools that are there, they are very good for, like you say, menial tests, small bits, things that have just sort of taken you out of that creative flow that you just kind of want that thing. ready to then start poking around with like, great, really good at proof of like getting a proof of concept together quick when you don't care about the quality of the code. I still care. Like I'm looking at this and going, I don't want something that's an absolute, you know, rat's nest of code to then support if this app actually goes anywhere, right? Like that, that's just going to be a future mess to sort out for me. Like that annoys me. That doesn't feel good. So I think like, unless it's actually going in the way that I really want it to go, then I have to stop and go back to the point I said before as being another option, which will be right. New project, start again, pull in the core bit that I know is working well. That's great. And probably just ditch a whole load of these tools and pare down again. So then kind of go, well, OK, what do I really want to bring in alongside me? You know, if anything. But I'm not there yet and it feels like it's kind of working, but it's felt like that before. We had that show a couple of shows back that was straight off the back of successful proof of concept. Right. Daniel (28:12) Mm-hmm. And I still feel like I wrote, I think I wrote a message to you about this at some point. Like I wrote to you, like this feels like playing a video game. wanna see like one more thing, like I was gonna play around with it more. And I think that's the thing, like AIs, like especially the coding AIs are kind of catnip for programmers because it feels so incredibly productive and you're flying and stuff. I think like if you look... David Gary Wood (28:41) Yes. Daniel (28:43) in the, at the output in the cold hard light of day, it is not as effective and efficient and whatever, but it just is so fun to just like tell it a bit. And then you see so much, so many changes. Everything is like colorful and bright and it seems to work at first stance. And so that is like, that is like the, the, the siren call, the siren song of AI, which I think people will start to realize more and more like, then like, I think we were going to David Gary Wood (28:52) Mm-hmm. Daniel (29:11) get out on the other side of this with a new cool tool that is helpful in some situations and that's pretty nice. David Gary Wood (29:17) Yeah, I agree. As one of reasons I do want to, did want to keep talking about it on this show, right. It's because, you know, you've seen me come through the phases very quickly. think show by show almost. and you know, like, yeah, maybe on the next show, I'm like, dude, I got it in test flight. It's actually looking quite groovy. and you know, I've got a pattern down and it's working for me. but I think, you know, It's good to talk about the reality of it because, yeah, otherwise people are just sort of thinking, these things are really, really brilliant. And they can be, you know, context is key, right? In all senses of the word. And also, they're not they're not completely awful either. But I don't really know what I'm trying to say. I think really one way or the other, this is going to be the only sort of fully like vibe coded experiment. That was that was the frame of this experiment to myself. Right. Can I get something through this sort of process to test flight and then see how that goes? I think if the answer is it takes two months of spare time, well, I probably could have done that quicker myself, to be honest, at least I've. Daniel (30:21) But it is a good experiment to have, to experience this for yourself and see, okay, this works, this doesn't, so that's kinda cool. Also, I you deserve to play a video game every now and then. I've been rediscovering roller coaster tycoon. This is just for fun. David Gary Wood (30:31) True. Classic. Yeah, yeah, exactly. Exactly. And, and, know, again, that's why the experiment has, I am going to have to commit to a cutoff with it because I do want to get back into adding features to Govj and the apps I've got in the store already. Like I can't just sit in this sort of tinker mode all the time. So, you know, but on the other side of it, then I've learned some of these tools. I've had a bit of a go. That's not the worst thing. be honest with you. It certainly has felt more fun than going, I'm never going to use any of these things at all because they're awful. Like it's like, yeah there are a lot of awful sides to it and I don't really want to go into all of that end of it right now on the show. Daniel (31:14) I'm still expecting like half of these AI companies to fold once the initial gold rush is over. So who knows if we can even use them. Like once they actually have to be profitable, like it might become prohibitively expensive to use them. David Gary Wood (31:17) Yes, exactly. And that's a whole other reason to not tolerate it producing you rubbish code because you're going to be left with that on your plate in the future. know, this whole idea of like, well, you can just keep rerolling every time. And it's like, yeah, you can if you think you're getting somewhere. But, know, from my experience, chances are if you're not getting somewhere, no matter rerolling is going to help you because you're already in the entropy of a shit code base. Daniel (31:38) Yeah. David Gary Wood (31:54) basically. yeah. Daniel (31:55) Hmm. Yeah. Yeah. I think, I think like it's working for me. It's working the other way, like because I'm only giving it like small, small tasks, but at the same time it's trip keeps tripping over the edges and rough rough edges of my code base. So over the last few weeks, what I've actually done is like remove a lot of war or fix a lot of warnings, like improve the structures slightly so that it makes more sense or like, cause actually improving my code. Like I have to improve my code myself, but it makes me improve my code. So that's kind of nice. But do you want to, as a last thing today, maybe talk about the Apple event that happened recently? Because we haven't talked about it yet. David Gary Wood (32:24) Yeah, that is cool. Yes. Yeah, go on. Although I'm not sure it's going to be a long conversation, Daniel. ⁓ Daniel (32:35) No, no, but like, I mean, we are still like, have some Apple, like, or the name of the show is like, Waiting for a View and whatever. like, we should probably go a little bit into it. I found that I have way less enthusiasm for the whole show than I even did before. I thought my enthusiasm for Apple stuff, like, was at the bottom, but ever since he gave that weird... David Gary Wood (32:46) We should. Daniel (33:00) like block of gold with the, the piece of glass on top of it to the, to the guy. Like it has sunk even further. Like I was, I was like seriously considering just like not turning it on. Like actually I forgot, I was like, I was actually deep in roller coaster tycoon. And I was like, and then people started like talking on my social media and got notifications from it. I'm like, yeah, it's starting, it's starting. I missed the first five minutes of it, which so. David Gary Wood (33:06) You You You Daniel (33:26) like which would never have happened. David Gary Wood (33:27) He probably opened with Tim Cook saying good morning. You probably didn't miss too much there. Daniel (33:31) Good morning. What it didn't like, because I wanted to check, but I just like, according to the people on social media, what it didn't start with was some kind of environmental or social message, which on the one hand is a bit sad, but also telling. But on the other hand, it would have been incredibly, incredibly facetious. so I can, I can, I guess like that's, that's off. What it did have was. a 10 minute segment or 15 minute segment even of like, if you don't watch your wear your Apple watch, you're gonna die. Like you wear your Apple watch or you'll die in a horrible accident or because your heart's going out or whatever. So that was that was still there. Like they they've been doing this for a while. But yeah, let me go through the products maybe. ⁓ First up was the AirPods Pro 3, which I am David Gary Wood (34:06) Mm-hmm. right. AirPods Pro mate? Yeah. Daniel (34:26) Like the AirPods Pro are still my most used Apple product because like I have them in my ear right now. have them like they are in my ears like like hours every day. Like either I'm listening to podcasts or music or sometimes I'm just using them as for the noise cancellation. Like I'm not playing any audio through them. It's just like I like I'm someone who is like Sometimes if it's too loud, it's very distressing for me. So if I'm going through the city or whatever, I will usually have them in just to have a bit of quiet, just to dampen the sounds of the city. The sounds of the city. A little bit. I'm such a soy boy. ⁓ David Gary Wood (35:04) No, it's cool. I kind of use them similarly. I use them similarly as well. Like if I have to go around the supermarket, then yeah, it tends to be that I'll either use noise cancelling or I'll put transparency on, but I'll have some music up quite a way over the transparency. Right. So. Yeah, they're in my ears a lot. So no, it's cool that they're getting updated, but Daniel (35:19) Ahem. David Gary Wood (35:28) I don't see any reason to go straight out and buy a new pair. It's like. Daniel (35:31) So the thing was, before the event, I was like, maybe that's the only thing I might actually get because if the noise cancellation is significantly improved, then it might be a worthwhile upgrade. also, because I'm using them so much, the battery starts to degrade after a year or two. And so it might be something that I can consider. But then just like hours or just a day before the keynote, think, out comes a Mark Gurman article that says, well, the noise cancellation is basically not improved at all. And then Apple comes out and says, the noise cancellation is twice as good. So I don't know to believe. So what I'm going to do is I'm not going to preorder it or anything. David Gary Wood (36:13) All right. Daniel (36:16) I'm going to wait for someone, some publication that I trust to review them. And if they say, yeah, this is like a marked improvement, then I will con... my God, I am waiting for a review. David Gary Wood (36:23) name of the show, you're waiting for a review. couldn't resist, sorry. Daniel (36:30) that's fantastic. Yeah. But I'm trending toward not buying them, like also because like, I don't know, like I don't want to, I don't want to, like I have like better uses of my cash, I think. David Gary Wood (36:38) You You gotta wait until you put them the washing machine like everybody else. Daniel (36:45) Also, the other thing that is not announced is like my laptop is actually starting to die. parts of the, some of the ports are, are like getting unreliable. And it's ⁓ four-ish years, three-ish years, three and a half years maybe. But I've been using it a lot, especially traveling with it a lot. And it also had a motherboard replacement or main, main board. What are they called? David Gary Wood (36:54) What? How long have you had that one, Daniel? Okay, yep Daniel (37:09) Apple devices. I don't think they call it motherboard. Logic board, that's the one. So I had a full logic board replacement for some issue a while ago, and that also might have weakened the whole structure maybe. Because that's what it feels like, as if the metal structure is kind of weakened, and then things are just getting wobbly, and then the contacts of the ports are not holding properly. It's a bit of an adventure to charge this device right now. David Gary Wood (37:10) Yeah, logic board. Daniel (37:32) So I've been thinking, do I get a new laptop? Which they haven't announced at this event. And I've also decided I am not going to get one of the M4 machines. I can survive another half year. And then if they announce in the next half year or so, which is very likely, new M5 machines, I will consider those. But will also actually, because I will also look at the Linux machines a bit. But then again, I like Mac OS still too much. But anyway. David Gary Wood (37:32) rights. Daniel (37:57) That is not part of the Apple keynote, but so far I haven't seen anything that I want to buy, which is probably good because consumerism is destroying us. David Gary Wood (37:58) Yep. Mm-hmm. But on that note Daniel, what about that Orange iPhone Pro? They're almost... Daniel (38:12) That is actually very pretty. That is almost custom. Like, ⁓ the thing is, I bought an iPhone just last year. So even though I love the color and even though, like, I would love to that color, an iPhone of that color, I'm not in the market for a new iPhone at all. Because this one that you are watching me through is fantastic and good. David Gary Wood (38:14) I thought you might think that. Daniel (38:34) Also, these days it lives inside a case most of the time because I have a case that attaches it to my bike and so I wouldn't see the orange color anyway. But I love that they finally broke down and made something in exactly telemetry deck orange because we've been petitioning them to do that for years. ⁓ Also, I'm not going to go into other connotations of the color orange because that would be... David Gary Wood (38:51) Yep. Daniel (38:58) Distasteful. Yeah, I loved it. David Gary Wood (39:00) Mate, you mean like the soft drink Fanta? Daniel (39:04) yeah, of course. Anti-Antifanta. I love the color. haven't really... I watched the keynote. I watched it with one eye. The other eye was glued to Mastodon because so many cynical takes. That was peak comedy. Everything got taken apart very cynically. And people were commenting on like, yeah, this is... David Gary Wood (39:23) you Daniel (39:26) It used to be like people were like genuinely enthusiastic about the new Apple announcement and this time it's just like everyone is ripping it apart. But it was like really fun to watch that, to watch my feet while watching the keynote. But anyway, like I didn't see anything about the new iPhone that would be like an improvement for me personally. Like it has like some new features for professional filmmakers, which are not. It has a... like slightly better cameras with a huge like a hugely increased camera plateau with the iconic plateau. The iconic plateau is bigger, which fine, but I like the camera is pretty is plenty good after a year. So yeah, I'm not I'm just like, I don't see that I think that the dynamic island like the iconic plateaus larger, but the dynamic dynamic island is smaller. So that's kind of good. But and it has a vapor chamber, which is awesome technology. David Gary Wood (40:00) Yes. Yeah, no need. Daniel (40:18) But I don't have the problem right now that my iPhone gets too hot the way I use it. David Gary Wood (40:24) I mean I can make mine get very hot. I like the idea of that side of the technology that did look kind of cool. No pun intended actually. I don't know, I mean I may end up getting one eventually at some point because there's a whole circle of life of upgrade and hand-me-downs in my household and then that would kind of facilitate that. Daniel (40:27) Vapor chamber. David Gary Wood (40:48) so I get the new shiny, everybody else gets the next down shiny and so on. maybe but I'm still not fully sold yet. more like a probably again wait for the reviews and see where I am in another couple of months. no the orange looks cool. I'm not sure what they're really going for with the air. ⁓ like, yeah, I don't know that that just didn't quite. do much for me in terms of what it is like. Yeah. Daniel (41:15) The Air. Yeah, I can see that. When they first showed it, I was like, oh, this is thin. And then they showed the iconic plateau. It was just ludicrously, homongously large. And I was thinking, if you really want to make a thin knife, to say, OK, the camera is shitty because this is the super thin one, but just accept it. do this? It won't fit into pockets anyway. It destroys the look. David Gary Wood (41:26) Ha ha ha. Yeah. Daniel (41:44) Like maybe just not have a camera. Just have a front facing one, just have a shitty one, whatever. Show some courage. David Gary Wood (41:44) Yeah. You really could. Yeah, they really could actually. This is the phone you get when you're not really taking lots of photos when it is just a selfie for communication and selfies. I don't know. I really don't know. I sort of missed the days where there was no camera bump on the back of the phone at all. Daniel (42:07) Right, remember? Like, my niece has one, think. Like, she has an iPod Touch. And that thing was incredibly thin and it had an integrated camera. The camera was bad, of course, but it was unapologetically thin. And I loved it. David Gary Wood (42:19) Yes. Yeah, I kind of miss those days in some ways. I wouldn't want to go back to those sort of cameras. I'm spoiled these days in that respect. But like, yeah, I don't know, man. Like this felt to me like summing it all up because I didn't get up at 5am New Zealand time to watch it. It's like, no, I'm not getting up for an Apple event now. Yeah, it is for when it is definitely an infomercial. Daniel (42:26) Hmm. 5am is too early for this. David Gary Wood (42:46) sort of put together thing these days, right? It's not a live experience. Yeah, so I watched the Verges Apple event in 16 minutes. And my general vibe after watching it was like, feels like an S year. Which I think is probably an easy thing to say. But you know, like it's last year's models with some increments, the new cameras, think are going to be interesting. This whole sort fusion camera thing they kept going on about. That might be more interesting than I sort of dismissed it as when I first saw it. Because I don't know if that's going to lead to sort of more interesting things you can do, especially when you've got more than one of them on a device. But yeah, aside from all of that, it's another event. Daniel (43:15) Mm-hmm. David Gary Wood (43:26) Another year. ⁓ Daniel (43:27) Yeah, I feel the same way. Like S-years are interesting Apple wise because there's a ton of new technology that just gets quietly introduced like the vapor chamber, but also like they have a new way of addressing RAM basically that makes hacking the iPhones harder. And this is just cool technology, still, I won't feel it. David Gary Wood (43:38) Yes. Yeah, and that will filter through into the new laptops and things as well. Probably as an approach, maybe if it can translate. ⁓ But that's the sort of thing we see. Yeah, something gets tried in these various ways and then it gets pulled out onto all the other devices. But yeah, no, anyway, not terrible. And I think if anybody is going to get some new shiny, obviously enjoy the heck out of it. Get your money's worth. Daniel (43:57) Right. Right. David Gary Wood (44:14) your lifespan out of it. ⁓ Daniel (44:16) So yeah, new watch, slightly thinner apparently, but I haven't seen any numbers. And it has the new, the big new feature is it's gonna save you. Like you're not gonna die if you wear this watch because it will alert you to high blood pressure. The thing is though, I have high blood pressure and it's, so what it won't do is like actually tell you your blood pressure. It will just tell you like after watching your pulse for 30 days, I think you might have high blood pressure, which is not helpful for me at all. Like, wake me up if they actually have a blood pressure sensor, which is going to be way harder, of course, but still. Maybe this is a step on the way. So yeah, also not very interesting to me. The thing is though, like this is a rental watch with this rental, the hardware rental company. So I might upgrade it when it arrives there, because like if I could just send this one in and for the same monthly amount get the new one, like I might do it. But I might also not because this one is fantastic and I love it. if I can get that one in orange though, like, let's talk. I don't think they have the what in orange. Yeah, what else? Anything else they announced? David Gary Wood (45:18) Nothing else really left out to me. By the time this show goes out, they may already be on their way to people. I'm just thinking about the timing. So, you know, people, people's early orders and everything else will be shipping. So that's kind of cool. But the other thing is, of course, we will end up with liquid glass. Daniel (45:31) I don't ⁓ yeah, yeah, did actually, like I'm not running the betas on my phone because I didn't, I wasn't brave enough, but I am running the beta on Mac OS on this very machine that I'm recording on by the way. And I wanted to see some liquid glass and it's disappointingly unliquid. Like the most, like I sent you a screenshot earlier of just my desktop because you were looking funny and I thought that was cool. David Gary Wood (45:48) Right. ⁓ Daniel (46:00) But like the main thing that you will like discover on MechOS, ⁓ or least as I discovered, is that all the edges are way more rounded. And that's kind of cool. That's kind of cute. David Gary Wood (46:10) just opening that screenshot. Yes. that's kind of cute. Although does that, does that lose space out of the... Daniel (46:12) but that is the main thing also David Gary Wood (46:16) Does that lose much space? doesn't really lose any space. Daniel (46:17) Sorry. No, that's the other thing. Because if you have rounded edges with that high of a radius, you should probably move back the inside elements a bit so they have space to breathe. But they don't. So sometimes the elements on the very edge of a window look kind of weird, but I think they will update over time. The other thing is they are no longer allowing icons that have a non-squirreled shape, which is kind of disappointing. Look at them. David Gary Wood (46:29) Yeah. Daniel (46:44) in my dock, look at the things I can like fifth or sixth on their left. It's just like hidden in a gray jail now because they haven't updated the icon. David Gary Wood (46:48) Okay. Daniel (46:52) That's kind of like how I don't know. Like it makes my life easier because like if I only have on my website, for example, if I only have to display squirreled icons, I can add the shadows and the squirreled and everything in post. So this is very helpful, but it's kind of disappointing. Like MechOS icons were always the one where they had a little bit, little element that is kind of escaping out of the thing. So yeah. David Gary Wood (46:53) Right, OK. Yeah, yeah that's a shame. But hey, it's gotta be fun I guess to sort of play with some of the new stuff as it comes out. I have deliberately not ran any betas ⁓ this year, largely because I've been in tinkering with AI mode, right? ⁓ Daniel (47:25) Mm-hmm. Mm. David Gary Wood (47:32) At some point I'm going to have to and it's probably a job to do quite soon actually to just make sure that I don't have any really bad bugs or issues before the actual release comes out. But now I've kind of swerved it all this year. Maybe I should pick a sacrificial device this weekend and have a look. Daniel (47:51) I just wanted to see some liquid glass and the only liquid glass like I've really like I've expected the whole OS to be like very translucent and maybe it wasn't a previous beta or something like that. it's the only liquid glass I'm seeing is in the dock and in Apple Music where they have like the UI laying very translucent Lee and unreadably over the over the content. ⁓ So yeah, I don't know. David Gary Wood (48:13) Yeah, I've heard about that. Daniel (48:17) It's fine, it's stable, I have been using it for a while. The only thing that doesn't seem to work sometimes is screen sharing. But only like I assume they have improved or updated the API's for allowing an app to share the screen like they Remember that was last year. That was a thing Like where they had new permission dialogues for screen recording and stuff like that And so just yesterday I was talking to a customer over Microsoft Teams, which I never use And so I was like trying to my screen and then the dialogue pops up and I said like allow and then the dialogue says you might need to restart Microsoft Teams David Gary Wood (48:34) ⁓ Yeah, vaguely. Daniel (48:49) And I didn't because I never do that and always works and this time it didn't work. So maybe this is a symptom of that new permission style or whatever. David Gary Wood (48:55) Yeah. Your first mistake was using Microsoft Teams. Daniel (48:59) I have no choice in the matter. David Gary Wood (49:01) Anybody who is spoken by a true Microsoft Teams user, of which I am. Daniel (49:05) hahahaha But I have all the meeting video things. I have the obvious ones like Zoom and FaceTime. I also have the open source ones like Jitsi or BigBlueButton because that's somehow, sometimes how we communicate with people. I also have Tuple, which I haven't had a lot of use in the last few months, which is kind of sad. Yeah, but I have all of them. Same with messengers. David Gary Wood (49:30) I know we don't use any of those to communicate with to do the show because we use what we're using. Riverside, But the idea of connecting with you on all of those and then just choosing a random one when we next record, I'm kind of amused. You're not going to know where the call's coming from. Daniel (49:35) yeah, we use Riverside, which is also sharing video. I like that. David Gary Wood (49:50) Well, Daniel, we should wrap this show because clearly I've not got much more to talk about with the Apple event. Like it was an event that happened. I'm glad there's new things for people. That's cool. But it has been lovely to talk to you again, though, dude. And. ⁓ Daniel (49:50) All right. Mm-hmm. Yes, yes. It has been indeed. I'm gonna postpone telling you about namespaces until next time. Yeah, this has been fantastic. Thank you so much for your time. Thank you so much for listening. I will ⁓ only touch smurfs that want to be touched concentrually and enthusiastically. David Gary Wood (50:09) Yes. I shall hold you to that. Leave those Smurfs alone. Daniel (50:22) So all the Smurfs out there and all the non-Smurfs out there, that includes everyone I think, thanks for listening. Please rate us on iTunes and YouTube, send us emails at contact at waitingforreview.com and join our Discord, the link is in the show notes. Dave, where can people find you on the internet? David Gary Wood (50:35) ⁓ You can find me on my Instagram and that is just light beam apps on Instagram. So I'll link in the show notes. Please, we talk about the show notes so much. Just click the thing in your podcast player. Have a look at those beautiful show notes and find me at the bottom there. It will be linked. But how about yourself, Daniel, for people that don't want to go to the show notes? Daniel (50:55) plastic. Yeah, find me on Mastodon. Find me on mastodon at daniel at social.telemetrydeck.com because that is still the platform that I enjoy the most. Also, I had the opposite of a shitstorm the other day where someone was asking publicly about recommendations for analytics systems and tens of people were recommending and mentioning Telemetry Deck and it was like, that feels really nice. So say like a warm golden shower. No wait, like a warm shower. David Gary Wood (51:16) Aww. Yes. Wrong connotation. Daniel (51:21) Wrong connotation there. All right. Fantastic. See you soon. David Gary Wood (51:25) Take care Daniel. Daniel (51:26) Bye!