 All right, today is the 29th of December. This is the Minecraft DevSync, welcome. And let's go ahead and just get started with Chris Vair. What have you been up to? How are things going? I've been up to a couple of things. One was to get a bunch of data on a Selene database for wake words that existed in the old precise database. I've communicated this to Ken. He says there's enough there for him to do some of his testing and test, which is good. So I'm not going to do all 900,000, just to save some time. But I did notice that with a couple of 100,000 rows on that table that the response time of the UI has slowed a bit. So I'm looking into, I haven't done it yet, but I'm going to use and explains on that query we use and see if maybe adding some indexes to a couple of those tables will help that performance a little bit. It's averaging in the four second range to get that API call to come back. So a little long, I think. And that's only with 200,000 rows, not even with a million. So I worked on that and I also have a document now, a spec document for the agreements flow and how agreements are going to be, well, how they work now and how, and in some cases, how they're going to work with some of the new agreements we're talking about. So that is done and I will be moving on to memberships documentation for that later today and tomorrow. Okay. I'll share that document out and dev team too, so people can take a look at it and comment on it. That's great. Okay, thanks. I want to go to Derek, but he looks stunned. No, no, I'm just focusing. All right, so we've got Cusol in the recording booth making new prompts for our new setup. So I was finalizing that, well, I say finalizing, but really we just threw a bunch of extra stuff in there to give us options, because Gazz and I had kind of been going back and forth on some of it. So I went ahead and just made a bunch of alternatives and because relatively speaking, it's nothing to have say a few extra lines and we can make our decisions later. I've got all the pieces in that startup script and mostly Gazz has just been chiming in on that, but there's a lot of work for us to do to add all these extra little things or to piece back together what it used to be, but I think it's ready from scrutiny from the rest of the team. So I'll throw that out on the team, the dev team chat. And then other than that, we've got some things that came in recently, the cameras, and so I pulled some of those out to test and Josh's got a couple things he wanted me to change in the laser cut design to make it like almost impossible to assemble incorrectly. There was still ways to assemble it incorrectly, so I've been tweaking that as well. So yeah, that's been basically what I've been up to today. Okay, my understanding was that he wanted it actually impossible. I'll leave that for you guys. The one thing that we can't do is the, we can't do the, if they put the bottom plate down upside down, then you can legitimately, you can basically assemble it inverted, the whole, all the exterior surfaces in and the interior surfaces out, like the... I know that because I've done it. Yeah, if you flip every part 180, it can't be assembled inverted. But yeah, but I think that we solved that by on the bottom etching in this side up, right? And that resolves the problem. So if you can read this, you're an idiot, we put on the bottom. So yeah, anyway, that resolves the problem and it's like the XKCD cartoon. My gift is the entire universe with the exception of one set of headphones, right? Like I wrapped the gift in the entire universe and everything in it with the exceptional one set of headphones, I wrapped the gift inverted. So yeah, anyway, but yeah, I'd like to get that cut because we are now running up against like rush timetables and stuff in order to ship. So sooner or better, so we can send it off to the cutters. All right. And I'm, yeah. And one other comment, Derek, as long as he's recording stuff, can you get him to re-record the phrases that you already have, just in case there's any difference in tonality or whatever of the recordings? Yeah, we could probably do that. If you get Ranger School changed his voice to be deeper and more manly. Yeah, so, okay. Yeah, well, we'll ask to do that for sure. Cool. All right, Ken, how's it going? Difficult. So I've spent the entire day working on debugging this PanaCore images, seemingly a timing bug. I don't know. I have more questions and answers, but from a high level, what's happening is that precise is taking too long to load and it's falling back to the pocket spanks, wake word. And I haven't seen pocket spanks recognize a microf yet once, but that's besides the point. There's, I'm inside of a package. We have a lot of packages in this repository that belong to Matthew and they're in a Matthew repository or GitHub somewhere. I'm inside one now, it's a package extractor utility and it calculates the MD5 checksum on the image and it compares that against the MD5 checksum that's up in the cloud to determine whether it has to download or not. And it's just taking forever to calculate that checksum and it's not clear to me if that's because somebody's writing to the file at the same time or what, but it's just, it's got a 10 second time out. I bumped that up to, I don't know, a couple of hundred seconds and it's taking about five minutes to calculate the checksum. So what it's doing is it's timing out above it and not a wake word factory and saying the size is taking too long to load and then it's going to pocket spanks but the calculation routine continues to run for like another couple of minutes, that's wrong. And it just continuously, regardless of what tries to download and install the EXE every time on boot which I don't think is correct but maybe Gez can enlighten me on that. So I'm in the middle of this pretty good and I'll be in here until it's fixed. Okay, great, thanks for that update. That to me sounds like an issue that's just right for fixing the architecture but I'll let you figure that out. Yeah, I just, it doesn't make any sense why sometimes it would work and sometimes it doesn't, it's certainly not slow fetching the assets, it's slow calculating and I don't know how it ever worked which is why I suspect that there's some sort of timing issue where somebody is downloading and writing that file while the calculation guy is reading from it, but I don't know for sure. We shouldn't have to do any downloading. That's the thing, like if the system boots and all the assets are on device it shouldn't have to download anything. Yeah, that's the other problem is that this guy was paired and running and everything was fine. So, you know, it's a problem that surfaces if you power it down and power back up on first boot. And I think Ricardo may have discovered that if you shut down the, I don't know if it's the voice service or the audio service and started to get it soon to work which again makes no sense to me now that I see what I'm seeing. We could have multiple issues, but I don't know I'm about 10 hours into this and it's probably about a 50 album bug. So I'll figure it out. It's just timing and probably re-entrancy and who knows. Okay, well, I would take a second to look at whether or not they should be doing any of those massive calculations and whatnot to start with. Yeah, I think at a high level there's two things that there's one is that we try and preload so it doesn't have to download it as you said. The other thing is that we've had on the list to change the way that Precise has delivered for quite a while because the current system I think is kind of stupid. As in it was, it works most of the time but it's caused so many problems over the years or over the, yeah. So I think finding a better way to do that makes a whole lot of sense. Yeah, well I'm sure you're on top of this but why it would happen in the Panticore universe versus but not in the other implementations is also interesting. But ultimately, not really important because the architecture should be such that we don't have to download any files on boot. This should just work. Yeah, I agree. It's kind of timing related probably but maybe because they're running in the system de-services instead of just as applications, something like that. Gez, is it supposed to always try to download the latest version on boot? Let's chat after, yeah. All right, thanks. I think let's get through this and we'll stay on that. I'll see you after the meeting when I can talk a little bit. Yeah, yeah. Okay, great. Anyway, that's what I'll talk about and we'll be working on for the foreseeable future. Awesome, thanks Ken. So Gez, I heard you were in a meeting with Panticore. How did that go? Yeah, it was just looking at the flint management stuff because they went through a lot of that with Chris and already around the API. And so I mean, I got a high level of the API stuff but I didn't want to go into too much detail there and more looking at how we're gonna manage all those channels and release management and all that sort of fun stuff. We just parked it at a good point for today and we'll get back to it. But I'm gonna start using the flint management internally so that we can play around with that and probe it appropriately. So I'll give you guys a different link to use on devices shortly. Devices that we're just using for end user testing not for development purposes. Other than that, I created a new repo which feels a bit dirty. But as we've been talking about, wanted to pull the Wi-Fi stuff out of the Mark II skill given that it's not limited to the Mark II. It should be used on any future devices as well. So we now have a skill Wi-Fi connect which has all those UI screens and prompts and things for how to set up your Wi-Fi. I also added in a very quick intent in there. So if you say test Wi-Fi connect then it'll just run through all the screens quickly just so that you can see them all in order and all that sort of stuff. I still need some work and I still need to hook them up to the deboss messages for network manager so that they're triggered at the appropriate times but that'll make the device boot much prettier and nicer. And yeah, once we've got the Wi-Fi and startup script nailed down 100%, I've already shared the link with Panicor but there's still a little bit of discussion in there about the details so I wanna make sure we are giving them what we really, really want, not we think we want this but it might change. So yeah, once that's down, we can start pushing on them to get that end-to-end happening. But I also had a good discussion with them last night about things like this bug that Ken's working on and that we're not sure why that's what's going on there with the mic failing at the moment and until we know whether it's core or whether it's something in the Panicor's containers then it's everyone's problem and we all need to have that as our number one priority. And yeah, so, I think there's some progress there soon. Okay, is that something Chris and Eric could help out on? Which bit, sorry? The timing issue with the mic's not working. If, yeah, if you've got a red four, I think the limitation at the moment is we need red fours to test it on because there's like, none of it will work on my red three. So there's no way to like, you know, know what's going on there. So yeah, if he's got a red four, then that would then can help there. Yeah. Okay, you don't have a red four, do you, Chris? No, my red four is in a box. Getting ready to make his way to gas, so. Ah, well, maybe we could wait a day and you could spend tomorrow helping Giz out because it'll take a week to get there at least, right? Especially with the holidays. It may be useful. I don't know another set of eyes on it. We got to get this going. So if there's anything in the core that you could, you know, help figure out, you know, you know the architecture as well as anyone. Yeah, I could assemble, I could take it out of the box, assemble it real quick. Maybe without it in the plastic, just get the electronics all hooked together and take a peek. Yeah, that'd be good. That certainly takes priority over the documentation and the user management stuff. You know, we'll tackle that next year. Okay, I will start on that after this meeting. And Dan, if you want to. Well, maybe you can stay on with, Ken and the three of us can stay on and have a chat about what's going on. Okay. All right, great. Sounds good. So the only update I have is that tomorrow I'm going to place the order for the SJ201s. We're going to send out components for 400 units and we're going to have the first a test batch of 20 made to start with. And then if those come back good, we'll have them spend the other 380 and we'll be on our way. Okay. Any other things people want to talk about today? Other balloons might drop from the roof behind you. I haven't placed the order yet. So maybe. Two things or three things, I guess. So one, when are we going to have a couple of devices to send to Rollover? Two, when are we going to be across the line where we can set up devices in our kitchens and have them work? And then three, Derek, you and I need to get the plastics done today and out on order, like today. So if you can send me upgraded drawings, I will cut them quickly and then put them together, double check them, and then we need to send them out. So, and the bottom piece should have this side up on it. Was your fake SJ201, did it actually dock to the Pi okay? No, I stopped assembling it when I realized you could assemble it backwards. I can do that if you want me to. Did it dock to the Pi? No, I don't know, it has holes in it or not. It's good, I actually have the holes in it. Yeah, that's a pretty small part. Well, yeah, but the reason I made some changes was so you could access the ethernet cable and all that better, which you need to have the Pi in there to see. So I wanted someone to double check that for me. Okay, I will assemble everything if you can handle the reversal stuff, but I still would like answers to the other two questions. Rollover, date for shipping stuff to them that will work? Yeah, I'll give you that tomorrow. Okay, and then on the software side of things, when are we gonna be across the line to like, hey, we can start, non-technical people in the team can start using it. Well, that sounds like, you know, both, well, it sounds like we've got at least two software bugs that we need to figure out, right? Whatever Ken's working on. And then the mic issue. Okay, so then I'll, since questions one and two do not have answers, it sounds like the answer to one was we'll have an answer tomorrow. Two, is there a Chris, Gez, Chris Vair, Gez, Ken, like, when? Well, they're gonna talk about that issue right after this meeting, so. Yeah, if we knew what the issue was, then we could tell you that. Okay, I don't need to know when it'll be fixed. I need to know when I'll know when it'll be fixed. I'm happy to extract that away, or? Oh, you know, you know. I guess the best way to answer that is you'll know as soon as I know. Yeah. All right. Keep tuning on the DevTain channel. All right, I'm starting to get, we've been close for a while, so it'd be great if you could get the line. So all right, I will get out of your hair then and let you give you time to go and do that. Derek, do you wanna, I'll assemble this real quick and then get back with you right away. It's the old 80-20 rule. The first 80 seconds done fast is the last 20% that takes forever. Yeah, I know, but we, like the money unfortunately doesn't work that way. It flows out every two weeks. So yeah. All right, cool. All right, talk to you soon. All right, thanks everybody. We'll talk to you tomorrow.