mergeconflict390 === [00:00:00] James: Welcome back everyone to Merge Conflict. Happy holidays, almost happy new year. Oh my goodness. It's Christmas Eve here and I don't, we're not wearing any. Are you wearing any Christmas? I got nothing. I got a Microsoft shirt on. Merry Christmas from Microsoft. I [00:00:13] Frank: wear my galaxy shirt and people see Christmas stars in it. So that, that's, this is my Christmas sweater. This is, you know, as, as festive as I tend to get. I [00:00:23] James: don't, I don't think I have any Christmas attire. Heather, I bought Heather a, um, a holiday sweater. She wears once a year, every year, which is a Drake, a Drake one. Um, and it says, um, yeah, it has like, uh, uh, some rhymes on it, some, some holiday verses, if you will. But, uh, she just wears it every year. And I was like, you know, just for one day and that's it. But I was like, that makes sense. Cause you're probably not going to want to wear. This Christmas sweater in all year round. I [00:00:57] Frank: don't know. The problem with me is I, I love my one actual Christmas sweater and I do want to wear it all year long, but people give me weird looks so I don't do it. So I always look forward to the Christmas season so that I can wear. My sweaters. Yeah. Plus I'm a cold weather kid. I [00:01:13] James: have a lot of sweaters. I'm just going to, I'm going to put some, I'm going to put some on after this. All right. Well, let's get into it. Happy holidays, everyone. If you are listening is on Christmas. Awesome. Thank you. If you're after Christmas, that's great too. Um, before we get into it, this is episode. What is it? 390? 390. Who knows? Who [00:01:31] Frank: knows? Can you even count that high? I'm not even sure. Like, thank God we have computers that can tell us these numbers. [00:01:39] James: 390. Now we have, uh, put out 395, because we had some bonus episodes here and there. And we've had so far since our origin, 2. 5 million downloads of our pods. So thank [00:01:54] Frank: you everyone. 2. 5 million. That means like what we're saying is being heard by people. I have to keep that in mind. Sometimes I have to be careful of what I say from time to time, because wow, 2. 5 million, you're all awesome. Thank you for listening and watching. [00:02:12] James: Yes, we super appreciate it. And, you know, through the years of ups and downs, you know, there's highs and lows of episodes and have we've gone, you know, we have a. Very dedicated folks out there. You know, people are always interested in how we do, you know, every month we get about like 20 ish thousand downloads of all of our podcasts combined, which is great in our episodes, you know, pull in somewhere around four to 5, 000. The [00:02:36] Frank: episodes where you pick the topic, we get a lot of viewers and listeners. Well, no one shows up for. But I get it. I understand, people. You're here for James. It's the James show. [00:02:46] James: No, now we are specifically coming up on another big milestone besides 2. 5 million downloads, which is we're almost at 300 hours of content. We're at 287 right now. [00:03:01] Frank: How many CDs is that? Like if we wanted to do the box set, how heavy would it be? [00:03:06] James: That's a great question. It is 12 days. So, you know, I mean, I think I don't want to do the [00:03:12] Frank: math. I always mess up the math. [00:03:14] James: I think it's, you can fit, I don't know how many, like 75 minutes, 70 minutes, something like that. So it's quite a lot. So if you did two, two 87, I'm 60. Divide by 70. We're talking 250 CDs easily. [00:03:27] Frank: So that is worthwhile. I think you all should have save up your pennies to afford the box set that we'll probably never release. [00:03:34] James: Now that's what I call merge conflict. Number two, yeah. Anyways. So I just want to give some fun stats to kind of kick off the show, you know, and of course, thanks to our amazing sponsors. If you've heard throughout the years, and of course, our good friends over at Syncfusion who've been with us from the very beginning, which is amazing. We'll see if they're back in 2024. I have to imagine. Yes. Uh, other fun stats. Most folks are from the U S we do have 9 percent from the UK, 6 percent from Germany. Um, I guess that's going to be another 5 percent from Australia, 4 percent from Canada, 4 percent from Brazil, um, 3 percent from Poland. Ooh, 3 percent from Croatia, 3 percent from Sweden. Hello. [00:04:17] Frank: So we need to learn German, Portuguese, Serbo Croatian. What else was in there? We need to do some specialized [00:04:26] James: episodes. Whole bunch of stuff. Now, as always, Apple Podcasts is our number one platform in which people listen to us. However, uh, number two, there's a bunch of like other, you know what I mean? So there's a bunch of other. Uh, that's 29 percent Apple podcasts. That's quite a lot. [00:04:42] Frank: They still run the market, [00:04:43] James: huh? Yeah. 11 percent overcast, which means that that must be 40%. Just on iOS, as far as we know. And then there's Pocket Cast, which is the podcast application I use. And Podcast Addict is like, uh, 3 percent Castbox, 2%. I feel a little dumb. I [00:05:05] Frank: don't know Podcast Addict. I should look that one up. I [00:05:07] James: know Castbox. Yeah. So that's kind of cool. Um, let's see any other fun stats? No, that's all I got. So, uh, average duration, 43 minutes, holding strong. So. There you go. [00:05:19] Frank: For our 30 minute podcast. So in other words, we [00:05:22] James: never hit our mark. Great. That is correct. Yes. Uh, our very first episode was put out in, uh, 2016. Yes. June, July, 2016. So seven years ago, seven and a half years ago. Wow. That [00:05:39] Frank: was nearly, well, two administrations ago. A measured time in American administrations. [00:05:47] James: Now, if this is your first podcast, every 10 episodes, we do lightning topics. And that was our first one right there, fun facts and stats from marriage conflict. All right. Um, Frank, Frank, Frank, let's talk about holiday hacks. Uh, I got two, um, over here, we got. Lightning topics, by the way, six, six ish topics, five minutes each. There we go. Uh, two hacks. One's not a hack, but I did pick up, I'm really into these little emulation, uh, stations type of things. And we talked about this, the, the retro, um, one that I had before. This is a R 36 S these are a little, this is a Linux based one. My other one was an Android one. I had a little Gameboy SP one. This was. 35 from AliExpress. It did take about a month to get here, but 35. [00:06:34] Frank: How many viruses did it come with? How many viruses do you get with it? Uh, it came [00:06:37] James: with, uh, uh, it came with 32 gigs of games on it. It came with 5, 000 games. Those are 5, 000. I had to delete all those and put my totally legal, uh, actual ROMs on there. Now this is a fun hack because these little Linux, um, systems, unlike Android. Have on one end, they can do dual boot, but on one end I have, I'm pointing here on this little, this looks like it looks like a little Game Boy with like two little knobs. Um, one SD card for the OS and another one for the games. And what's cool is that you have to go through this crazy formatting process and you have like dual boot layers and all these things. And you sideload these images of this, uh, one is called ARKOS, A R K O S. I know that one. You grab the distributable from, you feel like you're hacking the Gibson. You're like, grab a file, an image, flash this thing. All these things scroll up the screen like crazy. And then boom. You totally have, uh, games on here and this can play all the way up to PlayStation one, which is bananas, um, in general, but it's a nice little one here. And I'm going to hold it up. Um, now this isn't like the most powerful system in the world. Uh, but it, it, um, for 35, I was like, I'll give it a go. It's got like a mono speaker here and has this really beautiful UI. That's the Arc OS. So I can go in, I just, uh, use this another, another application for all my ROMs on here. It's called, um, uh, I'm going to look it up. It's in my downloads folder built with NET Framework. It's called Scraper, S K R A P E R, and what it does is it will pull all the artwork and images for all of your ROMs and create the gameless profile for you. And this is pretty cool. It does a little box art. And if you let it hover, it does a little, uh, video on it. Um, and that took like eight hours to, to, to scrub. And it calls a server in France. Um, and I became a, I became a Patreon, uh, so I could download things faster. So, and you can become a Patreon or mergeconflict. fm where you get stuff, bonus stuff. Anyways, that was one of my holiday hacks. Other holiday hack before I get into yours. Let me [00:08:45] Frank: interrupt. Let me interrupt. You've been pitching these things for a few years now, and I got to say that's the best looking one so far, especially cause it's Linux. I think you might've actually sold me. On this one, I think I might have to pick one of these puppies up. Hopefully you'll put it in the show [00:08:58] James: notes. I will put it in the show notes. And I will, um, the Retro Game Corpse is the one guy that I listened to a lot on. There's a few different YouTubers, but I like him a lot. He's based out of Hawaii. Um, and. He has like his kind of favorite systems of the year. So I'll put that video in the show notes and it's all different ranges. Uh, cause you know, this was 35 bucks. Uh, the other one I have is like a hundred bucks and they can go all the way up to like several hundred bucks. You know, he talks about like the, the, the bigger ones, but yeah, this is really nice, you know, it's not a hundred percent pocketable, but if you wanted to go a little bit more, you could, I like this one a lot. Cause I wanted to have something really nice for Game Boy games. And, uh, this is, it looks like a Game Boy basically. So, and it's, it's pretty nice. It's a little. A little clicky. [00:09:43] Frank: Yeah. I mean, for 35, they're either going to cheap out on the buttons or the screen or the CPU, or there's so many places to cheap out. But, um, the fact that you were able to show that on video and the screen was coming through tells me it's actually a half decent screen. Yeah. So I'm pretty excited for that. Yeah. It doesn't want a tiny pocket Linux [00:10:01] James: machine. It's very true. Yeah. It's really crazy. And like the, the ecosystem, as you learn about it. It is very, very fascinating. So that to me, that's half of the hacking is like learning about the emulation and learning about the other stuff. Retroarch is the, is the system and the emulators and all this other stuff. It's really crazy. Interesting. Uh, in general, and I feel like I probably won't even play it at all to me. The fun is setting it up and I probably wasted hours upon hours. I mean, that, that's the fun part. You know, Like, do I, do I really want to play, um, you know, the Simpsons, uh, MAME? Maybe, you know, but I could. I played that. [00:10:34] Frank: I put a lot of quarters in to play the Simpsons. That one I would actually want to play just because I spent so much money as a kid. [00:10:39] James: Yeah. It's pretty neat. Uh, my other hack that I'm going to do live on my YouTube, which is, I'm finally, Frank, it's happening. I am going to migrate my cadence. Dot Net Maui. It's happening. It's happening, people. It's going to happen. [00:10:54] Frank: I'd like to say you're being forced into it, but we've all been forced into it. And we were given a one year reprieve, but, um, yeah, we all, we all need to be doing some live streams of porting our apps over. Uh, so I'll definitely tune in for yours. I want to see, I want to see what frustrates James. You're always so smooth and calm. I want to see what actually gets you. [00:11:15] James: I'm a little worried because it's, it's, it's, it's. It's not, it is a simple app, but it's also a little bit complex app, but I have a, a library that I forked basically that does all my charting and graphing. And it's all based on SkiaSharp and I'm a little worried about that one. So I might have to pull, pull in a library for my good friends at Syncfusion. Not sponsored this week's pod, but could be. Um, and just maybe swap that puppy out in general. So, but we'll see how it goes. What about you, Frank? You maybe have an update from last year? I don't even know what it was last [00:11:48] Frank: year. This has been a very, very busy winter, so I haven't gotten to my own holiday hacks. But I am now, um, staying with my mother. So, I'm going to have a little time to kill, but I haven't come up with a holiday hack. I do want to give an update. Um Those who pay attention to the show, I have my automatic magic AI machine learning thermostat. Now, most people are smart. They just go to the store, buy a smart thermostat that automatically learns, gives you all those features. Yeah, James, you're pretty smart. Um. Otherwise, there are people like me who just want to build everything from scratch. I built my own smart thermostat from scratch. It was awesome. It is awesome. It learns. It learns my habits. It sets my house temperatures. It's fantastic. Except it stopped working. This is always the problem with these kinds of hacks is, you know, it actually takes a decent amount of engineering to make sure something survives. And I was always wondering how long would my thermostat survive? And it survived one year. I'd like to tell you. Why it stopped working ? Yeah. Uh, it's kind of interesting. So my biggest design flaw was every request you do to the server, it examines all its data, trains all the machine learning stuff in real time, and then sends its response back to the little device. That was fine. It took a few milliseconds. Then after a couple of months, it took a second or two after a few more months, it took five seconds, few more months, 10 seconds, all of a sudden I was hitting TCP IP timeouts, how do you fix TCP IP timeouts? You increase it. So I increased my TCP IP timeouts to 30 seconds. And finally, I think Azure just decided, no, you're taking way too long on these requests. And so my biggest fault, uh, that I did was it was, uh, consuming all the data throughout the year, which is every five minutes, it has like 10 data points. For an entire year, load all that into memory, train something, give a response. And so I, I need to do a little bit of re architecture. I think the fundamental, um, the devices themselves are fine, the server is fine, it's running on Azure, it's going to be fine. Uh, it was just a bit of a design flaw that I put in there that it could never really summarize its old data, or it would never Retrain a network or it would never, yeah, you know what I'm saying? It couldn't incrementally update. It was doing everything from scratch every time. So I need to do a little bit of re engineering, but otherwise I want to say, A, it is awesome. It was awesome. It is awesome. B, it needs a little re engineering. I [00:14:27] James: think that you could probably just train it off the last month or last two months, cause seasons change. And that would probably be pretty decent. The other thing I'm really fascinated about, if you wanted to. Bring it in is to tie in other smart thingy. So for example, mm-Hmm. Uh, switch bot, uh, which is a, a little IOT devices. All of their devices, um, are API first. So you can actually get a key that's actually pretty unique in the world of iot. 'cause most IOTs are Rob Slap. Don't do it , but, uh, switch bot. They switch bot. They have a full rest, API and you can get a client consumer thing for your devices that you register. Yeah. So for example, that means that you could have a sensor. That you get off the shelf, or you could have other lights or things like that. And you could integrate it into your things. So you could say, okay, as I'm entering this room, turn this off, or as I'm entering this area, make it different things. And it could learn perhaps where you're at in your house. So that could be kind of cool too. [00:15:26] Frank: Yeah, that, that's the true smartness when it can start predicting you and your habits and things. The other little, uh, glitch that I've run into, it's not even a glitch, but for instance, I just bought a new space heater and the space heater itself is smart. It hooks up into the Amazon family of Dingus devices, but sadly, my thing doesn't talk to those. So now I need to start writing some API bridges where it talks to the Echo. Echo, uh, APIs and things like that. So it can interface with other smart things. It was always assuming that it was controlling a dumb device. I want it to be able to control a smart device also. So those will be my upgrades to last year's holiday hack. [00:16:07] James: Nice. Yeah. And my holiday hacking is pretty lightweight because I only just taking this week off. So really it's like. Christmas and New Year's, and there's only a few days in between family time. So I'm just trying to narrow it down, if you will, uh, which I think will be good. All right, Frank, what do you got for me? I'm ready. Uh, [00:16:25] Frank: I want to give, well, a little bit of an update again. I guess I'm in the update mood this year, but friend of the show, Nat Friedman. He has a little, him along with a few other people have the Vesuvius challenge. Have we talked about this much on the podcast? [00:16:40] James: I'm not sure. I know that Nat. While back, and this might be the same thing, maybe it's different. He wanted to find like these. Old things and do machine learning on it to like, it describes in the languages and the things off of it. Good, good, good summary. [00:16:55] Frank: No, not a great summary, but you did hit it. Um, there are a bunch of scrolls found, uh, papyrus scrolls. Uh, historians hate papyrus. Papyrus is awesome. It's our first paper. Great. Uh, but we hate papyrus because it deteriorates, falls apart, turns into junk. And, uh, a bunch of old papyrus scrolls were found, but they're so fragile. They're rolled up into a scroll. They're so fragile. You can't unroll them. So we can't read them because you got to unroll it in order to be able to read it. So the Vesuvius challenge was someone did a deep scan of them. I'm going to say x ray, but I'm not even sure if it was x rays or what, but some, some frequency of photons was tossed into these things and we got some 3d imagery out of them. The trick is a scroll is not 3d. A scroll is 2d. wound up into a cylinder. And so the Vesuvius challenge has been to train machine learning networks, whatever, use whatever technique you want, and try to go from these high energy scans into something readable. Uh, it's most likely Greek. Um, could be, it could be other languages, but it's most likely going to be Greek. Even though it was found in Italy, uh, they, it could be Latin or Greek. And so the challenge just had a really good breakthrough. Um, before, before this announcement, um, they had deciphered one word, the color purple. Someone was talking about the color purple. It could be anything. It could be a royal person bought a new shirt. Or it could be some trading of some mollusks from, uh, Phoenicia. Who knows? Who knows what that color purple was involving. Uh, but there has been a new breakthrough where someone did What I'm excited to see. They did a full 3D reconstruction of the 3D scan and then used nice mathematics to unroll, like virtually unroll the scroll. Wow. Which is really cool. Uh, so it, it's a real trick of figuring out what is text and what is damaged versus what is background. A lot of the math then to figure out what layers. 'cause if you think about something rolled up onto itself, it has multiple layers. Try to. Try to discretize those layers, unroll them, and then find someone who can read some ancient Greek and give a shot at it. And this should be a really decent corpus of text coming out of, I think it was just one scroll that this was done on, but again, you're developing a technique. So hopefully you can use this technique on other things. And just from a historian's perspective, it's really cool to get a whole new set of, um, Text to read, you know, um, archaeology is fun, digging through rocks and figuring out what people were doing with certain pots, but it's so much easier when we have written down words and people are like, I bought a purple shirt yesterday and you're like, great, good job, buddy. And so I'm excited for this new, um, well, translation and decipherment to come out. Really, bravo. Um, I'm, I'm sorry, I'm totally blanking on the person's name who did this, but how cool is that? Using machine learning for, you know. Good instead of just fun. [00:20:13] James: Yeah, it's wild. And I mean, there's like a million dollars in prizes and sponsorships and, you know, there was, uh, Nat was doing something else. Yeah, before this as well, it was like something similar, but different, but yeah, this is really cool. Um, and yeah, I think there, yeah, all the first letters were given away and there was the grand prize or whatever. Yeah. So this is, this is neat. I'll link to the main website, which is scrollprize. org, uh, which is neat. It gives you a lot of the background to like about Mount Vesuvius and then like when it was discovered and like the. I mean, this is totally not my world. I'm not a history person at all, but still cool. You know, when things that are like, you never thought in anyone's mind would ever become unraveled, right. It's sort of, uh, like a national treasure. [00:21:01] Frank: It is. It's also to see, um, one of the things I'm most impressed. Impressed about with archaeology is from such little evidence. If you just keep digging and thinking hard and digging and digging and digging, you can actually reconstruct these stories and things about people from ancient times. And it's just really wonderful to get a glimpse into the past. [00:21:20] James: And it's cool just to see how technology evolving enables us to do this, right? Just a few years ago, you wouldn't be able to do really inspiring stuff. Yeah. So cool. All right. On my list, I'm recording, I'm finishing a video right now, which is my favorite. Visual studio features of 2023. Um, I have a list here, but I do think I'm just going to break down. I'm going to give people two inside scoop and two. Okay. Um, and my number, my number one feature that I think. As a mobile developer, as a backend developer, is my favorite. Not only is it integrated into Visual Studio, but it's also a CLI, which is DevTunnels, Frank. Um, it's so good. Okay. So DevTunnels, it's like Ngrok, you know [00:22:11] Frank: Ngrok? James, you are saying words that mean nothing to me. You better explain yourself. [00:22:16] James: Okay. Do you know, like when you debug an API or a website, it runs on local host? [00:22:22] Frank: Oh, yes. I [00:22:24] James: love local host. What if you wanted to create a tunnel to your local machine for development purposes that would give you a full URL? That you could then use like in a mobile app. So you can like test on your phone. I could give it to you to test out. It could be a website. It could be an Azure function. It could be a teams app. It could be an API. That's a dev tunnel. So it's actually integrated directly into visual studio. There's like a VS code extension as well. And there's a CLI as well, but what's neat about it is that you just go into visual studio and you can create a name. It, it can be persistent. So if every time you open your solution, your dev tunnel URL never changes, you can add authentication onto it. If you want to, you can make it. Public private, or just your org and expose it. So this is amazing. And I show this, uh, often I've shown it off before in a video I did, which was my monkey, uh, application that, um, I often use, it has a backend and a front end it's like, well, the problem that you often have is that, yeah, I'm running everything on local hosts. Well, I can't run it on local hosts because HTTP certs and this, what if I went on a device? I can't talk. It's really, you can do it, but it's hard to talk to the local host. So when you debug your backend. You can give it this unique URL and then put that into your mobile app or do other things. Dev tunnels is just built right in. It's super duper cool. And the community edition, and there's a CLI, so it runs everywhere. So if you want to run this and integrate it to who knows what types of scenarios, you could totally do that, which is pretty cool. So. That's like beyond copilot and copilot chat, which has been around for a bit, but integrated super nice. Now, like those are my favorite IDE features. Do you have any new IDE features? Any IDE, it could be any IDE, Frank. It doesn't have to be, it doesn't have to be anything. You got a favorite Xcode feature? [00:24:06] Frank: Oh, okay. Well, Xcode feature. No, Xcode has no features, James. I don't know if you're aware of this, but Xcode does not do features. No, I'm just kidding. Um, I. If I were to say, well, I'll give an old Xcode feature. It's old to everyone, but Xcode Cloud has been a really nice CI CD. It's nicely integrated with TestFlight and Apple does keep increasing the capabilities of the CI CD part of Xcode Cloud, and they have really nice actual IDE integration into Xcode. As NET developers, we try not to spend too much time in Xcode. But when you're in there, you do want it to be a good functioning IDE and to get your job done and all that stuff. And so I'll say, actually, they did a really good job at the integrated part of the IDE where, um, you can go view the status of everything. Right in the IDE, right in a panel, everything updates really fast and they're going to start supporting other languages and things like that. So before it was really the purview of Swift and Objective C, but you'll be able to use those cloud features with other languages. Let's say like maybe C sharp. And so, so I, I think it's really cool that, um, Apple is expanding their minds and allowing us to use other technologies than their normal standard, standard Apple technology. So bravo to them [00:25:39] James: for that. That's super cool. No, I saw Tim Sneeth tweet about that. They were going to open it up to other programming languages. And that just had me thinking immediately of like, yeah, iOS y type stuff and Mac type of stuff to get it out there, which would be really neat. Uh, because you know, it's like. Honestly, to build your Donnet, Maui or Nette application, you need to install T net, which is really not that hard to do. No, hopefully. And then to install a workload is like super easy. I mean, I do it on GitHub actions all the time, right? The only difference is like, will Donnet be installed? Game of Switcher is like you just have to install it manually, which. Probably isn't even that hard either. And the cool part is that it's now not, it's not free, but there's free credits, um, for every Apple developer account. So I feel like, Hey, for 99, I'm getting some value finally. Yeah. Besides honestly shipping apps. Yeah. GitHub [00:26:29] Frank: Actions is good, but, um, you do have to do a bit of work to like authenticate with the Apple servers. If you want to do like a test. Test flight deployment. Yeah. Um, it's not hard. It just, it gets a little bit annoying. Honestly, with every repo you create, you got to add all your permissions, your search, your everything. And yes, I know everyone, there are automation solutions for all this. There's extensions for all this. I know. But each one is slightly painful in their own way. And so it's nice to get, uh, some first party support there. So I'm not, you know, whenever my CICD builds fail, it's 80 percent of the time. It's because one of my certs expired or something silly like that. And the more I can reduce that, the. Happier a developer I am. [00:27:19] James: Yes, I totally agree. All right. What else you got for me? You got another AI thingy? What's, what's on your list? I do. I [00:27:24] Frank: do. I'm sorry. I'm having my fun AI time. Another update. You ready for another [00:27:29] James: update? I'm ready. Full of updates, Frank. I'm ready. Hit me. [00:27:33] Frank: Uh, what, what do you imagine when I remind you of the Tesla Optimus robot? [00:27:39] James: Oh, it's, is this the. Well, no, I don't know. Is it, is it, is it when like Elon, like just brought out someone to dance on stage? Yes, I'm pretty sure. It's that, [00:27:48] Frank: it's that, except now it's actually a robot. [00:27:52] James: No, are [00:27:53] Frank: you sure? Yeah, probably a 50 percent chance human, 50 percent chance robot. We have, we have to have robot or not go see that podcast by the way, or go listen to that podcast robot or not. Uh, they demoed some new features on it and I honestly have very mixed feelings when it comes to. Humanoid robots. Yeah. Uh, number one, they're super cool because when we were all kids and someone said robot, what did we all imagine? Humanoid robots, like a person, but not a person. And the, but the truth is humans are a great general purpose device, but if we're building a robot, there's no reason it should be bipedal. Uh, balancing on two legs is difficult. Uh, very few animal species bother to learn that skill and develop the mechanics needed for that skill. It's true. Yeah. It's, it's honestly impractical, uh, coming from like an engineering robotics point of view. But, it's also super cool. Um, I even started a little YouTube series, um, Legged Robots The Balance, um, because it's just a fun topic. Like, from an engineering point of view, they're impractical, but fun, and they're really challenging to build. Well, um, uh, Tesla has released, let's call it V2, Mark 2, whatever, of this robot. My biggest complaint, number one, is it's still the world's worst walking robot. They are not, um, do you remember the old ASIMO robot from the Honda? I believe it was robot. It is the worst walking robot. It looks like a toy robot when it walks. It can only walk very slowly in a very jaded way and whatever. Um, I don't want to make too much fun because it's a very hard project and I would like to see there are better robots like, um, Atlas is a much better bipedal robot. It can walk and run and do very dynamic balancing tasks and all of that. Optimists, not so much, but I do want to give them credit because they're working on the more practical problem now of it has hands. It's a, it's a humanoid robot and those hands need to be able to do something. Otherwise they're kind of useless. And so it looks like that they're investing more time and energy into making those hands do interesting things. So sensors on the hand, uh, perception, generalization of like, it's pretty easy to teach a robot to pick up a pen, but can I pick up a pen and a cup and a two by four. And get me a beer out of the fridge. You know, it needs to be able to do all those things all at once. And it looks like while at the same time, um, I'm a little disappointed in its capabilities for walking, which I think is the cool part, it's neat to see them investing in the more practical side of a humanoid robot, the using its limbs for. Good and awesome. Using its lens to do actual useful tasks. And the AI problem of, um, you know, just build a tower of blocks and things like that. These are the classic problems in AI. Um, they're actually getting to those problems. I like to joke, anyone with enough money can build a bipedal robot, but is it actually useful? It looks like, um, they are starting to focus on the useful. Things on the show. We've talked about the, uh, Amazon little robot that we're all waiting to come out. That's going to be just a little like a Roomba, but with a cute smile on its face, and hopefully that will be kind of a utilitarian robot. It's neat to see them trying to take this bipedal robot, which by its nature is both scary and awesome and make it actually useful to humanity and make it worthwhile to buy and build also it's, it's, it's. It's, it's [00:31:41] James: just fun. I mean, we did a whole podcast about like handy and the Samsung ones that they're doing useful things. I was cool. It was like about, you know, doing stuff with eggs or whatnot. Yeah. It's very fascinating. I'll be, you know, interested to see how it evolves over time and not just robots do in general. Um, I, I didn't think they were going to. Continue on, that's fascinating, but you know, why not? You know, I, you know, there's obviously that it's a next evolution of things coming. Talking about things coming, you know, let's end this thing with a little tiny, um, pour one out moment and maybe a little protection moment. We can pour one out for Hyperloop. [00:32:18] Frank: Oh, pour one out for Hyperloop. There's your update. Yeah. That's cool. I want to say too bad, but there was. So much negative energy around it and a lot of people didn't think it was ever going to work. So, and it's a little bit sad that maybe some funds were diverted to it instead of proper high speed rail for the United States, but you know what? United States rail, it's a hard problem. We have mountains. We're big. Uh, it's really hard to build high speed rail here. Um, and Hyperloop is the first victim. Let's talk about [00:32:51] James: predictions. Uh, the humane, human, humane, humane AI pin. This is like from X Apple. You know, you've seen the thing you put the pin on, it projects things onto your thing. Um, uh, yay or nay, Frank, let's get a yay or nay going on. Oh, darn it. [00:33:06] Frank: This one I'm so mixed on, man. How can I do an up or a down? I really want to do a sideways, but you said yay or nay. Um, I'm going to go with nay. But I think it has potential to become a yay. Uh, I don't think that this V1 device is quite there yet, at least from the demos I've seen, but I really hope the company, I hope people buy it. I hope the company makes some money and I hope that they, um, keep working to improve it. I don't feel any need for the current one. It's going to get an A. But, but it's, it's, it's getting there. It's getting there, James. I, it'll happen, I think. [00:33:43] James: Yeah, I am in the same boat. I think it's a cool concept. I was spending a lot of time last night, like actually researching it and looking at a little bit more. And, uh, I also watched a few interviews with the co founders and stuff. I didn't really like their vibes too much. I know that's kind of a weird thing to say, but now that I buy all products. It's not that I buy all products based on vibes, but, um, yeah, I think that it does have some potential or some ideas there. I do like the idea of disconnecting more, but I'm not sure if it's all that I think it wants to be. All right. Next up, Frank. We gotta be quick. I gotta go. I gotta, we gotta go. All right. Let's talk about threads and matter. Um, this, is it, is it, okay, I just read a review. Which is Philips Hue, just integrated in with Matter and the review was everything in my house just got worse because of this basically. And it lost all these unfeatured, all this other stuff, Matter, Thread. Are we going to see the great unification of IOT devices in 2024? Is it going to happen or is this all just a pipe dream, Frank? [00:34:52] Frank: I will say it's not happening in 2024, so nay on that one, James. Um, this is the classic, yeah, yeah, we have too many standards, so let's invent a new standard. It's definitely facing that problem. But, um, it has potential, um, much like the humane device. Because Enough industry backers are backing it that even if it's a bad technology and you keep pushing it long enough, like USB was a terrible technology in the beginning. It barely worked, but people kept pushing on it, kept investing in it. Same thing for Bluetooth 1 and Bluetooth 2. Terrible technologies in the beginning, but people kept at it, kept pushing at it, and eventually we got Bluetooth 4, a version of it that actually worked and did useful things. Eventually we got USB 2, a useful version of USB. And it's possible for it to come out, but even though there's large industry buy in on them, um, not enough. We're not there yet. Um, it's just gonna, it's just gonna make things worse for a while, to be honest. And so I'm a [00:35:56] James: nay. I'm a hundred percent with you there. I think that, you know, it's hard. You need the old stuff. You need the new stuff. And we, I can't, you can't even get stuff on five gigahertz bands. You know what I mean? Nor do you want them? I don't know. Uh. It's really, it's just a mess. I think it needs much more time to shake out. I also believe that most people don't care, like, especially if they just have a few things, it's probably just going to work. So it's probably okay. It's really for more advanced setups. So again, that. That means like our mass people are going to understand these concepts. I mean, I barely understand even how Zigbee works. I had to watch some videos on it lounge, click a button. It kind of works. Right. So I think we'll get there at some point, like you're saying, where things becomes very synonymous, just like Bluetooth is, but I think we're many, many years off. All right. Last thing, let's end on a high note. What's the future of AI, Frank? What's next? What's next? How are we looking at 2024? Is it another year of AI or is it the next 10 years of AI? Or is it? Oh, [00:36:57] Frank: definitely 10 years, but I think we're still going to see a lot of good stuff this next year in 2024. I think a lot of us old AI people are waiting for the bubble to burst, because the AI bubble always bursts. Any kind of bubble always bursts. And it's really interesting to see how the technology matures after that bubble bursting. But right now, we are still definitely on an up, up slope. Uh, we are climbing up in the usefulness and the abilities. We're in a neat stage where there is a lot of low hanging fruit. Uh, we have a few new models out. We are getting really good at training them. We're getting really good at executing them, especially, um, multi modality. So text, plus images, plus video, plus audio, plus, I don't know, smell. We'll get smell o vision working. And I think that there's still enough low hanging fruit that we can easily have a very exciting 2024 because we haven't cracked video yet. A few companies are just starting out on video. So I think We have a runway for video. We know exactly how to do video. We just need more power. We need more money. We need all those things. And that's a good place to be. We need hardware to catch up to what we're capable of doing in software. And that's always a good place to be. What happens after we have video? That's when the bubble might pop a little bit. Um, but I think 2024, we are on solid footing for a good 2024 in the AI [00:38:23] James: world. I agree. I think even if you were just to say, here's what we have today, I believe that integrations are strengthening and getting better and will be more synonymous. Right? Like, I think if I look at the software I use, there's so many opportunities for it to just improve with different AI features. I've seen like a lot of mine. Start to sprinkle in different features like Descript, which is one I use for transcriptions as an AI feature that will like learn, like as you're correcting things, it'll go through and trends like, Oh, very cool. That makes sense. Right. Co pilot and co pilot chat of just from where they're at a year ago. are now it's a wildly different copilot. Chats didn't even exist. And then as you think about the features that get sprinkled in, Right. They had talked about, you know, um, Hey, we can generate your commit messages for you based on Copilot, right? And you Copilot is like now inside of the browser, right? And everyone else. So there's opportunities, I think, that will really see that even where we're at today, scale out and have deeper integration. And we're talking into writing emails. Just browsing the web. We're talking about, like you said, creating little video clips. I send these newsletters out internally and I use like Microsoft designer and Dolly together where you can generate image and it says a thing and just embeds. It's just like, this is so neat. And as I'm typing, it's like filling in my sentences, which isn't new, but like, it just feels like it's better. Um, and now I can right click inside of like I use edge, but if you highlight text, you can right click and you can say, rewrite with. Get with chat, GPT, and you can have it rewrite text for you and do different things. So those types of integrations I'm excited to see beyond the models and where things are going. What you're saying is I'm also excited from a developer point for productivity stuff that I've talked about, but I think that 2024 will. Continue to see the democratization, if you will, of simplification of letting more developers use this powerful stuff easier, right? You don't have to become a machine learning expert to integrate and sprinkle in some bits and pieces. That's my hope, you know, overall. I feel like we're in the. Beginning phase of that stuff. So I'm really excited to see how that stuff turns out in 2024. That's what has me excited because the DevTools stuff, I'm excited. You know, I, I can grasp that coding it and putting into my app. Like it's too advanced for me. Like I could do it, but I'm not there yet, but I know I will be as there's a bunch of smart people working on that stuff. Right. [00:40:48] Frank: Yeah, you just made me, yeah, I agree with you a hundred percent, um, and, and it's definitely those dev integrations, not development environments, but what we put into our own apps, um, right now we're in the awkward phase of we're just slapping a chat AI into every app. Whatever. It's a good first place to start, but as we, as developers get more comfortable with the technology and figure out how to really leverage it to do interesting features in our apps, we have these very powerful networks and we are catching up with the networks and a little bit, the networks can do more than even we can our imaginations right now. So it's like, we need a few generations, hopefully not generations, but a few years of programmers thinking on it and seeing, playing around. Seeing what kind of interesting things we can put into our own apps, aside from just putting a little chat box in the right hand side of your app. Yeah. And I should also say that, um, in the holiday season, I've gotten to see a lot of relatives and we do live in a little bit of a bubble here. All my developer friends know everything about these like chat GPTs and AIs and how powerful they are and how they can be used, but not everyone knows that right now. So I have plenty of cousins who I was. Explaining for the very first time, what is GPT and what can it do for you? And so I think we have a lot of, of room in 2024 to, uh, get it out there, uh, get more people comfortable with it, get more people using it. Because honestly, I, and I explained this to my cousins, this way, we don't have too many revolutions in computers. Uh, cryptocurrency was supposed to be a revolution, but smart contracts and all that, we all kind of saw the writing on the walls with that one. But this AI stuff, even if we do burst our bubble and we do have another AI winter or something, there is a foundational, fundamental, good technology here that we will use in the future. And I'm excited for that and getting everyone else in the world [00:42:44] James: comfortable with it. And I think that comfortable, getting people comfortable and figure out how. People best interact with these that is a newer model. So for example, like I'm a real fan of these new slash commands that they introduced and to get a co pilot chat, which, and also they have now contextual base, so you can kind of give it a workspace and things like this. It's kind of nice. You can kind of give it a little bit, you can help the AI, but I think it also is helping you, which is like. Here are common things that you want it to do. So you like put the guesswork in like, okay, you want to create a test. You want to document it. You want to, you know, like, okay, cool. Like here's some commands. Like, I like that. Let's have it do some really fun stuff and that'll just continue to improve. I bet where it feels even more integrated than ever. So I'm excited about all that stuff going forward. Well, that's going to do it for our 2023. We'll be back in 2024. We hope that everyone has an amazing holiday and the happiest of New Year's and we'll see you then. So until next time, wait, until next year, I'm James Montemagno. And [00:43:40] Frank: I'm Frank Cruder. Thank you for watching and listening. Peace.