Leo [00:00:22]:
So Gary, I hear there's some new Apple stuff. What's going on with Apple?

Gary [00:00:27]:
Well, you know, it's weird because we're recording this right in the middle of it because Apple sometimes does these things where it's like a week of announcements and we're recording this on Tuesday afternoon and Apple is expected to announce new things Monday, Tuesday, Wednesday. So we're 2 out of 3 so far. Uh, but we have some pretty good ideas of what's coming on Wednesday, right? The rumors point firmly in one direction. So, so far what we have, is it's a— we have a new iPhone, the iPhone 17e. So it's— this is their economy model of the iPhone. They had a 16e, but, uh, they came out with the 17 models, so the 16e was the current economy model, and now they've come out with a new, uh, version to fit the current models out there. So if you want an iPhone but you're not like a tech person, you don't need the fastest processor and the biggest screen and like all the latest specs, you just want something basic. The 17e is now the iPhone model to get, and no doubt will be the one that carriers are pushing.

Gary [00:01:32]:
You know, usually this is the model that all of the carriers, when they want to put an ad up saying you can get an iPhone for $99 if you sign up for a year, like, this is the one that you get. So they came out with that. They came out with new And I think this is actually a significant thing. They— Apple makes two displays. It wasn't that long ago that Apple didn't make any displays. Like, they made displays a long time ago, lots of them, and then they got out of the business. And then they got back into it with a super high-end Pro Display many years ago, and that was it. It was like, if you wanted a normal thing that a normal person would get, or just a creative individual or whatever, you would have to go to a third-party screen.

Gary [00:02:18]:
For your Mac Mini, Mac Studio, or to hook up to your MacBook, you know, as like an extra display. But then Apple came out with this really high-end one that was really good for like, you know, if you were like, you were getting paid by somebody, like you were a video editor and you had a big corporate company behind you.

Leo [00:02:35]:
In other words, if you could expense it. Yes.

Gary [00:02:36]:
Yeah. Yeah. You could expense a $5,000 monitor. Then they, but then people were like, well, we want something more affordable. So they came out with an Apple Studio Display, which is what I'm looking at right now. And it's, you know, probably the most expensive 5K screen of this caliber, but it is a 5K screen. So, you know, and it's nice and all that. What Apple announced this week was basically updates to both of those new versions of both of those with better specs, better stuff.

Gary [00:03:08]:
The Pro Display now has a camera. It originally didn't. It was like, hey, if you're like a pro video editor, why do you need a little webcam? You don't. You're working on editing video. But now they've got that and a newer version of the one I'm using. So they've updated it, which I think is significant because it shows that Apple is like, yeah, we're making these displays for the foreseeable future. These weren't experiments or things we threw out there. Just, you know, this is like we are in this business of having a high-end consumer display and a high-end Pro Display.

Gary [00:03:38]:
To go with our Mac Minis and Mac Studios and Mac Pros, I guess, too. The— so yeah, it's nice. I won't be upgrading mine because it's, you know, the, the, the new stuff isn't that much better than what I have now. But if I didn't have one and I was in the market, then I would definitely get it.

Leo [00:04:03]:
So, so it's interesting that monitors do seem to last a long time. Yeah, longer, longer than the devices, um, they're— they originally are attached to. Um, that Apple, that old Apple screen you were talking about, I've still got it down in my basement. Yeah, it's still connected up to my old, um, um, Mac Pro, the, the paint can version. And the monitor that I'm looking at right now is a Dell that I think is on its third new. I mean, I have a primary machine sitting next to my desk, and I think this is the third primary machine that this monitor has been connected to, simply because it just keeps working and it meets the, meets the need. Yep.

Gary [00:04:47]:
Yeah, so they came out with that, and then, then they also came out with new MacBook models. So new MacBook Air, new MacBook Pro, um, no major design changes. So basically, it's like the current models, but they put the latest processor, the latest memory and all that stuff in it. So, you know, again, if like you're not going to, you're certainly not going to replace one you bought last year. I may not, I'm probably still going to stick with my MacBook Air M2 from several years ago, not getting a new MacBook Air M5. But again, if I was in the market for one, this would be a great time to, you know, jump on one right now. Right. And so, it's good.

Gary [00:05:30]:
It's nice to have those. There are always, you know, sales always lag when people think, oh, it's been a while since they came out with one. I'll wait for the next version. You know, so it's nice to have like, oh, these are fresh new models that are out using the latest processors. Part of it was, you know, Apple came out with the M5 processor, the latest in their series, but they had an M5 and an M5 Pro. They came out with the M5 Max, processor to go in the higher-end MacBook Pros. So it's nice to have that. If you know, if you were, if you were looking at, if you're the type of person that just wants to buy the latest and greatest, no matter what the price, you were probably sitting on an M4 Max, you know, MacBook Pro or something.

Gary [00:06:11]:
And there was nothing to upgrade to. And now there is, there's an M5 Max. So that's what we have so far as Monday and Tuesday have gone by. For the most part. But unfortunately, we can only speculate what Wednesday is going to bring, and that is probably the new thing. And the rumor is— oh, I see. And this will be fun. This will be fun for people to listen to because by the time you listen to this, you'll know, and I don't.

Gary [00:06:39]:
So the rumor is that it's going to be a MacBook, a low-end MacBook, and the It'll be cheaper, maybe even down to the price of a Mac Mini. So it's going to have, you know, not a great screen, probably. It's going to be small, probably 13 inches. It's going to be, you know, the lowest end of all of that for home market, education market, that kind of thing. And it'll be fine for, for most people. I've got a MacBook Air as my second computer, and I know that that The MacBook Pro screen is far superior to what I've got. But that's fine if I really need to do something where the screen's important. I've got a good one on my desk.

Gary [00:07:23]:
But most of the time when I'm using my laptop, I don't. So the new MacBook is probably what they're going to announce on Wednesday. And the price will be the big thing. People that want to get a new one, that's going to be cheap. The rumors of how that price is cheap One of them is that it's not going to have a Mac processor in it. It will have a mobile processor in it, the same processor that's in the high-end iPhone.

Leo [00:07:51]:
Interesting.

Gary [00:07:51]:
Yeah, that's the rumor. I'm not convinced. And again, people can laugh at me now when they're listening to this. Oh, you know, Gary was wrong. But I'm not convinced because the iPads have— and not all of them, but the iPads in general have a Mac processor in them now.

Leo [00:08:09]:
I thought they were migrating everything to the M series of processors.

Gary [00:08:13]:
Well, definitely not the iPhones, right? But yeah, so why would you, like, why would the new iPad Air— oh, and by the way, that was another thing they announced was new, but they upgraded the iPads too.

Leo [00:08:25]:
Oh, okay.

Gary [00:08:26]:
I probably should have made a list before, but anyway, the new, like, the iPad Mini has an M M4, M5 processor in it? Like, why would not the low-end Mac have that, right?

Leo [00:08:38]:
Right, right.

Gary [00:08:39]:
So I'm not convinced, but maybe it'll turn out it's true. The other, you know, for a while people said this was gonna be a plastic Mac because Apple's done that in the past, so save some money. But most of the rumor sites now say no, it'll still be— it'll be aluminum. Aluminum is not gonna be a big thing for them to make. And the name is interesting because there's a last-minute rumor that instead of just calling it the MacBook, which is what was the name of the last time they came out with the low-end portable Mac, they're going to name it the MacBook Neo.

Leo [00:09:21]:
I will—

Gary [00:09:21]:
you'll know already when listening to this, but I I don't think it will be. The name coming from a piece of regulatory documentation for the EU listed a MacBook Neo as being released this week.

Leo [00:09:38]:
Interesting, right, right.

Gary [00:09:39]:
But Neo means new, right? And I don't think— I think a lot of these regulatory things, they don't need to say, here's the official name we're marketing this under, right?

Leo [00:09:50]:
Right. It's a code, not a name. Yeah, yeah.

Gary [00:09:52]:
So I think that they've maybe been using in various places, MacBook Neo. But we'll see, maybe, maybe I'll be wrong. I'll be— I'll be slightly surprised if it's MacBook Neo. It's fine if it is, you know.

Leo [00:10:07]:
Now it's not, it's just stupid. It just feels wrong. It feels kind of stupid. I've seen many things that have been named, you know, yada yada Neo, and it's just— no.

Gary [00:10:18]:
Yeah, I know. It's, uh, so I hope not. I— now the, the weird thing about the When Apple's done this in the past and come out with a cheap MacBook, the weird thing is, is that it ends up being smaller and lighter than the MacBook Air. So, you know, the whole idea is like, oh, MacBook Air, I want something really light, which is weird because it's half a pound lighter than the MacBook Pro.

Leo [00:10:39]:
Right.

Gary [00:10:39]:
So when you really, really try to compare them and say, oh, I want the lightest thing, but the MacBook Pro is not actually that much heavier if you get the same screen size. So do I really want to give up the power for just making it so just a little bit lighter, and I may not even notice. And then the MacBook comes out and it's like, oh, this one's cheaper and lighter. So I was willing to pay more for a light MacBook, and it turns out in order to get the lightest MacBook, I have to pay less, which is the situation I have been in before and might be in now if this thing is significantly lighter. Than the MacBook Air and it's cheap, I might get one just because I want this thing to be light. It sits on my lap, literally sits on my lap when I'm watching TV. It comes with— it gets thrown in my bag. Sometimes I go out for coffee or something like that in the morning.

Gary [00:11:35]:
I'll throw it into my bag and I may not even take it out, you know, but— or I may and decide to do like 15 minutes of going over some notes or something. So I want something really lightweight that's a Mac. So yeah, we'll see. I think if it is significantly lighter, I will just get one and make it my new laptop. And probably I'll probably do some videos on like what it can do. A lot of times when Apple comes out with a lightweight, lightweight in terms of like the power in it, people start to say things like, oh, don't get that if you want to render video. Don't get that if you want to, you know, run AI models or whatever it is.

Leo [00:12:14]:
Right.

Gary [00:12:15]:
And then I get it because it's— I want the cheap laptop and I'm like, it works perfectly. Like Final Cut Pro. Here's Final Cut Pro running on it perfectly fine. So I could see myself just wanting to get a $600 MacBook Neo and then just making videos that show like me rendering out 4K video.

Leo [00:12:35]:
And it's funny, given that it's a newer machine, right? It's going to have, you know, a decent processor. Not—

Gary [00:12:43]:
Yeah, well, see, that's the powerful— that's the thing. Yeah.

Leo [00:12:44]:
Yeah, but it's one of those things where, well, when you compare rendering on this brand new cheaper machine to like the machine from 2 years ago that people actually have been using, it's, you know, it's going to compare favorably. It just is.

Gary [00:12:57]:
So, or it's just going to be 20% slower, which for a professional is like, no, uh, nope. But for like, you know, people on a budget just trying to get something to like, you know, take their photos, make a slideshow out of or whatever, it's like, oh, I don't care about 20% slower. Anyway, we'll see and talk about it next week.

Leo [00:13:19]:
Cool. See if you're right or wrong.

Gary [00:13:21]:
Yeah.

Leo [00:13:24]:
You've been vibe coding.

Gary [00:13:26]:
Yeah. You and I and other friends have been talking about vibe coding and stuff. The interesting thing about vibe coding is I've been doing it for a long time. But usually I'd limit it to pieces of code. The idea is I'm creating something, maybe it's a complex web page, maybe it's working on one of my apps, whatever it is. I think, oh, I need a function that'll do something. All right. Well, let me just ask ChatGPT.

Gary [00:13:59]:
In this language, in this environment, give me a function that does this. Sometimes I do it for like, oh, I need like a MySQL lookup or a regular expression, something. It just gives this to me instead of me looking everything up. That's how I've been using it. Then I think I talked about this last week that I asked it for like a whole little website. Just because I'd heard other people vibe coding and being like, I don't know how to code, but I asked it to make this. I'm like, yeah, I'm sure there was more to it than that. But then I needed something just for personal use, just to basically, I wanted a little RSS reader that I could use to make it easier to read the news sources I want in the morning.

Gary [00:14:47]:
Just see a bunch of headlines. I had these ideas and I just wrote out what was supposed to be like a short paragraph explaining what I wanted. I just kept going and I wrote like 10 paragraphs describing a product.

Leo [00:15:01]:
Okay.

Gary [00:15:02]:
Then I hit submit. Then wouldn't you know it, it took about 5 or 10 minutes, but ChatGPT produced an entire web app. Not like, oh, here's the beginning, now where do you want me to go? No, it was like the entire thing done. I put it on my server and it worked. That was it, first prompt, and it worked exactly as I wanted. I spent a few hours excitingly adding to it because I thought, oh cool, now I can make this really neat.

Leo [00:15:30]:
So did you add manually or did you use ChatGPT to— to—

Gary [00:15:36]:
so at this point, I was— ChatGPT would, if it was a small change, it would say, here's what you need to change, like here, replace this function with this function, add this line after this line, that kind of thing. If it was big enough, it would say, here's a whole new file, replace this file with that file. It was doing that and I thought, wow, this is really cool. I tried a bunch of other experiments including even making a game, which is my thing. Create a game on the web that does this and it was doing really good. I was really finding it fun to build this way. Then I said, well, I think I know what the next step is. The next step is to actually give the AI tool direct control over the files.

Gary [00:16:25]:
Instead of it giving me code and me pasting it in and or modifying it, just give it direct control of the files. So as it turns out, that very weak Claude code was getting a lot of attention, right?

Leo [00:16:38]:
We were talking about it. Yep. I've been using it.

Gary [00:16:42]:
And so there are many ways to run Claude code. Actually, it turns out the I wanted one that would just be super fast for my workflow. What I did was I installed the Mac app for Claude, which is pretty good in and of itself, but there's a whole section for Claude code. I'm able to give it permission to work directly with files like in a folder on my Mac. Then on top of that, I know how on my Mac to basically say, oh, make that directory a PHP server. So it works just like a web server.

Leo [00:17:18]:
[Speaker] Really? Wow.

Gary [00:17:19]:
[Speaker] On my Mac. So now I'm not deploying anything to any live server. It's just sitting on my Mac, but it's producing stuff live. I don't have to work with the files, it's working with the files. So I could just ask for changes and it updates the files, adds new files, adds folders, just everything. It has this whole directory I've given it where it can work. Boy, is that fast. I thought just doing vibe coding was fast.

Gary [00:17:45]:
Now, this was like turbo speed vibe coding. So that was really cool. Then once I get what I want, I can upload it to a web page like on a site and get it going there. So I had the nice, my development server was local basically, and then I had a deployment server was the real server. Anyway, it's been a lot of fun. I did run into limits though because the question is, it's like, well, can I just ask it for anything? I did run into some limitations. I found a few times I got some error messages that I found other people were getting. Just sometimes it would, people are of different minds of why Claude Code gives you some of these error messages, but they seem to be persistent.

Gary [00:18:31]:
At one point I gave it an old game of mine and said, here's all the code. Make a new version of this that works on a web page. It thought for like 20 minutes and then gave me an error. No matter what I did, no matter how I broke it up, it just seemed to not be able to handle it at that point. It was like I broke Claude. Other people are saying the same thing. Then even if I went down and said, okay, let's just start, put a box on the screen, and it would still be like, no error. Starting from scratch, I was able to accomplish what I wanted, but it's a limitation.

Gary [00:19:09]:
It was a frustrating limitation. It's like, I don't know when that's going to happen again. I've like 5-coded 10 things in the last week and it happened with one of them. Why? Nobody seems to know why it's happened for some projects and not for others. So there's that. Also, of course, will people be doing this? Is coding dead? Like number one, will all coders just be writing code like this now? And number two, will anybody just become a coder? And I think a week of heavy use has taught me that it's not, I don't think people, just anybody will become a coder. There's still too many little things. There were still, as I would write the prompt, I would be like, I would think of other people in my life who do not know how to code.

Gary [00:19:57]:
And I would think, would they know to ask for that? Would they know to say that? Would they be able to, like, I think they would be able to ask for something in general. And then when it was there, it would be close to what they wanted, but then changing it to fit a specification to say this is exactly how it would work would stump them.

Leo [00:20:18]:
It's interesting because yes, that is the state of the world right now. I've already thought this about just AI in general, there's a tremendous amount of value in understanding how to ask for what you want. And that's true for just AI in general. And obviously, it's very true for coding. So in a sense, those folks who are being really successful right now are people that just know how to describe a problem in a way that will get them the answer they're looking for. What you're saying, of course, is that random people who don't have any programming experience don't really have the toolset to ask the problem, to ask for what they really want. Remember, this is the worst AI we have. Yeah.

Leo [00:21:10]:
In other words, it's only going to get better. And I fully expect that in some amount of time, I'm not going to say you know, how much time, I have no idea. But in some amount of time, I suspect that the AIs are going to get better at dealing with vagueness, ambiguity, missing specifications, right? So that you could ask for something very vaguely and the AIs will fill in the missing pieces and also realize that, well, you know, an app like that, that's also got to have this, right? Even though you didn't ask for this, we know that you're going to want it. When that happens, I don't know. But like I said, I think it's only going to get better from here. It's amazing what it does right now. Absolutely.

Gary [00:22:02]:
But, well, let me, let me— okay, this is an interesting direction.

Leo [00:22:07]:
Sure.

Gary [00:22:07]:
Because let's say let's get to a point where AI can totally do that. Yep. Right. So anybody can go and ask for any app they want.

Leo [00:22:17]:
Sure.

Gary [00:22:17]:
And we'll make it for them. So then there's two possibilities. One is you decide you want something. Let's use my example of an app that shows me all the news from the sources I want as headlines, and then I can click on the headline, and it works like an email client where I could say, I've read that, I've read that, I've read that, and it keeps like until I'm the inbox is empty. So that seems like a neat general-purpose thing. If somebody thinks, oh, that's what I want, right? I want to be able to go and handle my news articles like email, and they ask an AI for it, there's one of two possibilities. One is that already exists, like there's some product out there that already exists. So they're just recreating it from scratch, which is fine.

Gary [00:23:02]:
I mean, the AI may in the future say, oh, What you want already exists. You can just go here and do it. Or it may recreate that exact thing and give a person their own copy of it. The other possibility is it doesn't exist because nobody else wants it. Like it's very specific to what you want.

Leo [00:23:21]:
Right.

Gary [00:23:22]:
And that's interesting because then it will create an app that's just for you, which is really cool, but it's just for you. It's not ever going to be a product. No one else wants it. Somebody else is going to ask for something similar tomorrow in some other part of the world, and another AI is going to generate a different app that does things the way they want it. And neither of those two things will become products. They would just become things that each individual uses. So you have this thing where we cease to look for a general purpose solution that somebody is producing and only want custom solutions. Or it's just reproducing something that already exists and AI didn't need to go through it at all.

Gary [00:24:09]:
Maybe hopefully, for a lot of people, they realize it exists. And because it existed, they gravitated towards it. And they don't have to, like, if somebody wanted Pinterest, you know, they probably know Pinterest exists. And they're probably not going to think I would like Pinterest, but I don't want it to be Pinterest. I want it to be my own thing. It wouldn't even work very well because it's social, right? So then you would be like, well, I would be there alone by myself, you know? So you miss that aspect of it.

Leo [00:24:39]:
But look, so let's go back to the RSS reader because, um, there's a couple of interesting scenarios that I think, um, for example, I don't know why you needed to make your own because I have Feedly, and Feedly works very much, at least at that high level that you described. It's what I go into every day. It doesn't look like a mailbox, but it does act like a mailbox in the sense that I can read things and they disappear, right? So I suspect there's going to be a lot of that where people are going to reinvent the wheel. They're going to reinvent the wheel. However, what I also think is very likely is that you and I both know this, right? People have opinions. Features that they hate or, or miss or whatever. So I could easily see somebody coming along and saying, I want an app like Feedly, except I want it to do X, Y, and Z the way I want X, Y, and Z to be done. Um, I suspect that that is also very likely.

Leo [00:25:46]:
Now, the other thing that you're mentioning is that, you know, yes, these are all custom solutions for individuals, so never be a product. A, that's not necessarily a bad thing. But I think it's not—

Gary [00:25:57]:
I'm not saying it is.

Leo [00:25:58]:
Yeah, it does highlight though one of the things that I suspect is going to shape the market or shape what AI is used for over the coming months and years. And that is that some people are going to do exactly that. They just want a custom solution for themselves. They're willing to, you know, have it do exactly what they want. Obviously, they typically end up having some tolerance for, well, it doesn't quite do what I want, but I'm okay with that because it's just being and whatever. But there's another group of people, people that are actually associated with, who are really interested in using this vibe coding technique to generate product, to generate something that other people will pay for. And I'm thinking that that's an interesting fork in the road. It does have different applications.

Leo [00:26:50]:
Or different requirements or considerations for what you end up doing with AI and vibe coding, but it's just two different ways of looking at it.

Gary [00:27:01]:
Yeah, no, definitely. And we're gonna have to see what adoption is really going to be like. I mean, are people going to really build their own custom things? I mean, for me, for using something, there are many tools like Feedly just weren't doing it for me. Right? I was just not like, okay, this is not quite how I would have this work.

Leo [00:27:25]:
I still miss Google Reader.

Gary [00:27:27]:
Yeah, Google Reader is probably closer. So, you know, there's stuff like that. And, you know, will people— what will adoption really be like? I mean, there's just a lot of people out there, you know, with that are not looking to— they don't want their own version of something, correct? They want the—

Leo [00:27:49]:
they wouldn't even think of asking for—

Gary [00:27:51]:
they wouldn't even think of asking for anything. Yep. And then also, when you talk about products, it becomes something we've already run into, which is it's not necessarily the product itself but the user base. Like, you know, when you look at something like Twitter, you know, Twitter There's a bunch of different ones. There's Blue Sky and all the different—

Leo [00:28:16]:
Mastodon, Threads.

Gary [00:28:17]:
Threads and everything. Yeah. So it's not necessarily about the product anymore, but it's the user base, right? Twitter's worth something because there's so many people on it, so many organizations on it that it maintains this weight and the same with other social media networks. So if you create a product like a widget and the widget had just because you were first there or, you know, yours was most creative and you just had a head start on everybody and you got all the users to use your product. Everybody else could be like, oh, I could vibe code that in a weekend. Great. So there'd be 100 vibe coded clones, right, that wouldn't have any users. And then, and then sometimes it's, well, that's what makes the one valuable because it has users, but also it makes that what makes you want to use that one.

Gary [00:29:07]:
It's like if somebody said, oh, you want a, like a social network that has all these features? I've got one for you. Oh, there's nobody else there. None of your friends are there, but it has every feature you want, you know, as opposed to the other network that only has half the features you want, but everybody is there, you know? So, so there are other things to products.

Leo [00:29:27]:
Absolutely. There are other factors.

Gary [00:29:28]:
Yep. Yep. Yep.

Leo [00:29:31]:
So, yeah. So let's continue the AI talk here for a second. Um, it's interesting that you, um, added this. I have been— I don't want to say I've been fighting, but I've gotten a couple of notices from what boil down to copyright trolls. They notice that I have a stock photo on one of my sites And they tell me that, okay, we represent the person who currently owns that stock photo. Please prove to us that you have a license, which is annoying as hell. It's guilty until proven innocent. And if for some reason you can't provide the license, like say the stock photography outfit that you actually used for that photograph 15 years ago went out of business.

Leo [00:30:28]:
What do you do? Asking for a friend. The, so yes, what I've done, you'll, you've probably seen it on Ask Leo and I've been doing it on other sites as well, is I'm almost exclusively using either my own screen captures, of course, for the how-to stuff, but for the, what they call the hero image on articles or the featured image on various things, I've been using exclusively AI-generated images. There's nobody to claim a copyright because as you pointed out— well, go ahead with that.

Gary [00:31:08]:
Well, yeah. Well, first I wanted to say, when you first brought this up, it really made me think about it, about this idea that, yeah, what a lot of people in— a lot of artists, I know a lot of artists who talk about AI image generation and they hate it and all of that. And then the arguments are, you know, you should have a human do your art. It's better. AI looks, you know, fake. Artists need to get paid. They've stolen our artwork and have been training their models on it. And now you're getting the benefit, stuff like that.

Gary [00:31:45]:
But what a lot of the same artists don't realize is for people that publish things. Copyright trolls are a real problem. That's exactly what you're talking about. People that just make it their business to try to just extract money, usually fraudulently, by just looking around and saying, oh, you're using this image, pay us some amount that's just low enough that you're not going to—

Leo [00:32:10]:
Typically a few hundred bucks.

Gary [00:32:11]:
Yeah, you're not going to go and hire a lawyer.

Leo [00:32:13]:
Of course, so I've been told.

Gary [00:32:15]:
Yeah. So So, and it's annoying, and I may—

Leo [00:32:20]:
I'm— and some— a friend who is me may have in fact engaged an attorney.

Gary [00:32:28]:
Yeah, okay. So it, it's annoying and it's scary, uh, to have that out there. You know, you, you do the right thing, you, you pay for the royalty-free service, you— or back before, and I did this a lot, you buy a license. You know, they used to sell these things as CD-ROMs and DVDs. I remember those, yeah. Yeah, you know, and you would get a license and here would be 1,000 images of whatever, you know, and you're supposed to have the rights to do this. But of course, you know, 25 years later, do you still have the CD that had the little leaflet in it that had the little license that, you know, that company's not around or anything anymore and all of that? And you may not, you know, you took the image and you produced a JPEG for the web that's smaller and is devoid of any of the identifying information. So where, you know, it's all this problems we shouldn't have, but AI images not only are, you know, useful in this that they can't be copyrighted, but they also avoid this problem.

Gary [00:33:38]:
And the story out this week that relates to this, and the reason I thought we should bring it up is that the Supreme Court ruled or didn't rule, as it turns out, there was a ruling about that AI images can't be copyrighted. And federal court had made that. And it went to the Supreme Court. And the Supreme Court this week said, yeah, we're not going to say anything about it. In other words, let it stand. Now, it's a— I don't know, it's hard to know exactly what it means because obviously there are some ways in which there is copyright, but in the way that we're talking about it here, generating an image to use as like for your blog post or something, that is, yeah, no copyright. And even if there was, the copyright would have belonged to the AI image generator or the prompt creator. Which one would have been you, one would have been the company you're paying.

Gary [00:34:41]:
So you are still good, right? And you can very easily create an image. I, this week, did something based on like your talk about it before, is I unsubscribed to Adobe Stock.

Leo [00:34:57]:
I had subscribed to Adobe Stock. That's an expensive one too, yeah.

Gary [00:34:59]:
It's $30 a month. The reason I subscribed to Adobe Stock for $30 a month when they were better. Well, it's pretty good. It's a pretty good stock photography library.

Leo [00:35:09]:
Not the best.

Gary [00:35:10]:
Oh, really? Pretty good. Well, like Getty would have been the best, right? You know, or any— actually, Getty owns like half of the ones that are out there are actually owned by Getty, and that's why they have Getty Images. It's not because, oh, I could pay a little bit for a service that gives me access to Getty Images. Oh, I'm getting away with murder here. No, Getty owns that service. That's why it's so cheap to be able to get any Getty Images through it. But Adobe had a really good license at the time, years and years ago when I started with them, that understood digital. A lot of the stock photography services were still mired in legal agreements that had to do with print.

Gary [00:35:53]:
So you read their legal stuff and it's like, print this, print that, how many copies of this, books, magazines, newspapers, all this. It's like, no, I want to put this on my webpage. And you're not really addressing how that works or it's very vague. And Adobe right from the bat said, yeah, we're a digital company and we're gonna have that right there. It's gonna be really clear how you can use it. So I said, I'll pay extra because I want to make sure I know exactly how I'm allowed to use things. But I unsubscribed because AI image creation, it's like I can use that for just about anything I need. It's getting so good.

Gary [00:36:32]:
Heck, Adobe Stock was recommending AI images that it was generating most of the time unless you unchecked the box. They were like showing you all this stuff. And then I was thinking of your thing. It's like even if I'm getting the image from Adobe Stock and I have the license and it's all good, I'm worried about 10 years from now.

Leo [00:36:52]:
Yep. And it's a pain in the butt. And unfortunately, you and I both have this sunk cost. I've got hundreds and hundreds of stock images across my site. And I just, I don't know how, how much of a risk that is going to be in the long term. Interestingly enough, besides the copyright aspect of this entire scenario, I'm finding that I can get, I'll call them better images. Now, by better, I don't necessarily mean photographic quality or whatever. I'll talk about that in a second.

Leo [00:37:31]:
What I mean, I get an image that actually represents what I want. Yeah. Right? If you're searching stock photo libraries, you're looking for, especially in our world, you're looking for concepts or you're looking for computers and there's only so many hoodie hackers you can use on your website. And it's, but with, with AI image generation, I can actually describe a scenario that more closely represents the actual article that I'm using this image with and come up with something that just fits better. The other thing that I do, and I'm very explicit about this, is my images never attempt to be photorealistic. I'm, I've fallen back on, um, 3D animation style. So it definitely, there's a consistent style across all my images, but nobody will ever, uh, mistake them for reality, which I think is one of the more controversial aspects of what's happening with AI in general is that people are using AI image and video generation to try and make something that looks like it could be real. That's— I'm not interested in that.

Leo [00:38:48]:
That's not what these images are for. So at least I'm sidestepping that. But, but yeah, it's, it's giving me a lot more flexibility, a lot more power. And I'm sidestepping some, some potentially gnarly issues down the road.

Gary [00:39:07]:
Yeah, definitely. So anyway, I thought it was worth talking about. Again, I'm curious to see though how this ruling is going to affect other things. I mentioned about how copyright, it's weird because either the argument is that, well, I wrote this prompt. So writing this prompt is like making the brushstrokes and drawing a painting or pointing the camera at something and deciding when to click the shutter button.

Leo [00:39:40]:
Your prompt is probably copyrighted, right?

Gary [00:39:41]:
I don't know, according to the article I read, you're not, it's not.

Leo [00:39:45]:
The prompt, I would expect it because it's something that you've written.

Gary [00:39:48]:
I know, it seems weird, but now—

Leo [00:39:51]:
What the prompt generates might not be copyrightable, but it seems like the prompt itself should be.

Gary [00:39:55]:
But let's say for a second that, okay, you write a prompt describing an image you want, you send the prompt into the AI, the AI produces an image. The image isn't copyrightable. You didn't draw it, you didn't photograph it. You just gave a prompt to an AI, the AI generated it. Okay, let's say that's been decided. Now go to the next step, video, right? The next step is video, which now there are dozens of video generators out there. So it becomes a lot more fuzzy when you say, okay, I wrote a script, I made choices, I designed characters. I went to these image generators and I described people, and it drew them for me.

Gary [00:40:39]:
It just drew scenes for me. I fed that all in, I made choices, I used the script, and I generated all this video with AI. And, you know, do you own the copyright on the film that you just produced? And of course, if your mind is immediately going to like, oh, you're taking jobs away from actors and filmmakers and stuff by doing that, Consider, I'm talking about animated. It looks like Toy Story or South Park or whatever you like. Basically, you were using computers to generate images, CGI, and before using tools, you're now using tools on the computer to generate video the same. It's just a different process. You're writing a prompt, and believe me, it's a lot of work. It's lots of prompts, lots of attempts, lots of changes, lots of specifications, sometimes drawing stuff onto the frames and indicating things.

Gary [00:41:40]:
I mean, it's not simply, you know, here's, here's my 3 sentences, my pitch for a short film, go, you know, it's, it's tons and tons of hours.

Leo [00:41:50]:
If you take it to an extreme, if a movie studio Yeah. Say one of the big ones generates as part of one of their movies. I'll just say it's an 8-second, since that seems to be a common length right now, an 8-second sequence that was completely generated by AI. Is that 8-second sequence from this major motion picture house public domain? Yeah. They would probably say no. Right. But if we read the, the legal side of things, the implication is that, well, maybe it should be.

Gary [00:42:36]:
Yeah. Well, anyway, these are, these are things people are going to have to decide. We're going to have to figure out. And I think that there's going to have to be a line drawn. I don't think it's ever going to go to like no matter what you create, if you use an AI tool, no copyright. But on the other hand, I don't think it's going to be like, oh, you can give— you could say, you know, a rose on a vase on a table, go, and then the image you get and say, I own that now. Like, that's no— also, so the line is going to have to be in the middle somewhere. And it may be that you have to copyright certain things.

Gary [00:43:11]:
Like, if you want to produce a film, what you may want— need to do is write the script, for it. Of course, AI can write that as well. But then you copyright the script.

Leo [00:43:24]:
Sure.

Gary [00:43:24]:
You know, who knows, we may end up with a weird set of rules. Like I could see movie, animated movie studios with artists that their only job is to draw the base character.

Leo [00:43:36]:
Right?

Gary [00:43:37]:
Okay. So they come out and they, they produce hand-drawn, here are the characters, they get a proved by the director and the producer of the film and the writer and all that says these are the right characters. Now that you've produced those, we can copyright those characters, the artwork you drew, and we could feed that into AI and nothing beyond that point has a copyright. But you can't just steal the film, right? You would have to create your own characters and you can't steal the story. Well, the story you could, but you couldn't steal like the actual script because you copyrighted that as well. You know, so he basically makes it— that would be like probably a good way to do it because then you basically can't steal the film and you can't outright copy the film without ever saying, well, you know, any of the stuff that AI did was part of the copyright.

Leo [00:44:29]:
It's funny because we're kind of in that world already without AI when you think of what has fallen into the public domain in recent years. The original Mickey Mouse.

Gary [00:44:39]:
Right.

Leo [00:44:39]:
Yeah. Um, is, is now public domain, but prior to that, that character was copyrighted, which means that nobody could create any film or any animation using that character. Right. And I think that's just—

Gary [00:44:57]:
or trademarked. Wasn't the character trademarked?

Leo [00:45:00]:
Uh, whichever. It had legal protection of some sort.

Gary [00:45:03]:
Okay.

Leo [00:45:03]:
Trademarks I'm not sure about because Um, I don't think trademarks actually expire the way copyrights do.

Gary [00:45:08]:
No, no, they don't. You can keep renewing the trademark forever.

Leo [00:45:11]:
Right. But the point is that he did fall into public domain. So it's, it's probably wasn't a trademark issue anyway. Point being though, that that's exactly what you just described is people will create some piece of intellectual property, a character, and then that character is copyrighted. And if you were to duplicate or steal somebody's, you know, if they were to use AI to generate a film. Containing that character, it's not copyright-free just because it was generated by AI. The character within that AI-generated thing is still copyright, and therefore people can't steal it. Fun times.

Leo [00:45:47]:
I do think that there's going to be a tremendous amount of change here because I think a lot of these tools, uh, eventually are going to become the norm rather than the exception. Right now we're all on the bleeding edge, right? It's all new stuff, and it's all stuff that we're still trying to figure out. But over time, I suspect that as these AI generation tools get better, they're just going to become more of the norm. This is how you create a movie. This is how you do images. Just, yeah, just like it, you know, back in the day, cars— at some point, horses were how you got from city to city, and then cars came along. Yep. So, uh, I had it— I've had an interesting week.

Gary [00:46:30]:
Yeah, tell me about it.

Leo [00:46:31]:
A very, very interesting week. So this is very, very geeky stuff for those who are so inclined. I have two servers on the internet, web servers. I have a bunch of different sites. I mean, obviously I'm known mostly for askleo.com, maybe things like 7takeaways.com or notallnewsisbad.com, but there's another collection of sites that I host on one of my two servers. And I was running into an issue where the version of some software that I use, PHP, was out of date and I couldn't upgrade it because the version of the operating system that I was on was also too far out of date. So I actually needed to first update the operating system so that I could then update PHP. And the correct solution to that is not to try and update in place, it's actually to do a server move.

Leo [00:47:29]:
So you basically spin up a brand new server and start migrating. Uh, it took me probably 3 or 4 months to migrate askleo.com. And that's because of a variety of issues, uh, where I would try to migrate it and it wouldn't work. And I would fix something and it wouldn't work. And I would fix something and it wouldn't work. And I would give up in frustration and I'll say, okay, fine, I'll try again in a couple of weeks. Finally got all that handled. Then I basically now have askleo.com on a server running Ubuntu 24 long-term support.

Leo [00:48:14]:
It'll be there for a while. The version of PHP is all hunky-dory. I made some tweaks so that the server itself is running really, really well. And I'm actually looking at a scenario probably later this week where I'm going to be able to make the server smaller. These are all virtual servers that I use out at Amazon Web Services. And for a long time, I think I mentioned this a while back, again, AI related, I had to increase the capacity, double the size of the server because it was getting hit so hard by AI scrapers. And this new server with its new configuration seems to be handling all of that so much better. That, like I said, I should be able to just say, okay, AWS, make this server half the size, half the processors, half the RAM, that kind of stuff.

Leo [00:49:05]:
And that'll be fine. It's one of the reasons I like virtual servers, because if you want to make a change in its configuration, it's a button you push and a reboot, and all of a sudden you have a different looking server. So that's going to happen. The other server, on the other hand, had 50 websites on it of various sorts. Some of them are services, some of them are sites, some of them are like example sites. Like I've got yourveryowndomainname.com, which is something I use as an example in Ask Leo, but I actually have it hosted. So if you go there, it tells you that, hey, I'm an example site. I actually finished migrating all of those one at a time over the weekend.

Leo [00:49:55]:
What's neat about the current state of software is that very often moving a website— in the past, moving a website from one server to the other basically meant laboriously copying all the files, recreating the environment, firing up a new database, making sure that everything was pointed at all the right places. Essentially, you're almost rebuilding the thing from scratch. Patch. Today there are tools that let you migrate things. So for example, your very own domainname.com is a great example. On the old server I said back it up, on the new server I said restore it, and poof, there it was. All I had to do was go out to the domain name, you know, the DNS settings, and change the IP address. The IP address will become important here in a minute.

Leo [00:50:42]:
The— unfortunately, I was also— I shouldn't say unfortunately, but my original configuration had things organized improperly. So I took this opportunity to rearrange a few things, which meant that, well, instead of having the server make a move from one site to the other, I would have to make a different kind of change and maybe use a tool that would allow me to move the site from one place to another, but without all of the bells and whistles. So there was still a fair amount of cleanup to be done. The example would be on one server I had, well, like mp3.tehpodcast.com. Is where we host all of the audio files for teh.com. Um, what I had it on the old server as a top-level domain, which is the wrong configuration. What it really should be is there should be a tehpodcast.com, and then mp3.tehpodcast.com would exist as a subdomain, literally in the server configuration.

Gary [00:51:44]:
Yeah, that's the way I would do it. That's why I do it for lots of things.

Leo [00:51:48]:
Sure. But since teh-podcast.com lives on a completely different server, just treating mp3.teh-podcast was a quick way to get the thing up and running. Anyway, those kinds of rearrangements, because I had a bunch of them on askleo.com and askleomedia.com and a few other places, that ended up taking a little bit more work because in some cases, if they were WordPress sites, I could just tell WordPress to back up and then try and do a restore. In other cases, I had to manually copy things over and recreate. Anyway, all that was working. I got that all finished over the weekend. It was great. I was able to turn off two old servers, save myself some money, get rid of some hard disks, and I've got two servers running.

Leo [00:52:31]:
What I discovered when looking at things this morning is that the IP address of the server hosting 50 sites was by default a dynamic IP address. Which meant that should I ever reboot the server, the IP address could change, which meant that all the DNS records for all 50 of those sites would be instantly wrong. So I did the appropriate thing on AWS, which is to take a different kind of IP address and assign it to the server purposefully. Permanently. I had expected that while I did that, the, uh, both IP addresses would be— it would respond to both IP addresses, the dynamic one because I hadn't rebooted the service, the server, and the permanent one that I had just assigned to it. Sadly, I learned that that is not the case, and at the click of a button By assigning this permanent IP address to the server, all 50 websites were instantly offline because I couldn't reach them anymore. They were at a different IP address. So I spent the next hour running through all of my DNS settings and updating the IP address for every one of those sites that was hosted on that second server.

Leo [00:54:06]:
I think I got them all. I think I got it right. But it just kind of goes to, you know, for folks that are not in this world that don't deal with server and website hosting, the number of things that have to go right for things to work, especially when you're moving from one server to another, is pretty intimidating and pretty scary. It's gotten better over time, but it's still pretty intimidating and pretty scary. And as this little example shows, it's very easy to take your world offline with a single click of a button.

Gary [00:54:43]:
Yeah, well, at least it was a click of a button. I hate it when you didn't do anything.

Leo [00:54:49]:
You know, that actually would have happened.

Gary [00:54:51]:
Date changed or something, you know?

Leo [00:54:53]:
Yeah, yeah, yeah. You know, it would have happened at some point because the server, it's going to reboot at some point and it probably would have gotten a new IP address., and I would have spent probably an hour figuring out why aren't things working before I realized that that's what had happened. So I'm glad I discovered it. I mean, I'm glad ultimately I'm glad that things are fixed. It's just that the road from, from there to here was a, was a stressful one.

Gary [00:55:22]:
Cool.

Leo [00:55:22]:
Speaking of cool. Speaking of cool. Um, I, I, as you As I've mentioned before, I am a classic sci-fi buff. Yeah, I cut my teeth on folks like Asimov and Herbert and Heinlein and Clarke. And one of the things I did just before I terminated my Audible subscription last year is I used up my credits. And one of the audiobooks that I got was the collected stories of Arthur C. Clarke. Um, it's not, you know, not the novels and the novellas.

Leo [00:56:00]:
These are mostly the short stories and things that are smaller. It's like 51 hours of audio. There's a bucketload of stuff in there, and I started listening to it last week, and I am enjoying the hell out of it. Um, they're starting with his older stuff, and the man had a sense of humor from the start. There are some great stories in there. I'm really enjoying it. Of course, You know, there's a book, there's a, you know, the Kindle/paper book if you want to, but this is exactly the kind of stuff that I really like to listen to when I'm doing long-distance driving, which is starting to happen again. So the Collected Stories of Arthur C.

Leo [00:56:38]:
Clarke, highly recommended if you're into that kind of sci-fi.

Gary [00:56:42]:
Cool. I thought what was cool this week is I saw the movie Rental Family on Disney Plus this weekend. It's Brendan Fraser. Takes place entirely in Japan. He's an actor that's hired by a company to basically act out roles in people's lives. Like, for instance, he's hired by the daughter of a famous actor who's very old to pretend he's a journalist wanting to write a story about the actor. Because— and so it made her— making her grandfather feel important, you know, that kind of— so tons of different stories like that. And it's about his kind of journey.

Gary [00:57:23]:
And as an American, In Japan, he's in demand for specific roles where they happen to need an American-looking actor. So anyway, really, you know, it's the type of movie that either brings out emotions, brings out the feels in a good way, you know.

Leo [00:57:43]:
So I have mixed feelings about Brendan Fraser. He's done some really interesting stuff and he's done some really weird stuff that just doesn't appeal. So I'll have to see if this one applies. In terms of self-promotion, so this is an odd article that I want people to pay some time, spend some time with. It's How Does Cloud Storage Work? It's askleo.com/75658. And on one hand, you would think that cloud storage is just, as so many people have said, somebody else's computer. It's a place you put your files. But what I have come to understand is that people really don't understand how tools like OneDrive and Dropbox and Google Drive and iCloud all work.

Leo [00:58:30]:
They don't understand synchronization, and that ends up— it often ends up biting them because they'll end up deleting something and they don't realize that delete will be synchronized across all of the other things that it does. And they end up getting very, very frustrated when in fact the tool that they're using, its primary reason for existence, especially originally, was to do this kind of synchronization with cloud storage. So it is well, well worth understanding how these types of tools work before you load up all your valuable data. How does cloud storage work?

Gary [00:59:18]:
Askleo.com/75658. Cool. I believe I talked on this podcast not too long ago about being able to get your iPhone or any phone out and start recording video fast in case something is going on around you that you want to record. Yes. And I, I published a video on that subject. And actually what I determined is there is— there isn't one way like this is the way to do it. There are a variety of different ways. And it really depends on like, are you— do you feel you're really good at like hitting an icon on the screen? Do you want it to be a physical button on the iPhone? Various different things.

Gary [00:59:55]:
There were a bunch of techniques. So instead of saying this is the way to do it, I just went through a bunch of different ways that anybody can look at and maybe pick one that's close to how they do it now, but maybe my tips give them a way to even get it going faster. So I've got a video on that called 10 Ways to Start Recording Video with Your iPhone Fast. And this video is getting a lot of traction.

Leo [01:00:19]:
So I can imagine.

Gary [01:00:20]:
Yeah.

Leo [01:00:20]:
Yeah. Well, like when we talked about it, I remember, I mean, it's one of those things where there are unfortunately these days situations where that actually might be called for and speed can be of the essence. So. Having a shortcut to make that happen, or whatever solution you're using, um, can be very, very valuable. Yep. Cool. Alrighty. Well, I think that pretty much wraps us up.

Leo [01:00:45]:
Almost an hour to the exact, um, wraps us up for another week. As always, thanks for listening, and we will see you here again next— or next week, maybe real soon if not. Take care, everyone. Bye-bye. Bye.