Leo [00:00:23]:
And how's it going, Gary? How's life in Denver these days?

Gary [00:00:27]:
Good. How's life in the Pacific Crest?

Leo [00:00:29]:
We are finally seeing some sunshine. I'm gonna be taking a little bit of a trip here in a few days, which I'll talk about later. But, it looks like the, the roads are gonna be clear, and I'm actually gonna be able to see where I'm going.

Gary [00:00:41]:
Excellent.

Leo [00:00:42]:
So it's funny. I was listening, you know, before we started recording, before we started the show, I've been listening to a couple of podcasts, AI related, and one of them has been doing kind of sort of some of the things that we've done. You may remember, I don't know, months ago, I took my voice and your voice and, former co host Randy's voice and through recordings of each of us at ElevenLabs.

Gary [00:01:13]:
Mhmm.

Leo [00:01:15]:
And then use that to train ElevenLabs so that it could then say whatever we wanted it to say in whomever's voice we wanted it to be. At the time I did that, they suggested that you upload, say, a maximum of, say, ten minutes. Anything more than that was probably gonna be a waste. And it did a pretty respectable job. If I remember right, I'd have to go back and look and see what episode it was. But there are some examples of ElevenLab speaking in my voice and in your voice, and they're actually coming across pretty reasonable. The podcast I was listening to, actually this morning, the AI fix, it's called. Mhmm.

Leo [00:01:51]:
They, they actually did an on air test. One of the host tested the other, in doing essentially exactly the same thing. Eleven Labs has apparently progressed dramatically. And I say that because, not only do will they now take, hey. You know, give us a couple of hours of voice so that we can train our model for your voice. And, you know, if you're not if you don't know that one of these might be AI, a side by side comparison, it's almost impossible to tell the difference. It was spooky. How realistic the, the AI generated voice of one of the co hosts was.

Leo [00:02:41]:
I just thought that was fascinating. You know, we're still using Eleven Labs for our, explicitly synthetic announcer at the beginning and the end of the show. Mhmm. But, it's funny. The the clearly, it's one of those cases where one of the cohosts of that of that podcast does the post episode, editing like Connie does for this for this podcast. And he had admitted he's using Eleven Labs to replace his cohost's voice when there's kind of a trip up. Like, when something gets stepped on or isn't said properly or whatever. I've never noticed.

Leo [00:03:24]:
And it's a great use for that. Anyway, I just thought it was absolutely fascinating that things continue to progress in that, AI. You you just can't trust AI voices anymore.

Gary [00:03:35]:
Well, I I've heard of, yeah, filmmakers or, you know, documentary filmmakers also using it, particularly if, you know, if you have a long interview with somebody and, maybe some of the sound is great, but you've got a really good piece of set like, a really good quote, like, something that they say. Like, oh my god. The the there's a horn honking outside right in the middle. You know, something like that. That what they've been able to do is well, they have a huge sample of the person's voice. Right. Put put it into Elven Labs, regenerate that quote of what they said, as a way to clean it up. Right.

Gary [00:04:12]:
And, and then even at some point, you know, I think documentary filmmakers are are so, careful about things that they even then present that to the person and say, I here's the original, and you can hear it out. Sounds messed up. I generated this using AI. It sounds just like you.

Leo [00:04:27]:
Is that okay?

Gary [00:04:27]:
And then the person Is that okay? And then the person usually, like, well, yeah. I mean, that is what I said. So and it does sound like me. So sure.

Leo [00:04:36]:
Yeah. Yeah. I I think that there's a lot of I'm sure that it runs the gamut from people who are just cavalierly doing it without asking

Gary [00:04:44]:
Mhmm.

Leo [00:04:45]:
To people who are, you know, getting permission beforehand to say, hey. If we need to, can we use synthetics to replicate your voice or whatever? But the the concern that I have, and I've actually I think I I'm trying to remember. I think I've gotten an article or two on it. It's certainly, an issue that I think a lot of people need to be aware of is that, scammers are using this technology as well.

Gary [00:05:09]:
Yeah. Yeah. They are.

Leo [00:05:11]:
Yeah. They're using it to, basically simulate the voice of someone you know. It implies that they've gotten enough of that voice to make a simulation, but they're then doing the simulation in real time and then trying to scam you over the phone.

Gary [00:05:25]:
Yeah. It's definitely a concern. But, you know, for people worried about that kind of thing, I mean, I think we always have to separate very targeted, scamming versus mass scamming. Sure. You know, the idea is you could definitely use something like that if you target somebody. It's like, I wanna simulate the voice of somebody they know and pull this whole scam on that particular person. I'm gonna take the time and effort. I'm gonna set this up.

Gary [00:05:49]:
I'm gonna do all this and target that person. It's but most of the scams, Matt, vast majority of scams are not like that. You know, they're just trying thousands and thousands of people to try to get, like, one person to send them a bunch of iTunes gift cards. You know?

Leo [00:06:03]:
It's interesting though because the barrier to doing this, it used to be really, really hard. Now, if you've got ten minutes of your target's daughter

Gary [00:06:15]:
Yeah.

Leo [00:06:15]:
Speaking That's Target. Or on Instagram or something like that. On social media, it's not that uncommon. It's not everybody, but there definitely are individuals who are spending a lot of time on social media recording video, speaking on video, that the scammers have a lot to choose from. And it's just not that hard anymore to do.

Gary [00:06:34]:
It's still targeted, though. It's still, like, they're testing humans I am doing. It's not, like, I wanna try to I wanna have a computer run this scam a thousand times or 10,000 times for me tonight, wake up tomorrow morning and find out I've got a bunch of, like, money in my Venmo account or something like that. You know?

Leo [00:06:49]:
There's the other side of it too. It's that, you know, if I'm getting a if I'm getting scam calls from my daughter, I know that it's a scam call because there is no daughter.

Gary [00:07:00]:
Yeah. Yeah. No. It's, yeah, it it it's definitely gonna be a bigger problem for people that that could be targeted, but it's not quite to the point where it could be mass rolled out. So

Leo [00:07:14]:
in other podcasting news, I've been experimenting with this very podcast. It I heard some time ago that, the single the the biggest podcast publishing platform is, of all places, YouTube, which when I heard that, surprised me because you don't think of YouTube as being podcast but when you think about how difficult it is to to publish pure audio on a platform, on any of the platforms, what do you do? You turn it into video. So if you're gonna turn it into a video, then you publish it on YouTube. Anyway, it is, and, we are there now too. I've started taking the episodes after, Connie has done her magic and published them to our website and you and, podcast feeds and slapped, you know, our logo on there and and played a little bit with, of all things, DaVinci Resolve. There's a bunch of tools that will take, an audio with spoken word and generate a transcript or captions or whatever you'd like to call it. DaVinci Resolve will do it and burn it into the video when you generate the video, and it's just it's amazing what that software can do, and I've been having a lot of fun with that. CapCut is another one that does something similar.

Leo [00:08:43]:
DaVinci Resolve is, more generic video editing software, but it definitely does, a lot of these neat things. So, anyway, I've been having fun with that. So, anyway, there'll be a link in the show notes for those of you who are YouTube focused. Unfortunately, you can't just drop, a YouTube podcast link into your podcast player of choice. Your podcast player needs to be YouTube Yeah. Or or I think YouTube music. But there is a playlist, at the TEH podcast YouTube feed that is, the the podcast eventually. So, anyway, I thought I'd mention that.

Gary [00:09:18]:
Cool. Yeah. It's it's just probably convenience thing. People just are used to they're at YouTube. Yeah. Maybe this a search thing, they come up with stuff. People have playlists. And a lot of people do listen to music, from YouTube.

Gary [00:09:32]:
So makes sense.

Leo [00:09:33]:
I it's funny. Since I since I have YouTube premium, I also have YouTube music included for that. Yeah. And, I find YouTube music frustrating. And I say that in part because the app for YouTube Music is probably the worst music playing app that I'm aware of. You know, things like Spotify and Amazon Music and so many of the others. I don't play with I with Apple Music, obviously. But, you know, they've got good interfaces.

Leo [00:10:06]:
Searchability, findability is really easy. I just get frustrated with YouTube music all the time. However, everything's there if you know where to look. Right. It's pretty incredible. Everything is somewhere on YouTube. And if it's music on YouTube, it's music on on the YouTube music app as well.

Gary [00:10:22]:
Cool. So here's an interesting story sticking with kind of AI, but maybe the other side of it. Apple, of course, is a little different than the other AI companies is that they're trying to remain focused on privacy. And, there are two ends to privacy when it comes to AI. One is when you're using AI like ChatGPT, how private are your prompts? Like, is that being shared anywhere? How well protected is that? But the other end to it is privacy that involves people that aren't even using AI. It's the privacy of your information out there. What's being used to train the AI models? And it's interesting to note that some of the biggest players in AI, like Meta, like Google, and, like, the the site formerly known as Twitter, all have tons of content generated by people over the years that they could draw on before training their models. So how does Apple do this? Because Apple, first of all, doesn't have its own big, like, social media, you know, and search engine content like that.

Gary [00:11:26]:
And the other thing is is Apple's saying, oh, we're gonna be, you know, privacy based. So if they actually went and said, oh, we're just gonna take all the information from Google or wherever, then that would kinda be going against their own privacy rules. So they have come up with a variety of different techniques to try to train their models. And what's in the news yesterday, kind of a minor thing because the but Apple announced it themselves. They said, here's some things we're doing to better train our models while still keeping privacy. And I'll link to the article because it's a really interesting article. Differential privacy aggregate trends. Anyway, it's using two methods.

Gary [00:12:07]:
One is this, I think is interesting because of this, randomization method that they're using. Basically, they're they're using it for generating gen emoji or gen emoji where you want an emoji character that doesn't exist, so you describe the emoji character and it gives you an emoji character that fits the description. So they're usually just a few words. And Apple is basically needs to train it, but the problem is they can't train it unless they know what people are creating. So what's useful to Apple is what are people wanting from Genmoji? If we knew that, then we could test it and make sure it's creating good Genmoji, and if it's not, we could maybe give it more images of, you know, hearts or flags or faces or whatever. So but it can't do something like, oh, let's just get everybody searches for Genbo you know, when they type a Genboji, let's send that back to Apple servers. So then we can have a list and see what those popular ones are because the idea is, well, privacy. Right? You type something, I want a Genmoji that's a, you know, a a smiling face or something.

Gary [00:13:14]:
That shouldn't be shared with anybody. So AppLabs is randomization thing. So they come up first with a term that they think might be used for generating gen Genmoji, like smile. Right? And they wanna test it out to see how popular the term smile is. So what they do is they send it out to all the iPhones and Macs and iPads. Now you have to have your, like, share data thing turned on for this. You know, help us improve our stuff thing turned on. So all the people that have that turned on, they all their devices get like this word smile.

Gary [00:13:49]:
And basically, your device then goes and says, okay, let me look at the Genmoji that have been created. Has this word been used? And if it has, then report back that yes, smile has been used. Now, of course, the privacy problem is there, then they know, oh, you tried to make a gen emoji with the word smile in it. They're doing the normal things where they they anonymize it. So in other words, it comes back positive saying, yes, you did search for the word smile, but I'm gonna immediately forget who you are, what IP address this came from, whatever. I'm just going to add a one to the smile list and that's it.

Leo [00:14:28]:
Smile counter.

Gary [00:14:29]:
But yes. Smile counter. And then they could see how popular word is. But Apple goes one step further than this because that that all that so far, that's not very fascinating. That's easy to do. Right? Here's the fascinating part. Apple randomizes this by actually falsifying the information a bit. So I read it, and I I think a good way to explain this technique is imagine if, the answer is yes, you use the word smile.

Gary [00:14:57]:
Instead of sending back a yes, a positive, it's going to 90% of the time send back a yes, and 10% of the time say no. The word smile was not used. It's going to lie some of the time, which means that if the data is ever intercepted, the yes or no means nothing because it could be lying, could be false. Mhmm. So imagine let's go further because 90% is a lot. Right? Let's say it's 60%. Let's say the word smile is used 60% of the time it's gonna send back a true, and 40% of the time it's gonna send back a false. If the word smile is not used, 60% of the time it sends back a false, and 40% of the time it sends back a true.

Leo [00:15:47]:
Alright.

Gary [00:15:48]:
If the information that positive, is intercepted between you and Apple, it's not very useful. It's showing that there's a tiny bit more of a chance that you did actually use the word smile. But if you wanna know, it's useless. It's like Right. That could almost just as likely be that. But if you think the flood of data of thousands, a hundred thousand of these coming in to Apple, The 60% means that there's going to be a bulge in the data. The data is going to be much more positive for the word smile than, say, for the word frown. Right? Because not as many people are using frown, so it's gonna be, you know, the other 60% for false and 40% for true.

Gary [00:16:34]:
So they can tell, looking at all that data, that, yes, the word smile is one of the most popular words used.

Leo [00:16:39]:
Right.

Gary [00:16:40]:
Even though the data they're getting is very you know, it's kinda noisy is what they're calling it. Very noisy. You know, and this kinda makes sense be if you think about how statistics work and and just pulling. Like, if you pull one person, that data is really useless. You have to pull like 300 people or a thousand people to actually get good data. And it's the same thing here. You're making the data really noisy. You're making that one bit of data coming back useless.

Gary [00:17:05]:
And even if you intercepted that piece of data, it's still pretty useless because it may just be wrong. But enough people, thousands and thousands of responses, and you see a pattern in the data, and you could see which words are really used. So Apple can actually get this even though your iPhone or whatever is lying to Apple almost half the time.

Leo [00:17:25]:
Most time.

Gary [00:17:26]:
It's so it's fascinating to me. So an interesting way to not to anonymize the data and also make the data so noisy that a single interception is useless.

Leo [00:17:36]:
It's interesting because I think the the the term aggregate analysis Yeah. I I've used this in you know, for years saying, you know what? You you are not very interesting. Yeah. 10,000 people like you. Yeah. Yeah. That's interesting. But I think that is a really, really difficult concept for people to grasp Sure.

Leo [00:18:05]:
When they're concerned about my privacy. Right?

Gary [00:18:09]:
Right. Exactly. I think I think it's pretty interesting. I mean, they're using two methods here. One of which should be good as long as you trust Apple to throw away your identifier, like your IP once it gets to them. But the other one even works even if you don't trust Apple to throw that away. Of course, if you're at that point and you say, I don't trust Apple to actually implement what they said, then, you know, there's all of that. But at least they Apple can hold this up and, you know, perhaps it could even be tested in the field in in many different ways.

Gary [00:18:38]:
Now there's a second way that it's done, for more complex things like actual prompts. So prompt doing a an AI prompt or just asking Siri something or even in this case, the example they gave was an automatic response. Somebody texts you and you already get, like, this thing where you can respond with one tap, you know, if it's, you know, obvious, what, like, it might be. Apple goes and looks at the real prompts, the real things people type in response, and wants to mimic those. They kind of can make a guess, and they already are making a guess that somebody would phrase a response in this way, but they wanna get better at it. But how to anonymize that? Because if they actually take a real response from a person, send that back to Apple, they've got that real response. So instead, they're doing something similar where they're saying, okay. We're going to generate three responses, based on, like, the question, what do you wanna do for dinner tonight or something like that.

Gary [00:19:42]:
And, we're gonna generate three possible responses, send those to different devices, and it's not going to look at the specific words like the nouns. Like imagine saying I think the example they gave is, would you like to play tennis tomorrow? You know, the word tennis isn't important. It's soccer, it's football, it's whatever it is. It's the rest of the sentence, the structure saying, you know you know, would you like to play something, some sort of sport tomorrow? And they send three variations. And if you happen to use one of those three variations, in in in a different way, would you like to play backgammon tomorrow, it will go and get that as a positive that, oh, one of our three was actually used and do the same an automated anonymous noisy thing sending it back. And then it does that enough times, it can go and say, okay, out of these three ways to ask that question, way number two was used more often. That's the more popular one that people actually type. And so that's kind of interesting.

Gary [00:20:49]:
It's using AI to train AI in some ways because the AI is used to generate the suggestions. But I think then the idea could be that if number two is suggested, then at some point variations of number two can be suggested. And then some point after that, whatever the winner of that is, variations of that could be suggested. And, you know, you could get better and better, and that the real data of what humans actually type to ask that type kind of question will eventually kind of make its way as the real suggestion. And not that's something that is awkwardly phrased because an AI created it.

Leo [00:21:25]:
Fascinating because you there's so much you can do when you have a a large enough sample size.

Gary [00:21:30]:
Yeah. Well, yeah, that's exactly what you know, I'm

Leo [00:21:32]:
using them all. Their advantage. Yep.

Gary [00:21:34]:
Yeah. And then they just set up processes to run these so they don't have to come up with all of these. They can just have basically the AI constantly, you know, making itself a little bit more accurate in how it phrases things. Also, they could come back with, like, here are, you know, some Genmoji based on what people popularly put in for Genmoji. And then, you know, they can even have AI go and say, which one of these are kinda awkward? You know? And then put a list of, like, oh, would you ask for, like, a frowny face with a thumbs down or something? These don't look great.

Leo [00:22:11]:
Right.

Gary [00:22:11]:
You know? So let's go and play with the model and maybe make that look better, and then we'll get better results instead of trying to make all sorts of things look better that already look good, that, you know, already producing good results. So, anyway, some cool stuff and just showing, you know, how Apple is trying to overcome its problem of being private and trying to get an AI better and better.

Leo [00:22:38]:
I love that they're doing this. And you know me, I trust that they're doing this. But I think something you said earlier is that, you know, if you don't trust them to do that, you don't trust them to do this, if you just don't trust them, none of this helps. Right? Right. I it's it's the reason it's frustrating for me is that, I have a lot of feedback from a class of reader who absolutely refuse to use, I'll just say OneDrive, Microsoft specific.

Gary [00:23:12]:
Yep.

Leo [00:23:12]:
Because they are absolutely convinced that Microsoft is pushing OneDrive as hard as it is because they want to analyze your documents. They want to use your information to train their AI. And how do you prove a negative? Right? How do you prove that they're not doing that? Well, you can't. Yeah. And it's it's it's frustrating. Mostly because I don't believe they are, and they claim they're not, of course. But even if they did things such as, you know, what Apple is doing, we just described Apple is doing, nobody believed them anyway.

Gary [00:23:54]:
Yeah. I think people perceive the, the value of doing something like that is far greater than it is compared to the risk to Microsoft or Apple or whoever. Of course. I mean, the risk of of getting caught

Leo [00:24:07]:
Yes.

Gary [00:24:07]:
And finding out that, yes. Oh, all of your Microsoft OneDrive documents have been, you know, used to build this AI model. That's a huge risk, and they have other ways of getting, data to train Yeah. AI, including just having people write stuff. I mean, there are people that you know? I know for coding, there are, like, people that are paid to actually just code and solve coding problems, so that AI can use that to train itself to write better code. And That's good. Yeah. The same thing for writing, I assume.

Gary [00:24:40]:
Plus, there's tons of of public domain works and tons of probably, libraries and examples of writing that they can get access to. Yeah. You know, legitimately with, you know, licenses paid for, signed, agreements made, all that kind

Leo [00:24:55]:
of

Gary [00:24:55]:
thing. It's just it's just be weird if Microsoft will be like, oh, we can make this process 3% better by risking everything. You know? It's like, why?

Leo [00:25:07]:
Right. Change the topic.

Gary [00:25:11]:
Yeah.

Leo [00:25:11]:
I learned a new word this week. Yeah. Quishing. The we've heard, of course, of phishing and smishing and a few other ishings. Quishing is, an attempt to mislead using QR codes. And as it turns out, misleading QR you using QR codes to mislead is really, really easy to do.

Gary [00:25:38]:
Mhmm. Sure.

Leo [00:25:39]:
The scenario that is used as an example is very often these days, you may find a QR code on a parking meter as the gateway to pay for your parking.

Gary [00:25:52]:
Oh, yes.

Leo [00:25:53]:
Good one. And, of course, you know, nine times out of ten, ninety nine times out of a hundred, that's exactly the right thing. It goes to the payment processor, does all the things. But it could be maliciously overwritten with a sticker

Gary [00:26:09]:
that

Leo [00:26:09]:
has a QR code on it that then leads you to a website that looks like the parking place or looks convincingly enough like a parking place to actually take your money. So not only do you not do they you know, do you not give your money to the right person, but all of a sudden, you your car gets ticketed because you didn't pay your parking even though you thought you did. But that's just an example. There's a lot of places where QR codes are in the wild, in public, and in a way that is easy to, like I said, put a sticker open them. I think what a lot of people don't understand is that there's really nothing special about a QR code. A QR code is just text. Yeah. That it happens to 99 times out of a hundred contain a URL that your QR code scanning app knows to do something with, like, knows to hand it off to your browser to go somewhere.

Leo [00:27:05]:
That's more of a coincidence and a convenience than it is a design. I've got an article, what is a QR code, where I've got the Gettysburg Address, the whole thing in a single QR code. It's like I said, it's just text. The problem is that you and I can't read a QR code. Not we don't know where it's going just by looking at it.

Gary [00:27:30]:
Yeah.

Leo [00:27:30]:
And as it turns out, not all QR code scanning apps are great at previewing what it is you're about to go to. And a lot of people that use QR codes, I'll say, misuse them by using URL shorteners like bitly or others so that not only do you not know where the QR code is going, even if you can see the URL that the QR code represents, you still don't know where it's going because it's going through a URL shortener. So anyway, I just wanted to bring up a new word, quishing, and have people be a little bit more skeptical when it comes to scanning random QR codes. One of the things that I did do for myself is I confirmed that my on device security software, which in my case is Bitdefender on my Android phone, it actually is monitoring not QR codes and not necessarily even URL shorteners, but it is monitoring the URLs that my browsers are going to go to. So it is watching for known malicious URLs before they actually get processed. But that's about the best we can do right now. QR codes are great. I use them.

Leo [00:28:49]:
I love them. They're wonderful. But, they do require just a little bit of extra attention and caution in the wild. Yeah. I just like the word, questioning.

Gary [00:29:01]:
It it is, and it's interesting, scam there. Of course, one of the things about it is that it can happen just as easily with a URL, I mean, because it's the same thing. So you could have somebody could somebody put a sticker over a parking meter that has a QR code that goes to their site, and underneath it has, you know, some URL like paypark75dot, you know, whatever. And that URL there, if you were to type it in, it doesn't protect you at all. The QR code and the URL go to the same place.

Leo [00:29:32]:
My impression.

Gary [00:29:33]:
They've replaced both. I have a hard time, I have a lot of people because I have videos on how to do QR codes, how to make them. Because there is there's a way in shortcuts on your Mac to make it. And the thing the problem I've got is I hate when I see QR codes that are obviously generated by a free QR code generator that puts a little image in the middle. Like, I think Google's does it. Puts a little image so you can identify the QR code as being created by Google. Okay. And it's just a little you know, it doesn't hurt anything.

Leo [00:30:01]:
Okay.

Gary [00:30:01]:
The URL is still just the URL. They're just putting a little picture there. But I'm like, you don't need to go to a free, you know, QR code generator that is then going to put a little image or promote itself in some way. You can go to Apple has a little thing in the shortcuts app where you can create a QR code and it's completely generic. There's nothing in the QR code saying Apple or that it was generated this way it's just the text so I show people how to use that but then I get lots of comments for people that don't understand that all it is is a representation of text and they wanna know how do I get it to be, number one, dynamic. Like, I wanna create a QR code, but I wanted to point to a different URL later on. Like, I want it you know, if I change my website or whatever and I have to explain to them, it's just like URL. It doesn't work like that.

Gary [00:30:49]:
Like you have to create the web page at a URL and then this is a representation of representation of that text. The other thing I get is people without any kind of website they wanted to represent something else. Can I make a QR code so that when somebody scans it it downloads this PDF? Like well yes but you have to have a web server. You have to put the PDF there and then you have to have the QR code be the URL to that PDF. It can't just be a PDF, like, the QR code can't somehow encode the PDF unless it's a massive QR code. Technically, I guess, it could. But but it is interesting how people don't people don't understand. It is just it's like the easiest way to explain it to another language.

Gary [00:31:32]:
Like, if you were to write a sentence in English Mhmm. Then write the sentence in French. Right.

Leo [00:31:36]:
It's the

Gary [00:31:36]:
same thing as writing a sentence in English and then the sentence is a a QR code. It's different language that nobody could speak. It has to be decoded by the computer. Although, you can actually decode it manually. If you really wanna go to the trouble, if you wanna if you wanna a fun way or maybe not so fun way, just spend a Saturday afternoon. You can look up how to do it, and there are websites that explain it. And then, you know, it's basically kinda like doing a Sudoku, but even less fun.

Leo [00:32:03]:
It's funny. It's it amazes me how much redundancy is in Oh, yeah. QR codes. It's you know, you can, like, lose depending on decisions made when the QR code is created. You can let use lose, like, half the QR code and it

Gary [00:32:16]:
still works.

Leo [00:32:18]:
I wanted to touch bases real touch back real quickly on dynamic

Gary [00:32:22]:
Yeah. Destinations

Leo [00:32:24]:
because, it's it's totally possible.

Gary [00:32:28]:
Oh, yeah. With a redirect.

Leo [00:32:30]:
Exactly. It's not the QR code that is dynamic. It's the destination of the QR code that is itself dynamic. Now, this is ever so slightly frustrating for me because some browsers actually will let you create a QR code of the current whatever's in the current address bar.

Gary [00:32:46]:
Yeah.

Leo [00:32:46]:
Right. So you go to a web page, you do something. I think Edge has it built in. It says, make me a QR code of this, which is great, except, depending on how you got there, it's way long. It's got tons of tracking information encoded in it permanently. That by definition, that becomes invalid because it's not tracking what it was tracking originally anymore. So it still requires this step of saying, you know, make sure you are encoding what you want to be encoded. Redirection comes into that because, for example, the nonprofit that I work with, we have, you know, organization.org/donate.

Gary [00:33:24]:
Mhmm.

Leo [00:33:24]:
That's not the page. The page is something much longer that has a lot of other information in the URL. So if you actually go to the donate page, that's not the page you wanna create a QR code of. You actually actually have to take a step back and create the QR code of this shorter version. But, yes, reader or dynamic is is very, very possible. It's just not the QR code like you said.

Gary [00:33:49]:
Yep. A small note here because I've I've mentioned the TV show Mythic Quest a few times on this podcast. It's a Apple TV plus series. It's had four seasons, and it's about a video game company. It's a sitcom about a video game company, and it's it it's thought of it as pretty good. I loved it and all. It's canceled. And, you know, it's four seasons, and it's canceled.

Gary [00:34:15]:
No big deal. What's interesting is it was canceled after the fourth season just ran, and there's a bit of a cliffhanger at the end of the, you know, the end of the last episode of the four season. And the creators didn't wanna leave it like that. So with permission from Apple, they redid the ending, and that's gonna be released later this week. There'll be a revised edition of the last episode. I highly suspect it is not like they took the last five minutes away and replaced it with something else. I'm I will bet anything, it is going to be the same episode, and then they'll tack on two minutes. You know, do something where suddenly it jumps forward in time ten years.

Leo [00:34:57]:
Oh, right. You

Gary [00:34:58]:
just see, you know, the what what how that cliffhanger kind of resolved itself and where the characters something like that. I don't think they would actually change any. I think they're going to add a few minutes to it. This is the same show that's done a bunch of different things, like, when the after the first season when the pandemic hit, they jumped on it really early and did an episode where everybody was isolated. So they sent iPhones to all the different characters, had the writers write a script. Each actor got in their home an iPhone and a script.

Leo [00:35:30]:
Remember you're talking about that. Yes.

Gary [00:35:32]:
Yeah. So they and they put it together, and, of course, they made it like it was real life. Like, they were all, you know, at stay at you work from home because of the pandemic, and they did a whole episode like that. And, so they they like to jump on things, and they like to do kind of different kind of different experiments with it. They eat also, this is the same show that, there's one episode each season that has none of the main characters at all. Like in the I think it's the first season. Maybe wrong or maybe the second. But since it's a game development company in the first season, I think there's an episode that goes back twenty years to the two the couple that founded the game development company in the early days of computer games.

Gary [00:36:14]:
Just them working from their home, coming up with an idea for a game and all this. And those characters are long gone as, like, is so often in real life. They have nothing to do with the current company. But you learn about, like, the how this company was founded at the very beginning. So it's a completely different style, completely different, you know, actors, everything is set up. You you watch the first couple of minutes thinking, is this the right show? It's like a completely different thing. And they have one episode every season that does that. So they they really have done lots of interesting things.

Gary [00:36:45]:
I'm not surprised they also wanna do this. Let's redo the end of the of the last episode and put that on there. It's a shame that it's, it's ending, but I think it was it's a quality show. It had a target audience of developers and gamers that it hit perfectly, and I think they got a big audience of exactly those people. And the hope would be, like a lot of shows, that it grows beyond that, right, and gets a bigger mainstream audience. And it didn't quite do that. So everybody that watched it and wanted to watch it was really happy with the show, and everybody else didn't ever hear that it existed.

Leo [00:37:23]:
My my hope is that this final episode I've not seen it, by the way. Yeah. My hope is that the final episode is not the cheesy thing where, okay, they come to the cliffhanger, and then they just scroll text that disrupts, you know.

Gary [00:37:36]:
No. No. I'm I'm, no. Based on that

Leo [00:37:38]:
this and so on on

Gary [00:37:40]:
No. Based on that, the deal big deal they're making about this and the people involved with the show Right. No. They they they definitely shot something. Even even if it it's they just spent their own money and just went and got everybody together for a few hours and did something with the set and shot something. They they definitely put something. So it'll be interesting to see.

Leo [00:38:02]:
So last week, I mentioned that I had sold my Tesla and gotten myself a Rivian. Yeah. I've been having fun with it. Like the Tesla before it, it is a rolling computer or rather, several computers rolling, multiple multiple CPUs in there for sure. One of the things we'll be doing here in the not too distant future is going on a bit of a trip. I pull a trailer, and one of the big concerns about switching to an all electric vehicle for pulling, and not not it's an Airstream. It's a not a small trailer, is range.

Gary [00:38:46]:
Mhmm.

Leo [00:38:46]:
What's the impact on range? On the the gas vehicle that we've been using for the past fifteen years, mileage gets cut in half easily. It's easy. Right? You know, you drive around in in my Toyota Sequoia, and it gets 16 to 17 miles to the gallon. And if you're pulling the trailer, it's like eight, which is a big cut. As it turns out, in theory, the, the Riven will do the same. I have a, maximum battery pack, which supposedly has a range of around 400 miles, which is actually pretty nifty. Yeah. And when you hook up the trailer, this is where the tech comes in that just fascinates me.

Leo [00:39:29]:
Because I did this for the first time this afternoon. I hooked up the trailer, and I drove it around the block Mhmm. Just to make sure that, you know, all the i's are dotted and t's are crossed, and we've got all the pieces we need. One of the things it does is, a, it notices you've hooked up a trailer. Mhmm. And after a few minutes, it tells you how much the trailer weighs, which, again, I get it. Mass, if if, you know, the the impact of the trailer on simple things like acceleration. Right? Yeah.

Leo [00:40:03]:
You have energy going into the motor compared to the, the speed that comes out of it kind of gives you enough data to make a rough guesstimate as to the mass of what it is you're pulling. And then it does a real time recalculation of your range. And sure enough, my 400 mile range got turned into a 98 mile range, which and the the other thing that'll be interesting, that was just based on a trip around the block. As I understand it, that information is going to get updated in more or less real time as we're traveling.

Gary [00:40:39]:
Yeah.

Leo [00:40:39]:
So as as, you know, my driving style when pulling a trailer impacts all of this, how fast you're going, what's the terrain, all that kind of stuff. So it'll be really interesting. I just thought it was absolutely fascinating that not only did it recognize it was a trailer, it did the math to figure out how heavy it was, and it did the math to infer the impact on the overall range. We'll see. I mean, talk to me in a couple weeks. We'll see how the trip went.

Gary [00:41:06]:
Yeah. No. I I I envision you, you know, you're pulling a trailer. You you have to find, somebody who's got, a Cybertruck and have them pull a trailer, and you guys kinda race for bink slips or something like that. So

Leo [00:41:20]:
yeah. No. No. As I understand it, if the Cybertruck tries to pull the trailer, the bumper falls off or something.

Gary [00:41:28]:
Well, there you go. Then you would win. I you know, it's funny. I was actually aware of Rivian pretty early because, there was a TV show. Do you you ever watch any of the shows with, you know, the actors, Ewan McGregor and Charlie Borman? They did a bunch of shows. The first one where they drove motorcycles across Asia, and then they drove did it across Africa.

Leo [00:41:50]:
But we never watched any

Gary [00:41:51]:
of that. The last one they did two, and then there was a big gap. And then they did one where they drove from, the tip of South America up the Pan American Highway. Yep. And on that one, since it was a few years ago, they decided they wanted to do all electric. Cool. So they got two electric motorcycles, but then they needed support vehicles, and they needed an electric truck, which didn't exist yet.

Leo [00:42:18]:
Ah.

Gary [00:42:19]:
And but Rivian was starting. It was a start up. Right. They they gave them two of their prototypes and I think some engineers to go along with it. So the whole trip, they're on their bikes. The support vehicles are these two Rivian, I forget if they were pickups or, you know, with the they had, you know, the SUV kind of designed to them or whatever. But they went and they, they went. And there was actually times when they, I think, they charged the motorcycles from the battery on the Rivian.

Leo [00:42:51]:
I believe it. Yep.

Gary [00:42:52]:
Oh, it's interesting. And that was, yeah, that was before you could buy one, before they actually had them out. These were, like, prototype vehicles that they had airlifted down to South America, and they were they were getting data and trying to, you know, make them better and stuff. And it's interesting. You could just buy them now.

Leo [00:43:08]:
These are full sized. Mine is an SUV, the r one s. It is the full size SUV. It's the equivalent to my Toyota Sequoia, which is a full sized gas powered SUV. The r one t is their pickup truck. But you're probably seeing more Rivians than you think because they also make, the delivery trucks that Amazon is using. Oh, yes. There's lots of ribbons at least out here delivering packages and pulling up to the driveway.

Leo [00:43:37]:
And I'm wondering if what they were using in South America as a support vehicle might not have been one of those.

Gary [00:43:43]:
May yeah. Definitely was low pro, you know, normal vehicle profile, not the tall things.

Leo [00:43:48]:
Okay.

Gary [00:43:48]:
Yeah. So so anyway, cool. I I I want one, but I have absolutely no use for a vehicle like that.

Leo [00:43:55]:
I understand. Yeah. No. I I get it. I'm and like I said, I'm downsizing. I've always want I mean, I've been I was thrilled with my my Tesla for I was thrilled with my Tesla for eight years. I loved all electric driving. It just made sense to replace it.

Leo [00:44:10]:
What I'm really looking forward to now is the ability to replace those both cars, the Tesla, which I used for day to day driving, and the Sequoia, which was used for heavy lifting and pulling the truck, replacing them both with a single vehicle and going for cool.

Gary [00:44:26]:
Cool. Alright.

Leo [00:44:28]:
Okeydokey. What's cool? What's cool, dude?

Gary [00:44:32]:
Well, so, Doctor Who is back. New season of Doctor Who started, on in America on Disney plus, everywhere else in the world on BBC. And they start right off with a very topical episode on AI. It's kind of interesting, because the story, it's not spoiling the story at all to say that, you know, some other planets, some distant place and time, there is a kind of an AI based civilization that somehow in its large language model gets some data from Earth, which messes things up a little bit because it doesn't understand exactly what that data represents. So yeah. Well, basically, it's one of those like you you buy somebody a star, you know, you you pay money, spend money to something. Right? And then you get a certificate that says that the star is now named after them. And somehow that ends up in the AI, language model on this planet, which and it it and it's its own star.

Gary [00:45:37]:
So it decides that that person is the queen, must be the queen. And the the planet is named after, like, named after the person and all that stuff. And there's a whole lot of different back and forth about AI, in the episode. So a little commentary going on Mhmm. About AI, and a lot of others regular Doctor Who stuff too. It's not all just about that, but

Leo [00:46:02]:
it

Gary [00:46:02]:
kinda has a little bit of a AI theme running through it. So and then we've got, I guess, a season of a bunch more episodes coming out. Also, Black Mirror has a new season too, but I have not checked that out. It's interesting that doctor and Black Mirror are basically debuting almost at the same time.

Leo [00:46:19]:
It's funny. So, we did watch the Doctor Who episode. We're we're right on on the Okay. That one, and we'll be following it as they come out. Black Mirror, my wife mentioned it to me just yesterday that, you know, we keep hearing about it. Maybe we should give it another try. I think we watched, like, one episode

Gary [00:46:36]:
Oh, okay.

Leo [00:46:36]:
From the first season. Yeah. Got so depressed that it was like, no. Maybe not. But

Gary [00:46:43]:
yeah. They throw in a they throw in a happy episode once a season just to keep you going, keep you from jumping off a cliff. Right. You know? So yeah. But yeah. Obviously, I yeah. I'm gonna I've watched all the Black Mirrors, and some of them do give me nightmares, nightmares at night. And sometimes when I think of them during the day, they give me nightmares.

Leo [00:47:06]:
You see, that's not that's not a great recommendation for where I'm at right now.

Gary [00:47:10]:
Yeah. Yeah. I gotcha. I gotcha.

Leo [00:47:12]:
Cool. Have you been watching Jack Reacher?

Gary [00:47:16]:
No. Okay. Do you

Leo [00:47:18]:
you know Jack Reacher from the Lee Child? Okay. Mhmm. It's a season. I think he just wrapped up his third season of Jack Reacher. And, we were talking about him, Jack Reacher, the actor that plays him, Richardson. Richardson. Keep saying his name wrong. And we decided, okay.

Leo [00:47:40]:
Yeah. What else has he been up to? He's been a busy, busy guy. He's done a lot of different things. And in doing that research, we stumbled across a movie that sounded really interesting called The Ministry of Ungentlemanly Warfare, which we had never heard of. It just came out last year.

Gary [00:48:00]:
Yeah. Well, it came out. I don't even know if it came out that long ago.

Leo [00:48:04]:
And like I said, well, it's dated 2024 in IMDb, for whatever that's worth. But, completely under the radar, it's got several familiar faces. The the way that I describe it, it's a World War two story Yeah. Where, you know, this team has to go and basically blow up or get access to something that would then, allow Americans to make it safely across the Atlantic more often and and join the war effort. It had kind of an ocean's 11 vibe to it Mhmm. Where it was almost a heist caper, where you had all these different people bringing their different skills to the table and doing things all at the right time in the right way under the under the noses of the Germans in these very anyway, it was just fun. We enjoyed it, had a good time with it. Like I said, I think the the, the the amount of time between learning that the movie existed and finishing the movie and thoroughly enjoying it was something on the order of about four hours.

Leo [00:49:09]:
So, anyway, Ministry of Ungentlemanly Warfare. It was

Gary [00:49:14]:
Yeah. I saw I saw two, and, I was surprised because I just was brought in by the premise. I didn't know anything about the actor. I was like, premise seems like it might be fun. Let me start watching it. And I was instantly brought into the movie. And then after watching the movie, I was like, wait for a movie I've never heard of that for all I know went straight to, like, being released, you know, streaming. The the, quality was really high.

Gary [00:49:38]:
And then, of course, the movie ends, and it's a Guy Ritchie film.

Leo [00:49:42]:
Yes. I didn't Yeah. I didn't realize that until the end either.

Gary [00:49:45]:
Yeah. I'm like, oh, okay. Well, that explains a lot of what I saw in this movie. It's a Guy Ritchie film. And it's funny to actually watch a Guy Ritchie film, you know, and he for those that know, he did lots of act British action movies and Sherlock Holmes movies and all that. And, he, it it's interesting to watch it without knowing it was Guy Ritchie, without being, okay. I'm gonna watch a Guy Ritchie film script. A lot of action, a lot of cool camera shots, a lot of quick paced dialogue.

Gary [00:50:12]:
And I didn't know that. There was a lot of action, a lot of cool camera shots, a lot of quick paced dialogue. I said, And then I was like, oh, that explains it.

Leo [00:50:19]:
The other thing I forgot until the end of the movie was that it started out by saying based on actual events. Yes. Because at the end, they show you the photographs of the real individuals on whom some of these roles were based and what happened to them. Yep. Which kind of my suggestion earlier about your your final episode of Mythic Quest. Right? There was kind of that. Here's a picture. Here's what they went on to do.

Leo [00:50:45]:
Here's a picture. Here's what they went on to do.

Gary [00:50:47]:
And I think if you, also, there's a tiny similarity between that movie and the Quentin Tarantino movie Inglourious Basterds, which is completely fictional. But I remember when that movie came out, he said he was inspired by some individuals that were real. And I believe some of those individuals are the people in this movie. Oh. So this movie is, like, based on a true story. Inglourious Basterds is not based on any kind of true story, but inspired by people that existed. So yeah. So that's, that's kind of why they feel similar.

Leo [00:51:23]:
Yep. Anyway, it was a good time.

Gary [00:51:25]:
Mhmm.

Leo [00:51:26]:
So where are you gonna point people this week, Gary?

Gary [00:51:28]:
Yeah. I'll point to a video that's, out when this podcast comes out called 10 reasons your Mac may be acting strange and how to fix it. Just decided to do a compendium of, like, the things people ask me, like, something's something's wrong, something's strange, and it's like, oh, you've turned this setting on. Turn that off. I compiled enough for them to make a whole, video of 10 of those, for people to watch. And it's kinda thing that it may help you out, and it may just be good to watch and kind of, like, keep in the back of your head that, you know, when you see it, you're like, oh, wait. This is one of those things. There's a way to fix this.

Leo [00:52:06]:
I'm sure I mean, you've got that that gut feeling too. You end up with a question, and it's like, oh, yeah. That's this really obscure thing over here.

Gary [00:52:14]:
Yeah. Yeah. Exactly.

Leo [00:52:17]:
The one I'm gonna point folks at is, I use BitLocker. Are my backups encrypted? There's a lot of confusion about how whole disk encryption works and what its impact on things like backups are. And in some cases, the wrong assumption can be kind of dangerous. So I walk through exactly what's encrypted when you're using something like BitLocker. And then when you back that disk up, when is that backup encrypted? When is it not encrypted? Spoiler, most of the time, it's not, but there are some scenarios where it can be encrypted by default. So it's something to, to be aware of. I use BitLocker. Are my BitLocker or Are my backups encrypted? Ask leo dot com slash one seven nine five zero three.

Leo [00:53:03]:
Okeydoke. I think that wraps us up for another week.

Gary [00:53:08]:
Yep. Sounds

Leo [00:53:10]:
good. As always, thank you everyone for listening. And, yes, hopefully, you may even find us out on YouTube. No. We're not publishing the actual video of us talking to each other. It is one of those things with the yeah. We may do that sometime. We we did that for an anniversary episode years ago, if you remember.

Gary [00:53:30]:
Yeah. There's a bunch of us that were like Bunch

Leo [00:53:32]:
of yeah. It's one of those things where okay. You know? K. Fine. I have to make myself look a little bit better before I'm willing to go on camera that way. And I have to turn on my lights and all that kind of stuff.

Gary [00:53:43]:
Yeah. Yeah. Exactly. Anyway, we could put a background, all that.

Leo [00:53:46]:
All that stuff. Alright. Well, there we go. Thanks again for listening, and we'll see you again real soon. Bye bye.

Gary [00:53:52]:
Bye.