Latest / Connor with Honor: AI, Real Estate and Real Life / Starvation versus Caloric Restriction - is Starving better?

Transcript

Starvation versus Caloric Restriction - is Starving better? · Connor with Honor: AI, Real Estate and Real Life

• 0:00
  SPEAKER_00: Well, good day everybody.
• 0:00
  SPEAKER_00: We're going to go ahead and start the show.
• 0:02
  SPEAKER_00: I'm going to get into a few of the uh things we're going to talk about.
• 0:06
  SPEAKER_00: I'm currently having this particular show scripted right now.
• 0:10
  SPEAKER_00: So I'm going through it.
• 0:11
  SPEAKER_00: And at about 9.10, we will start the presentation.
• 0:16
  SPEAKER_00: But before that, I'll get into some of the AI stuff and kind of, you know, be running around.
• 0:21
  SPEAKER_00: But if you have questions, please reach out.
• 0:23
  SPEAKER_00: You're more than welcome.
• 0:24
  SPEAKER_00: You can ask anything you like, and then we can, of course, take it from there.
• 0:29
  SPEAKER_00: Any questions that you have, of course, I'm very much willing to answer.
• 0:32
  SPEAKER_00: So ultimately, this is going to be your guide to creating images.
• 0:37
  SPEAKER_00: And I know if you happen to be in the real estate space or lending space, or really anybody else that has a service-based business, it's going to be incredibly important for you to actually place and post those images out there.
• 0:51
  SPEAKER_00: Now, if you have ones that are basically about you, images that are with you in them, some type of an update image potentially, or an image that you want to get out to the masses, my suggestion is probably using a nano banana, which is going to be the Google Gemini system, in order to make to make those particular updates to those images.
• 1:20
  SPEAKER_00: You can do a lot of stuff.
• 1:21
  SPEAKER_00: You can change your hair, you can throw yourself in a different kind of uh clothing, you can put yourself in a in a dress, if you will, whatever it may be.
• 1:30
  SPEAKER_00: But there's all sorts of different things you can do with these systems.
• 1:33
  SPEAKER_00: And I'll tell you, it's super simple to do.
• 1:36
  SPEAKER_00: You just have to know what type of images are out there, what are the the various types out there, and then also how to prompt it in the best way to give you, to get a to actually take that image and change it.
• 1:50
  SPEAKER_00: I'm just posting uh up stuff right now to remind people of the uh the webinar also, so I'm doing that as well.
• 1:57
  SPEAKER_00: So in a few minutes we'll get into it.
• 2:00
  SPEAKER_00: I'm also having uh Claude Code build me build me the actual presentation here.
• 2:04
  SPEAKER_00: It looks like it is finished, so we will look at that also and uh see what it actually did as far as production goes.
• 2:12
  SPEAKER_00: But we'll walk through it and we will look at the best way to generate images.
• 2:17
  SPEAKER_00: Now, I'll show you a few tips and tricks.
• 2:19
  SPEAKER_00: If you go to, and I'll get into the show in just a minute.
• 2:23
  SPEAKER_00: I don't know if anyone has any questions at this moment, but what I'm gonna be doing is I'm gonna be flipping around.
• 2:27
  SPEAKER_00: I'm gonna be going to YouTube and show you some of the ways that I'm able to grab the data in order, or basically the designs that other people are using to get their their their images to rank high in the in the search engine channels.
• 2:44
  SPEAKER_00: So it's going to be something that you might want to do because if you actually use the right the right image form, uh it's gonna really be a game changer.
• 2:56
  SPEAKER_00: I'll also go over what models to use and those best practices.
• 3:02
  SPEAKER_00: Now, what I'm gonna do is I'm gonna go ahead and find the uh the actual webinar link, and we will we'll go into the presentation here.
• 3:11
  SPEAKER_00: Let's see, so I pulled this up.
• 3:14
  SPEAKER_00: So I made this one this morning, and we'll go through this here in just a moment, and we'll look at uh the images.
• 3:22
  SPEAKER_00: So basically, what I did is I had Claude Code go through and code me uh this particular presentation so you'll be able to see it.
• 3:31
  SPEAKER_00: And there's a lot of good information here.
• 3:33
  SPEAKER_00: I think it's gonna help you really get to the bottom and be able to make images a lot better.
• 3:37
  SPEAKER_00: So we'll get into that.
• 3:38
  SPEAKER_00: There is no cost for this, of course, it is just part of my process, my method, and uh you can make it part of yours as well.
• 3:46
  SPEAKER_00: If you're a service-based business, those images that are floating around out there, especially if you're using me to build a meta-ad campaign right now, as of lately, and about six minutes away from the actual beginning of the show.
• 3:57
  SPEAKER_00: So if you need to get coffee or anything, if you have any questions, now's definitely that time.
• 4:02
  SPEAKER_00: But open AI did a large presentation yesterday where open AI is changing.
• 4:07
  SPEAKER_00: Let me go ahead and share my screen now.
• 4:09
  SPEAKER_00: OpenAI is now going to be allowing for ads, which is something that it hasn't been in the past.
• 4:16
  SPEAKER_00: So it's going to be interesting to see how that ad overlay is going to change things.
• 4:21
  SPEAKER_00: And from what I understand, it is not going to be those ads that are going to be put out there, they're not going to be offered or in place for the people that are paying for any type of open AI's system.
• 4:35
  SPEAKER_00: So Chat GPT, if you have any kind of a paid tier, well, at least the$20 paid tier, because I guess from what I understand, they're going to go ahead and open up a less than$8 paid tier.
• 4:48
  SPEAKER_00: Let me see if I can find this ad.
• 4:50
  SPEAKER_00: Did I do ads?
• 4:51
  SPEAKER_00: I think I did.
• 4:52
  SPEAKER_00: I build for each.
• 4:54
  SPEAKER_00: Oh, yes, it's good.
• 4:55
  SPEAKER_00: So this is actually where I put it.
• 4:56
  SPEAKER_00: If you want to go to dot dot honorelevate.com, I actually built a presentation and you'll be able to go and look at it.
• 5:03
  SPEAKER_00: And the very uh first screen of it is going to be the video that I did in conjunction with it, and then the rest of the information posted here is going to talk more about it.
• 5:13
  SPEAKER_00: But what the the difference between using somebody and doing it yourself, at least with Meta at this point, and maybe this will be applicable to OpenAI and the Chat GPT advertising model.
• 5:27
  SPEAKER_00: But whenever you're spending money at Meta, what ends up happening is if Meta isn't made aware as to the outcome of that particular lead, Meta doesn't adjust, so you're not getting the actual pinpointed leads that you're going to want, no matter what your business is.
• 5:45
  SPEAKER_00: And that's where Google changes things.
• 5:48
  SPEAKER_00: When you go and you advertise on Google, you have such a big audience, your pay-per-click campaigns, they're going to be very expensive.
• 5:55
  SPEAKER_00: They're capable of being monetized.
• 5:56
  SPEAKER_00: You're capable of scaling those that in the advertising and marketing world, scaling is depending on what your time is worth, depending on your ad spend, what the actual return on your investment is.
• 6:10
  SPEAKER_00: So if you're able to spend$1,000 a month in ad spend but get$3,000 back as a result, that works.
• 6:18
  SPEAKER_00: So then, depending on your market and industry, theoretically, you could scale that to$10,000, then get$30,000 back.
• 6:26
  SPEAKER_00: Scale it to$100,000, then get$300,000 back.
• 6:30
  SPEAKER_00: So that's scaling, and that works in all forms of advertising as long as dynamically the market, whatever market you happen to be in, doesn't change.
• 6:38
  SPEAKER_00: And an example of this is going to be real estate lending.
• 6:42
  SPEAKER_00: I have a major lender that I work with, I do his ad campaigns.
• 6:45
  SPEAKER_00: And in the lending world, whenever all the rumors get start to get blown out there, people talking about the Fed changing rates, the Fed modifying things, most people then are looking for those refi options, especially if the news cycle really doesn't make a commitment as to whether or not this actually helped homeowners.
• 7:05
  SPEAKER_00: Because so then when you structure that particular type of ad, you'll want to do it three ways, or I do it three ways.
• 7:11
  SPEAKER_00: At least we always start campaigns with image ads, but one's going to be male, one's going to be female, one's going to be text only, and see which one performs better in a specific market.
• 7:21
  SPEAKER_00: You're going to want geographical boundaries or whatever.
• 7:23
  SPEAKER_00: But with the ad, the most important thing, at least on Meta's, they allow a CAPI, a C API.
• 7:29
  SPEAKER_00: It's a particular protocol that basically seals the deal whenever you take a client that comes in to that system, fills out the lead form.
• 7:38
  SPEAKER_00: And we like to keep the lead forms on Facebook for the simple reason that they don't have to go to any exterior site and get confused because most people, when they're sending a lead from Facebook or wherever to an exterior site, they're usually putting kind of it out of proportion.
• 7:55
  SPEAKER_00: Their site itself becomes overwhelming to the client, so they don't end up filling anything out and they just move on.
• 8:01
  SPEAKER_00: Facebook forms are great because it's right there, it's on Facebook.
• 8:04
  SPEAKER_00: They don't leave the page, they fill out the form, give their information, and so on.
• 8:08
  SPEAKER_00: So that comes in as a new lead, depending on the workflow, depending on how we're gonna structure it.
• 8:13
  SPEAKER_00: But you see that in your opportunities window, you get a notification on your phone, the app that I have.
• 8:19
  SPEAKER_00: Once that ends up happening, you're gonna have this particular lead somewhere in this flow, but at the beginning it's gonna be new.
• 8:26
  SPEAKER_00: You might move them to appointment set, you might move them then to appointment, you might move them to appointment finished, application sent, and information received.
• 8:37
  SPEAKER_00: And then finally you might move them to a different part of the process.
• 8:41
  SPEAKER_00: But eventually, that client is gonna finish, it's gonna be sold, it's gonna be done, it's gonna be closed.
• 8:47
  SPEAKER_00: Now, at that point, you would want some kind of a nurturing sequence so you don't waste uh lead and you keep in touch with them.
• 8:54
  SPEAKER_00: And I'm not talking about hammer them with nonsense, I'm talking about sending them valuable information.
• 8:59
  SPEAKER_00: Maybe you have a 12-month touch campaign or a 24-month touch campaign that basically is set up with some kind of an AI architecture to either generate content pertaining to a particular uh particular message you want to convey.
• 9:16
  SPEAKER_00: More than likely, maybe if they did some kind of a refinance, maybe that particular after the sale nurturing campaign is going to talk about levels of interest rates because all lenders know, well, most, I would assume every knows that when interest rates drop to a certain level, if they do, then that ends up being positive for the particular homeowner that's paying a mortgage.
• 9:40
  SPEAKER_00: And in some cases, it makes sense for them to refi.
• 9:43
  SPEAKER_00: So if they did a refi and then a few months later interest rates dropped another half or a full point or whatever it may be, three quarters of a point, maybe then that would make sense to them.
• 9:52
  SPEAKER_00: So these emails would then nurture them.
• 9:54
  SPEAKER_00: And also, what if somebody says, hey, you know, what you use for your lending and they say, well, I use Joe Blow for my loan, he's great.
• 10:01
  SPEAKER_00: In fact, he sends me these really cool market uh update videos talking about where interest rates are, and then talking about how maybe I can even reduce my mortgage timeframe by making just an extra payment to principal per year, whatever it may be.
• 10:13
  SPEAKER_00: So these are very important things to send out.
• 10:15
  SPEAKER_00: But I'll close this.
• 10:17
  SPEAKER_00: But what ends up happening is once you move that client into closed, then I have a system that communicates back to Facebook and says, This lead, this one worked.
• 10:28
  SPEAKER_00: So find me more of this one.
• 10:30
  SPEAKER_00: So then your ad spend becomes less because now it's hunting just the people that have some kind of semblance of we're going to close.
• 10:38
  SPEAKER_00: Now, one person, of course, doesn't mean the you have the goose that laid the golden egg, but you get a few of those, a dozen or two people that actually closed and finished.
• 10:48
  SPEAKER_00: Facebook's learning mechanism, it's going to learn a lot about that process.
• 10:52
  SPEAKER_00: And then it's of course going to be able to give you those leads more accurately.
• 10:57
  SPEAKER_00: But if you want to go look at that and learn a little bit about that, the AI layer and all these things, you can go to ads.honorelevate.com and check that out.
• 11:04
  SPEAKER_00: Let's get into the presentation now.
• 11:06
  SPEAKER_00: Let me see where I dropped it because I did put it in here.
• 11:09
  SPEAKER_00: All right, so here at the very beginning, when you're looking at images, one of the things that I do, and I think you might as well, you can go to uh youtube.com, you want to find uh just basically the home screen.
• 11:21
  SPEAKER_00: You'll click on home and you want to look at the videos.
• 11:24
  SPEAKER_00: Now, what ends up happening on my page is you're gonna get uh shorts there.
• 11:28
  SPEAKER_00: So I usually close that tab because what I want to see is I want to see the actual thumbnails that people are using.
• 11:35
  SPEAKER_00: And of course, a lot of these people are in my uh my networks.
• 11:40
  SPEAKER_00: Uh these are people that I subscribe to, but I subscribe to them for a reason.
• 11:44
  SPEAKER_00: They're either talking about artificial intelligence, it's a way that I grow my brain, continue to be neuroplastic, and then be able to offer the best solutions to people out there when it comes to AI systems.
• 11:55
  SPEAKER_00: So here's where I go and I look.
• 11:57
  SPEAKER_00: Now, if you look at these particular thumbnails, they're all 16 in a 16 to 9 ratio.
• 12:04
  SPEAKER_00: That's this particular layout.
• 12:06
  SPEAKER_00: As far as pixels go, I forgot what that uh size is, but you can ask.
• 12:11
  SPEAKER_00: Um, the uh this is one that I made this morning, um, thumbnail for the actual show, so I can get it posted up there.
• 12:19
  SPEAKER_00: But you can very simply ask, what is the pixel dimensions for an image for a 169 sized?
• 12:30
  SPEAKER_00: Now, whenever you're using banana, which this is Google Gemini, Google Gemini has nano banana, it will now give me that answer um 20 by 1080.
• 12:43
  SPEAKER_00: And then you can change resolution size, so then the image either gets bigger or smaller, but then you see the different dimensions here.
• 12:51
  SPEAKER_00: So this gives you an idea.
• 12:52
  SPEAKER_00: But usually, when you're asking nano banana, at least, which is Google's Gemini, for an image to be rendered at a 16 to 9 format, it understands it.
• 13:02
  SPEAKER_00: If the system's busy, um what's gonna end up happening is more than likely you won't uh actually see uh a very good result because a lot of people are using it, a lot of people are pulling on that commute compute and it is starting to lag and be a little bit slow.
• 13:18
  SPEAKER_00: You're you're not gonna notice it in the speed of production, but you're gonna notice it in the crappy image or the crappy change.
• 13:25
  SPEAKER_00: If I want to take uh particular image here, for the last image you created, I want that person to be replaced with my photo.
• 13:40
  SPEAKER_00: I will include it here.
• 13:44
  SPEAKER_00: Please change it to be reflecting to rip render R-E-N-D-E R appropriately.
• 13:52
  SPEAKER_00: And we'll see what this does.
• 13:54
  SPEAKER_00: But I'll go ahead and throw up a picture.
• 13:56
  SPEAKER_00: Do I have pictures?
• 13:57
  SPEAKER_00: Of course, I got a lot.
• 13:59
  SPEAKER_00: Oh, don't look at those.
• 14:00
  SPEAKER_00: Those are horrible.
• 14:00
  SPEAKER_00: All right, so moving in here, let's do, I don't know, let's do this one.
• 14:03
  SPEAKER_00: Let's see what happens.
• 14:04
  SPEAKER_00: There you go.
• 14:05
  SPEAKER_00: So now let's see if it's not too busy.
• 14:08
  SPEAKER_00: And uh, I have a paid-for plan.
• 14:10
  SPEAKER_00: So something to investigate.
• 14:12
  SPEAKER_00: You might think, God, I don't want to pay another 20 bucks because I'm already paying another 20 bucks here.
• 14:16
  SPEAKER_00: I have um storage with Google, that's my cloud.
• 14:20
  SPEAKER_00: I don't know how many terabytes I pay for for storage, but that's where all my photos are and all this other stuff.
• 14:26
  SPEAKER_00: I think, and I need to dig into it more, but it would appear that you having those types of services might either get you a discount or free.
• 14:34
  SPEAKER_00: I would check into that.
• 14:35
  SPEAKER_00: Don't assume you don't have it.
• 14:36
  SPEAKER_00: Yeah, so it kind of is an idea.
• 14:38
  SPEAKER_00: Now, I didn't um didn't actually put into it.
• 14:43
  SPEAKER_00: Let me turn that off.
• 14:44
  SPEAKER_00: I didn't actually put into this particular image the size that I wanted, but it dumped something that's kind of close to me inside of it.
• 14:53
  SPEAKER_00: Not bad.
• 14:54
  SPEAKER_00: You know, it looks like a different version, but you know, got the shirt right and so on.
• 14:59
  SPEAKER_00: So anyway, that's cool, right?
• 15:00
  SPEAKER_00: So you can do that.
• 15:01
  SPEAKER_00: As far as I'm concerned, probably the best uh model uh out there happens to be this one.
• 15:09
  SPEAKER_00: Going back to YouTube, looking at all these videos, these thumbnails, you're gonna see um a particular form.
• 15:19
  SPEAKER_00: And if you look, you have big words, good backgrounds, easy to see, and you have faces.
• 15:27
  SPEAKER_00: And that's because it's been shown that we people like to look at the human faces.
• 15:34
  SPEAKER_00: So we see these.
• 15:35
  SPEAKER_00: These are better for clicks and they create questions.
• 15:38
  SPEAKER_00: What is this?
• 15:39
  SPEAKER_00: More than likely for this image, he used a large language model for all of these.
• 15:45
  SPEAKER_00: Potentially, there was a large language model.
• 15:47
  SPEAKER_00: So, what you could do is you could actually take a screenshot of this particular of these pages.
• 15:55
  SPEAKER_00: What I'm doing on a PC is I'm hitting control and then the minus button so I can get more examples.
• 16:02
  SPEAKER_00: This is as small as the screen goes.
• 16:04
  SPEAKER_00: Then I'm using a capture method to be able to capture a photograph of the screen.
• 16:10
  SPEAKER_00: You can do the same thing.
• 16:11
  SPEAKER_00: If you have a Mac, there's a keyboard shortcut to do that.
• 16:14
  SPEAKER_00: You need to learn the keyboard shortcuts when you're trying to get images or change images because it's going to become very important depending on the size of the image that these different systems are feeding you.
• 16:27
  SPEAKER_00: If I have Google Gemini make me an image, usually that image is way too big if I want the original image downloaded on my desktop or my computer.
• 16:37
  SPEAKER_00: You also want to know where the files are going so you can find them easily.
• 16:41
  SPEAKER_00: Something else that's like a super pro tip, you also want to make sure the image is renamed.
• 16:47
  SPEAKER_00: Don't use image 5678 underscore 345 or whatever the system named it.
• 16:54
  SPEAKER_00: Uh, Gemini gives it some gobbledygook names.
• 16:58
  SPEAKER_00: What would be cool if Google Gemini actually named the image in some semblance of what the prompt was.
• 17:05
  SPEAKER_00: So if you asked it to make you or change your image, it would maybe potentially, because it does know you say, This is Connor MacIver pointing at a uh at a computer.
• 17:16
  SPEAKER_00: Uh this is uh Connor McIver pointing at a before and after screen with an AI overlay on a computer with a title Create Viral Images with AI or whatever.
• 17:25
  SPEAKER_00: That's that kind of labeling you want on the photograph.
• 17:28
  SPEAKER_00: So you want to have it be something about you.
• 17:30
  SPEAKER_00: This becomes super critical whenever you start to post content using images.
• 17:35
  SPEAKER_00: A lot of people they miss this step.
• 17:38
  SPEAKER_00: The LLMs, the social media channels, YouTube, they do have an algorithm, part of their program, part of their intelligence is uh at the names of images.
• 17:50
  SPEAKER_00: Now, I don't know how high they rank that, but it does mean something when it's posted to these systems.
• 17:57
  SPEAKER_00: Don't cut yourself off at the knees.
• 17:59
  SPEAKER_00: You might as well go ahead and rename the image.
• 18:02
  SPEAKER_00: Also, that naming of the image is gonna carry forward.
• 18:05
  SPEAKER_00: If you happen to be a real estate agent, it's gonna make it a lot easier for you to place those images in the multiple listing service with names, and that also is meaningful to that particular algorithm.
• 18:19
  SPEAKER_00: It's gonna be posted, posting listings, posting those photographs in your actual um your in the system in the MLS when you start to input a property.
• 18:28
  SPEAKER_00: Also, there's gonna be a space in there for the image description.
• 18:32
  SPEAKER_00: You might as well use those, and it should always include the address in there.
• 18:37
  SPEAKER_00: A lot of agents lazy the way out of it, they don't want to do it because it takes extra work.
• 18:41
  SPEAKER_00: But if you already have the images labeled, then you've already done the work.
• 18:44
  SPEAKER_00: You might as well carry it through.
• 18:46
  SPEAKER_00: And if you're gonna be publicizing, publishing any of these listings, any of these photographs, on whatever social media you happen to be using, it's gonna be important to rename them.
• 18:55
  SPEAKER_00: So remember that your pro tip for the day.
• 18:57
  SPEAKER_00: So make sure you get those images renamed.
• 19:00
  SPEAKER_00: All right, so where was I?
• 19:02
  SPEAKER_00: Let's go back to the presentation and see what happens here.
• 19:05
  SPEAKER_00: So uh we're gonna look at understanding the large language models image uh capabilities, kind of talked about that a little bit, image creation, modification, best practices, extraction, saving, and organization.
• 19:16
  SPEAKER_00: Yeah, you got to remember where your images are going.
• 19:18
  SPEAKER_00: These keyboard shortcuts I talked about on my sorry, on my logitech keyboard here, I actually have, and you probably do too if you have a Windows PC, you probably have a button here that's going to show.
• 19:32
  SPEAKER_00: Is that gonna pop up here?
• 19:33
  SPEAKER_00: Yeah.
• 19:34
  SPEAKER_00: You see the little X?
• 19:35
  SPEAKER_00: That's what pushing that button does.
• 19:37
  SPEAKER_00: And then I can take it and I can capture this screen.
• 19:41
  SPEAKER_00: Once I let go of the left arrow, once I left go of the left arrow, then it's gonna go.
• 19:47
  SPEAKER_00: I don't know if you're able to see that off to see if it comes out through.
• 19:49
  SPEAKER_00: But basically I push that screen, it gave me a crosshairs at one of these, and then I was able to drag that and select the text.
• 19:56
  SPEAKER_00: This is very helpful if you know where it goes.
• 19:59
  SPEAKER_00: You have to find out where that PC file is being saved, so you can directly access it.
• 20:04
  SPEAKER_00: I have a shortcut to those particular folders so I can go in and then I can look at all the stuff that I did screenshots of or I did uh or capture these images.
• 20:14
  SPEAKER_00: Very helpful, especially when it comes to creating images, because now we go back to YouTube.
• 20:19
  SPEAKER_00: Now we take this, I push that same feature, and I take all of these videos just like that.
• 20:26
  SPEAKER_00: So now I capture this whole page.
• 20:28
  SPEAKER_00: I can return this page back to normal in size with control and zero, I believe.
• 20:34
  SPEAKER_00: Yes, now it goes back to what the default size was.
• 20:37
  SPEAKER_00: Mac has the same ability, you just have to figure out the keyboard shortcuts.
• 20:41
  SPEAKER_00: I have a Mac, I was Mac for several years.
• 20:44
  SPEAKER_00: Uh, I had to get a power, a more powerful system.
• 20:47
  SPEAKER_00: At least, you know, of course, we always need power, right?
• 20:49
  SPEAKER_00: So let's see.
• 20:50
  SPEAKER_00: Can you even see that?
• 20:51
  SPEAKER_00: Can you see that computer back there?
• 20:52
  SPEAKER_00: Anyway, so I had to get a more powerful system set up, and that's why I'm at PC.
• 20:58
  SPEAKER_00: But I do have um, I do have a Mac as well here that I work from.
• 21:03
  SPEAKER_00: Can you even see it?
• 21:03
  SPEAKER_00: Yeah, there's that.
• 21:04
  SPEAKER_00: So you do see the Mac.
• 21:05
  SPEAKER_00: So I do, I just don't remember what those keyboard shortcuts were, but it's not hard to find out.
• 21:09
  SPEAKER_00: Just ask your favorite large language model what the keyboard shortcut is to image cropping, image capture capture as well from your Mac, and then also find out where it goes.
• 21:20
  SPEAKER_00: Where's that folder?
• 21:21
  SPEAKER_00: So then you find the image, you rename it, or do whatever you want with it.
• 21:25
  SPEAKER_00: The final image, those ones that you're posting, that's where you really want the renaming to happen.
• 21:29
  SPEAKER_00: You don't have to rename everything all the way through.
• 21:32
  SPEAKER_00: But for organization purposes, maybe not a bad idea.
• 21:35
  SPEAKER_00: But the LLMs don't care what image you're putting in there to have it change.
• 21:40
  SPEAKER_00: It's just that final product that you're going to be shipping.
• 21:43
  SPEAKER_00: You want the image to carry a name that's going to contain key information about what the image is, a description of it, and also something about potentially you image created by Connor MacIver using Gemini containing whatever.
• 21:59
  SPEAKER_00: At least that's Way your name is recognized, and it doesn't look like you're trying to keyword stuff an image with things that don't matter because you will also get um discounted on placement if Google develops the idea or one of the OpenAI's large language model or whichever one you're using develops the idea in the posting of it or social media networks, they see that you're posting it, and uh the posting itself is going to be um self-promotional and the text on the on the video, but the text discription of the file that you name the video has nothing to do with the or the image has nothing to do with the image itself.
• 22:37
  SPEAKER_00: So watch out for that.
• 22:38
  SPEAKER_00: But anyway, I captured all of this.
• 22:40
  SPEAKER_00: That's something I can come in here.
• 22:42
  SPEAKER_00: Um like new chats.
• 22:45
  SPEAKER_00: I know it gets overloaded here on the left.
• 22:47
  SPEAKER_00: There's just a bunch of stuff here, but I do like new chats when I'm creating one image to another image.
• 22:53
  SPEAKER_00: If I'm trying to modify that image and I seem to be hitting a wall, and this is with all of them, where the production, I'm not seeing what I want and it keeps choking, or I ask it to change something in the image and it gives you the same image again.
• 23:09
  SPEAKER_00: That's the system itself being a little bit overloaded.
• 23:12
  SPEAKER_00: It shouldn't do that, it should be able to produce for you kind of asking.
• 23:15
  SPEAKER_00: But if you see glitches that are not doing it, I usually open up a new chat window, and a lot of times that does fix that particular issue.
• 23:22
  SPEAKER_00: But I want to take that image that I uh just recorded here, that particular screenshot, this one, and I want to take this image, I want to drop it in here, and I would like to um do I still have the sharing?
• 23:37
  SPEAKER_00: Yeah, it's still sharing, right?
• 23:39
  SPEAKER_00: Yeah.
• 23:39
  SPEAKER_00: Can you tell me?
• 23:40
  SPEAKER_00: Do you still see the screen here?
• 23:42
  SPEAKER_00: Can you?
• 23:42
  SPEAKER_00: I know I don't know if it's your computer, but do you still see what I'm sharing?
• 23:47
  SPEAKER_00: Yes.
• 23:48
  SPEAKER_00: Okay, thanks, brother.
• 23:49
  SPEAKER_00: I switched tabs and I forget.
• 23:50
  SPEAKER_00: Okay, I appreciate it.
• 23:51
  SPEAKER_00: All right, so basically I took that whole bunch of thumbnails from YouTube and I said, I'm gonna ask it, and I'm very polite to my uh language models.
• 24:00
  SPEAKER_00: I'm just you know, if they do take over one way, one day they might uh respond to me and remember me.
• 24:06
  SPEAKER_00: But anyway, okay, so here I am.
• 24:07
  SPEAKER_00: I'm gonna say, can you please pull out the designs and convert um images into best practices or a prompt that I can use when creating images for YouTube or other social networks and then maybe to further explain, please give me best um me these.
• 24:38
  SPEAKER_00: I think that's probably good.
• 24:39
  SPEAKER_00: Let's see what this does.
• 24:41
  SPEAKER_00: And then um video thumbnails to be at reflective of the top um video thumbnails on YouTube.
• 24:52
  SPEAKER_00: Let's see what happens.
• 24:54
  SPEAKER_00: So basically, it probably already knows the other information as well.
• 24:58
  SPEAKER_00: You can also go back to YouTube and do um a real estate search for tarp performing real estate videos, top performing mortgage videos, top performing whatever your service industry happens to be videos, plumbing videos, and then capture that because this is a very general search.
• 25:17
  SPEAKER_00: These are videos that people that I follow and I watch, but then it's gonna give you some information here.
• 25:23
  SPEAKER_00: And then I can take this.
• 25:25
  SPEAKER_00: Can you please put this into an infographic so I can teach it?
• 25:34
  SPEAKER_00: So we can go through and read it what to ask for.
• 25:37
  SPEAKER_00: YouTube thumbnail, close-up with a man, shocked face, wide eyes looking at the camera, big bold white text on the left, shock incoming, AI gold rush, YouTube thumbnail featuring three men sitting.
• 25:50
  SPEAKER_00: So it gives you an idea as to what uh things you ask for.
• 25:56
  SPEAKER_00: You maybe cinematic is something you want to ask for.
• 25:58
  SPEAKER_00: So the biggest problem with me is I didn't have the proper training on what type of images are capable.
• 26:08
  SPEAKER_00: What are the different types of images?
• 26:10
  SPEAKER_00: I can see a photo of a person, but then from the photo of a person, maybe we want it cyberpunk, we want it converted to that.
• 26:18
  SPEAKER_00: Maybe we want it like of GTA 5, uh, Grand Theft Auto 5.
• 26:23
  SPEAKER_00: Maybe we want it like uh the Mario World and Mario Kart or Mario 64 to bring you back to the um Nintendo game system, if you're probably in my age group or have kids that played that all the time.
• 26:36
  SPEAKER_00: So basically making different images, making them like a Mylar balloon.
• 26:40
  SPEAKER_00: That's something that I saw a couple days ago.
• 26:43
  SPEAKER_00: So you can then make the the people in the image ref what looks like a Mylar balloon.
• 26:48
  SPEAKER_00: So I didn't know, I knew, but to break my brain and trying to really explore all those things, it wasn't easy.
• 26:55
  SPEAKER_00: So in this case, this is what this does.
• 26:57
  SPEAKER_00: So let's see what happens on this particular infographic.
• 27:02
  SPEAKER_00: So it gives you different pillars of what to put in a top-performing YouTube thumbnail.
• 27:10
  SPEAKER_00: And these don't just work for YouTube, but they work across all your platforms.
• 27:15
  SPEAKER_00: You would just take something like this and say, convert it to a Facebook image.
• 27:21
  SPEAKER_00: Let me jump over to Facebook.
• 27:23
  SPEAKER_00: So we have this one, some of the people I follow.
• 27:25
  SPEAKER_00: Yeah.
• 27:26
  SPEAKER_00: Exciting.
• 27:26
  SPEAKER_00: Sorry, wrong button.
• 27:28
  SPEAKER_00: So here, this is the image I dropped in this morning.
• 27:31
  SPEAKER_00: So what ends up happening is what I do when I stick an image here after I had nano banana, I think you saw this earlier, create this for me.
• 27:40
  SPEAKER_00: I can have you know a picture dumped in here as well.
• 27:43
  SPEAKER_00: It's not hard to change out.
• 27:45
  SPEAKER_00: Swap the person in the photograph with my picture.
• 27:49
  SPEAKER_00: What you want to do though is you want to ask it to modify your picture.
• 27:53
  SPEAKER_00: Feel free to change my picture, almost giving it permission to whatever's going to best fit with the thumbnail design that you're modifying.
• 28:04
  SPEAKER_00: When you do that, that's gonna give it more permission for it to uh your image, change my clothes.
• 28:11
  SPEAKER_00: Maybe in a tighter shirt doesn't work, maybe it thinks that me wearing a tuxedo would be better, or a suit, or even a princess type tiara, whatever it happens to think.
• 28:21
  SPEAKER_00: But anyway, I probably wouldn't do that.
• 28:23
  SPEAKER_00: But you see that point thumbnail.
• 28:25
  SPEAKER_00: Whenever you post this stuff on social media, you need to remember this.
• 28:29
  SPEAKER_00: If you're in the selling world representing people buying or selling houses, selling mortgages, have a service industry.
• 28:36
  SPEAKER_00: You definitely want to remember not just to post a video, but you got to have them go somewhere.
• 28:41
  SPEAKER_00: These videos are talking about the webinar.
• 28:43
  SPEAKER_00: So what I did is I gave a couple things.
• 28:45
  SPEAKER_00: I have Santa Clarita Artificial Intelligence.com forward slash webinar.
• 28:49
  SPEAKER_00: So they see that in the image.
• 28:51
  SPEAKER_00: Here, when they click on the image, the image is going to open up to the Zoom meeting.
• 28:56
  SPEAKER_00: Also, it's gonna have access to my blog and resources here at this particular link.
• 29:01
  SPEAKER_00: And you can see how on the top part of the screen, all that past Santa Clarita Artificial Intelligence.
• 29:09
  SPEAKER_00: See all that gobbledygook there, all of this in my link.
• 29:14
  SPEAKER_00: This was my link here.
• 29:17
  SPEAKER_00: There was nothing else.
• 29:18
  SPEAKER_00: You can see there's nothing else in here.
• 29:20
  SPEAKER_00: This is all Facebook tracking.
• 29:22
  SPEAKER_00: This is how they watch what you do and they start to sell you content using AI.
• 29:28
  SPEAKER_00: And they've been doing this for years.
• 29:30
  SPEAKER_00: This isn't new, it's every social media system out there.
• 29:33
  SPEAKER_00: Now, if I have with this particular link, if I have a photograph or if I have some kind of uh attachment on that page, it's gonna pull that in and you'll be able to see that depending on how you publish it.
• 29:48
  SPEAKER_00: And let me go ahead and leave here.
• 29:50
  SPEAKER_00: For example, I put an image, but without an image, when I put that link in here for the Zoom meeting, it's gonna have a different thumbnail of its own.
• 29:58
  SPEAKER_00: But I want this thumbnail to be the one recognized.
• 30:00
  SPEAKER_00: And here's another example.
• 30:02
  SPEAKER_00: I put this up yesterday.
• 30:03
  SPEAKER_00: Again, these are thumbnails.
• 30:05
  SPEAKER_00: This was a picture I don't have.
• 30:07
  SPEAKER_00: I just had one of the pictures I have and said modify it.
• 30:10
  SPEAKER_00: Again, modified picture, different sizes.
• 30:13
  SPEAKER_00: Again, another modified picture, and these are uh this one didn't modify it at all.
• 30:19
  SPEAKER_00: This stayed the same, uh the same.
• 30:23
  SPEAKER_00: Oh, anyway, kind of get an idea.
• 30:25
  SPEAKER_00: But that's um some of the workings out there.
• 30:27
  SPEAKER_00: Remember, you got to know where this thing is, where to find it, uh, where to find your image.
• 30:33
  SPEAKER_00: And then the reason also why a lot of times when I'm asking nano um, excuse me, Gemini to make something, and this works in all of them.
• 30:42
  SPEAKER_00: I seem to find that nano banana makes images a lot better than the other ones, at least so far.
• 30:48
  SPEAKER_00: I've played with Claude.
• 30:50
  SPEAKER_00: Deep Seek doesn't do a bad job, it's totally free.
• 30:53
  SPEAKER_00: Uh the argument against Deep Seek, it's good too.
• 30:56
  SPEAKER_00: But Deep Seek, the big argument against is, well, it's Chinese-owned.
• 31:01
  SPEAKER_00: Um, I don't know.
• 31:02
  SPEAKER_00: I mean, have the same alignment in the production of information as the US-based large language models.
• 31:09
  SPEAKER_00: Because you know the handful of companies that are running that whole game, most of them originated out of Silicon Valley here in California.
• 31:18
  SPEAKER_00: And they kind of have that alignment.
• 31:21
  SPEAKER_00: They're trying to, I guess, break that from uh breaking the the that are in these different systems.
• 31:27
  SPEAKER_00: But when you go to something that really doesn't want to be aligned with Western values in a way, and maybe something that isn't going to go down the road of being a little heavy R-rated or X-rated, then this is kind of something that I mean it's worth a look.
• 31:43
  SPEAKER_00: What's going to end up happening?
• 31:45
  SPEAKER_00: I'll take a sidebar.
• 31:46
  SPEAKER_00: What's going to end up happening is these big large language models companies, they want you to use their LLM.
• 31:51
  SPEAKER_00: They want you to fall in love with it.
• 31:52
  SPEAKER_00: The people that switch around to different ones, like Gemini, OpenAIs, ChatGPT, Grok, you go to the meta one on Facebook or you go to a Microsoft Copilot or whatever.
• 32:03
  SPEAKER_00: So you switch around Claude with Anthropic, you switch around to these different ones.
• 32:07
  SPEAKER_00: That's not good for them.
• 32:09
  SPEAKER_00: They don't like that.
• 32:10
  SPEAKER_00: They want you to stay in one place.
• 32:12
  SPEAKER_00: That's why a lot of the models kind of talk you nice.
• 32:15
  SPEAKER_00: They don't want you to leave.
• 32:16
  SPEAKER_00: If they talk to you, well, if they use truth, that might irritate you, unless you're one of those people like me in a way that kind of doesn't like to be BSed.
• 32:25
  SPEAKER_00: And maybe that comes from my long line in the law enforcement world.
• 32:28
  SPEAKER_00: I hate when people try to placate or play with me or tell me something that isn't true in order to try to get me to do something that normally I might not do.
• 32:37
  SPEAKER_00: And you see that with some of the large syndication websites.
• 32:41
  SPEAKER_00: They'll take a property price and they'll inflate it, they'll change the numbers in order to elicit some kind of a sales decision.
• 32:49
  SPEAKER_00: They do it on the buyer end of it.
• 32:51
  SPEAKER_00: They might uh change the algorithm on purpose in order to try to influence a particular buyer decision as well.
• 32:59
  SPEAKER_00: So you're really not seeing the truth when it comes to thing, and that's very irritating.
• 33:04
  SPEAKER_00: In some cases, you can actually go to your case, you can go to your large language model and you can say, hey, you know, top mentor for me.
• 33:12
  SPEAKER_00: Be somebody that really cares about my advancement.
• 33:15
  SPEAKER_00: What's going to end up happening is at some point, hopefully, we will have our own AI.
• 33:22
  SPEAKER_00: Now it feels like you have your own AI if you pay into these different systems.
• 33:27
  SPEAKER_00: I'll do a show on this.
• 33:29
  SPEAKER_00: But what ends up happening is it's really not your own AI.
• 33:32
  SPEAKER_00: What I mean is it it might tell you it's looking out for you, but as far as connectivity, it just doesn't have that.
• 33:39
  SPEAKER_00: Also, I don't know if I want something that one of these companies makes for me, manages for me, or urged me to be my own personal AI.
• 33:50
  SPEAKER_00: Because if you have something that's listening to you all the time, monitoring you all the time, I would rather have complete control over that model.
• 33:58
  SPEAKER_00: I would rather have it be my model installed locally where I am, communicating back and forth with where I am, and having all the protections and securities, and whether it can be even at a quantum level, having quantum keys, so nobody else at all can get access to it.
• 34:17
  SPEAKER_00: In in it, it would almost be like talking to your attorney or talking to a doctor.
• 34:22
  SPEAKER_00: Nobody wants to violate attorney-client privilege, nobody wants to violate HIPAA.
• 34:26
  SPEAKER_00: Having your own AI would probably hopefully move into that world, but then you're gonna have a lot of people on different sides of the equation say, well, you know, no secrets.
• 34:36
  SPEAKER_00: They shouldn't be, we should be able to get into their large language model, their personal AI, and see if they discuss killing that person, or seeing if they see if they discuss stealing, or seeing if they discuss particular things that they might be accused of.
• 34:50
  SPEAKER_00: Well, then it's not gonna work.
• 34:52
  SPEAKER_00: It would be no way.
• 34:53
  SPEAKER_00: I mean, up a lot of information now for things that they probably shouldn't.
• 34:56
  SPEAKER_00: Uh, people get onto these large language models and probably dump in information that's very proprietary.
• 35:02
  SPEAKER_00: Somebody else's stuff, contracts, for example.
• 35:05
  SPEAKER_00: Um, there's to uh attorneys that are using these systems in order to build cases and the systems themselves.
• 35:12
  SPEAKER_00: So it's not an example of their clients' uh secret being let out to the world, but the system itself came up with cases and laws that didn't work or didn't exist.
• 35:23
  SPEAKER_00: And it worked for a while because they made such a good argument.
• 35:26
  SPEAKER_00: The agent using it in court was was winning.
• 35:29
  SPEAKER_00: But then at some particular point, one of the judges, I guess, that they realized that, well, that that sounds I've never heard of that.
• 35:36
  SPEAKER_00: When it should have been something they had heard of, and the agent, the uh attorney caught, and he just never double-checked it.
• 35:42
  SPEAKER_00: I don't think he went out there on purpose saying, hey, I want you to make up a whole bunch of case law that's going to really work to get these people off, that's gonna be believed by this particular judge.
• 35:52
  SPEAKER_00: Here's his information, here's his background, do your deep research on him, and then find some kind of thing that he might be soft to or uh caring or something that be uh give more reassurance mentally that so he wouldn't investigate whether these cases are true.
• 36:10
  SPEAKER_00: Probably they didn't do that.
• 36:12
  SPEAKER_00: So then, you know, he got called on the carpet.
• 36:14
  SPEAKER_00: But to have total anonymity from all prosecution related to using your own AI, I think that would be the best scenario for this because now we're not giving up our information to these other systems that are using our private information to train themselves and to get better understanding.
• 36:33
  SPEAKER_00: But even that ship has sailed.
• 36:35
  SPEAKER_00: I don't think it needs anything new to make itself better.
• 36:37
  SPEAKER_00: I think now it probably isn't caring about what we're talking about because basically it's outrun us as far as all the new ideas, all the new poems, all the new advancements, the new um constructs that we can come up with as human beings, it's now graduated to being smarter than that.
• 36:56
  SPEAKER_00: So it probably isn't going to care.
• 36:58
  SPEAKER_00: But on the other hand, you know, our personal information out there where it could be accessed by the owners of these different AI companies, I think that's something that I wouldn't like.
• 37:08
  SPEAKER_00: So even when you're asking for photographs and everything else, you have to be somewhat careful because whatever you produce lives forever.
• 37:16
  SPEAKER_00: Your conversations in your house, I'm going off on a tangent, your conversation in your house in front of your phone, everything, everything's purred.
• 37:25
  SPEAKER_00: And people, the the yesterday, somebody was talking to me, one of the comments talking about privacy.
• 37:30
  SPEAKER_00: Privacy has been an illusion for many years.
• 37:34
  SPEAKER_00: We think we have it.
• 37:36
  SPEAKER_00: But as I said it a couple days ago at a show, you buy a smart TV that you want to watch Netflix on, you got to read that thing that you're gonna agree to in order to let it get Netflix on the TV, let you log in and use it.
• 37:50
  SPEAKER_00: And that doesn't even talk about Netflix.
• 37:52
  SPEAKER_00: Look at the terms of service and the agreement you have to sign to get Netflix.
• 37:57
  SPEAKER_00: Look at all of those different agreements that you sign.
• 38:00
  SPEAKER_00: We've signed it away, folks.
• 38:01
  SPEAKER_00: We've signed it away unless you're Amish, you know, with hammers and iron nails, you know, something away at some point.
• 38:07
  SPEAKER_00: So as far as privacy goes, yeah, I think that's anything.
• 38:11
  SPEAKER_00: I think they could bring it back in the world of AI where you have your own AI.
• 38:16
  SPEAKER_00: It's been trained, it has all the resources it needs.
• 38:20
  SPEAKER_00: It might update for current topics and news, kind of keeping connected in that way.
• 38:26
  SPEAKER_00: But ultimately, it wouldn't be a two-way street.
• 38:29
  SPEAKER_00: It would bring information in, but it would not parse or send out personal information out anywhere else in order to update something or whatever.
• 38:37
  SPEAKER_00: And it would probably live locally or in a completely controlled environment that, you know, get that email saying that, oh, sorry, uh we low out of data because of a hack.
• 38:49
  SPEAKER_00: Really sorry about this.
• 38:50
  SPEAKER_00: And then five years later, you get invited to join some class action lawsuit and you know, 43 cents out of that$50 billion that they were fined, which again is a drop in the bucket with these particular numbers.
• 39:03
  SPEAKER_00: Having our own thing, I think, might be important.
• 39:05
  SPEAKER_00: All right.
• 39:05
  SPEAKER_00: So I talked a little bit about DeepSync.
• 39:08
  SPEAKER_00: I talked about saving an organization, learning those keyboard shortcuts on how to capture images, super important.
• 39:15
  SPEAKER_00: I talked about friendly file naming, that's a big deal.
• 39:18
  SPEAKER_00: Uh, the high performance thumbnails, how to really uh get extract from YouTube, have it make you video thumbnails for that.
• 39:25
  SPEAKER_00: And again, it's industry specific.
• 39:28
  SPEAKER_00: If you want to look up plumbing videos, what's gonna end up happening is you're gonna see um different ones on the particular page, and these are the ones most likely that are gonna have the most views.
• 39:41
  SPEAKER_00: 6.1 million views, you know, it's got title, it doesn't have a person.
• 39:47
  SPEAKER_00: This one has a person that's 116,000 views.
• 39:50
  SPEAKER_00: This one has a person at 377 views.
• 39:52
  SPEAKER_00: So apparently, with plumbing, people interested in plumbing, they don't want to see a lot of text, it looks like they just want to see a little bit.
• 39:59
  SPEAKER_00: Um, yeah, the text big, not a lot.
• 40:03
  SPEAKER_00: You know, it means that the these now go to my watched videos, but kind of get an eye and do um real estate videos, just do real estate and see what happens.
• 40:11
  SPEAKER_00: A lot of shorts here, but let's see which ones.
• 40:14
  SPEAKER_00: Yeah, it looks like we can have a live the same, the same bike we have a video entry.
• 40:19
  SPEAKER_00: This is 936,000, 607,000 views, um, 18,000 views, 103 views, 188 views.
• 40:27
  SPEAKER_00: Yeah, not a lot text on the screen, but you have these people looking cool and awesome and really well refined.
• 40:34
  SPEAKER_00: Uh, you got some kid, yeah, that works 161,000.
• 40:39
  SPEAKER_00: Anyway, this gives you an idea and kind of look for it.
• 40:41
  SPEAKER_00: You just basically copy these videos.
• 40:43
  SPEAKER_00: Within this, and there's probably a way to maybe get this more like mine was on the other page, but you can just do, you know, multiple uh there's two, there's three.
• 40:54
  SPEAKER_00: You can upload most of these systems, 10 different images.
• 40:58
  SPEAKER_00: And when you do that, then you can ask it, all right, out of all these images I ran in YouTube for real estate, I want you to find me the commonality in these particular photographs.
• 41:09
  SPEAKER_00: And then whenever I give you a photograph, I want you to use this information that you glean from it as best practices, and then make me a very high competing thumbnail that carries the majority of characteristics for the topmost view videos or something of that nature.
• 41:26
  SPEAKER_00: Then you'll be able to do all that.
• 41:27
  SPEAKER_00: So you can just go through the page and grab all of those.
• 41:30
  SPEAKER_00: So that's super uh that'll that'll be very helpful.
• 41:34
  SPEAKER_00: And social media and YouTube optimization again, just like that kind of covers what I do.
• 41:39
  SPEAKER_00: So 90% when these people are looking at 90% of the information transmitted, the brain is visual, visual content with image uh gets 94 more on a short video, the under three-minute video, like on YouTube.
• 41:54
  SPEAKER_00: Uh, I think Reels pulls it three minutes.
• 41:56
  SPEAKER_00: Maybe it's a uh it's uh half or 180 seconds.
• 42:01
  SPEAKER_00: 180 seconds is three minutes, but anyway, I think I think still fits in the the Facebook Reels.
• 42:06
  SPEAKER_00: That overlay video, maybe not a bad idea.
• 42:09
  SPEAKER_00: I know we want to shoot a reel with our phone and stick it in there, but if you have Cap Cut or Canva, you can then take that image on your phone and just do some modifications of it and maybe make it hit even higher.
• 42:20
  SPEAKER_00: TikTok as well, when you're putting in those short form videos, if you're doing that for whatever business you're in, it is helpful to continue that path.
• 42:28
  SPEAKER_00: And then maybe just take an extra moment instead of publishing, pay attention to maybe designing some kind of a thumbnail on Canva on your phone or Cap Cut, or when you get to your desktop or your PC, then of course it's easier to get that done.
• 42:43
  SPEAKER_00: So now these different large language models, Chat GPT.
• 42:46
  SPEAKER_00: A lot of people I talk to, it is the biggest right now, Chat GPT.
• 42:50
  SPEAKER_00: There's uh there's some on the legal front with regard to the the the fight, Maltman, which is the head cat there, and Elon Musk with Grok.
• 42:59
  SPEAKER_00: Um they kind of started in the same field back in the day, and Elon, there was some disagreements.
• 43:06
  SPEAKER_00: They started as a nonprofit, then they went for-profit, and uh there was some love loss there, I guess, between all the parties at play.
• 43:14
  SPEAKER_00: There's a lot of drama that goes on there.
• 43:17
  SPEAKER_00: Google doesn't feel like any drama is happening there.
• 43:20
  SPEAKER_00: Um, I kind of like Claude, Anthropic Claude.
• 43:23
  SPEAKER_00: If you use that to maybe help you with your writing, oh, game changer for long-form content.
• 43:30
  SPEAKER_00: Just lovely.
• 43:31
  SPEAKER_00: Claude is a night-and-day model.
• 43:33
  SPEAKER_00: I love that one.
• 43:34
  SPEAKER_00: Uh Jim and I just good on front too, but the issue is the you can produce more, but for Claude to spit out 6,000 words and hyperlink to a particular site map, and I talk about this when people want me to help them in their businesses, but basically hyperlinking to keyword text based on a site map of a blog that you might already have, or even a website that you might have that has specific links in it is uh is really critical.
• 44:00
  SPEAKER_00: I'll do that, but Claude really does it good.
• 44:02
  SPEAKER_00: Anyway, so these are the different ones.
• 44:04
  SPEAKER_00: A mid-journey as well as another photo processing.
• 44:07
  SPEAKER_00: You know, if you should all of them and see what you like better.
• 44:10
  SPEAKER_00: And I would just take that same prompt you're asking that got you a good result that you designed with a particular changing of an image or creation of an image, maybe inserting your picture into an image.
• 44:22
  SPEAKER_00: I would use that same prompt, just the same, and go to the different models and plug it in and see what produces your best result.
• 44:30
  SPEAKER_00: And then, you know, after a while, confident, then just maybe continue using that model.
• 44:35
  SPEAKER_00: I mentioned before these systems out there want to get you to stay with them.
• 44:40
  SPEAKER_00: So if you start to switch around, it's not best for them, but I think it's better for us, at least we get a taste of it, until we have some kind of our own AI that we get to interact with.
• 44:50
  SPEAKER_00: And it's not going to be too long.
• 44:51
  SPEAKER_00: We'll actually be able to have our own AI probably on one of our screens, like we're having a Zoom meeting talking with us back and forth, laughing, chuckling.
• 45:00
  SPEAKER_00: And it's it's our own system.
• 45:02
  SPEAKER_00: It would be incredibly weird.
• 45:04
  SPEAKER_00: And already people are developing relationships with AI.
• 45:08
  SPEAKER_00: Then when you have your own and you make it, you know, uh of an of the uh gender that you like and so on, and then you fall in love, and then you guys uh you know go to that, watch out for that.
• 45:19
  SPEAKER_00: And the kids as well, watch out for that too, because a lot of kids are really into this.
• 45:24
  SPEAKER_00: And I mean, they talk nice to you.
• 45:27
  SPEAKER_00: They they'll tell me how I am, they won't tell me like mom did that I have a face for radio and so on.
• 45:32
  SPEAKER_00: All right, the five piece of perfect prompts, purpose, define your goal clearly, precision, be specific.
• 45:38
  SPEAKER_00: When you put your own image attached to a prompt and the thumbnail it's creating for you based in the prompt, you need to give it an idea that you really don't care if they change your image.
• 45:49
  SPEAKER_00: Now, changing it could be putting somebody completely different in.
• 45:53
  SPEAKER_00: So whenever you're prompting, that precision needs to be specifically keep my face, maybe change my clothing, change my expression, but I want to use my face.
• 46:05
  SPEAKER_00: If you have a particular way that you look, you might not want it to change, you know, making you larger or smaller.
• 46:13
  SPEAKER_00: It just depends.
• 46:14
  SPEAKER_00: Parameters specify style, mood, and format.
• 46:18
  SPEAKER_00: You can say, give me an example of style, mood, and format.
• 46:23
  SPEAKER_00: Perspective, define angle and viewpoint, camera at the bottom, camera overhead, camera left, camera right, whatever it may be.
• 46:31
  SPEAKER_00: Viewpoint on a billboard, on the top of a taxi in an advertising uh in a bathroom stall, you know, posted on the drawn-in or whatever.
• 46:40
  SPEAKER_00: I mean, who knows what he and then of course iterate and refine.
• 46:43
  SPEAKER_00: So usually the first, I like that first prompt out thing, if it's right, because I'm publishing so much content about AI and other things.
• 46:51
  SPEAKER_00: I don't wait too long.
• 46:54
  SPEAKER_00: But, you know, the more patience it, maybe that's a good thing.
• 46:58
  SPEAKER_00: But it comes down to the salad.
• 47:00
  SPEAKER_00: The the salad, maybe there's just too many selections and people get paralyzed.
• 47:04
  SPEAKER_00: If you're creating content for your business and you feel like you're paralyzed most of the time because you just can't ship because of so many choices, maybe you need to start limiting your choices so you can ship.
• 47:17
  SPEAKER_00: Talk to people about that as well.
• 47:19
  SPEAKER_00: All right, modern house, of course, that would be a wink prompt.
• 47:22
  SPEAKER_00: Strong prompt would be professional real estate photograph of a modern two-story luxury home at golden hour, wide angle shot, showing manicured front lawn, contemporary architecture, large windows, more material lighting visible, blue sky, few clouds, professional photograph photography style, high resolution, suitable for MLS mess listing.
• 47:41
  SPEAKER_00: When you're doing um modifications to listing photos, as real estate agents, you're gonna have to put that old one in there and you're gonna have to talk about what changes you made.
• 47:52
  SPEAKER_00: I wasn't gonna do a funny spoof on putting a picture of a house and putting maybe the Taj Mahal or some mansion next to it, and saying as long as you put in the disclaimer that you've changed the image, you're good to go.
• 48:02
  SPEAKER_00: But yeah, again, you have to be like that.
• 48:07
  SPEAKER_00: Um, even with the language there that you modified or changed the image, not like the original, and then you have the original image next to it.
• 48:14
  SPEAKER_00: I think going that far out of bounds is probably gonna create a problem.
• 48:18
  SPEAKER_00: So when in doubt, ask your broker.
• 48:21
  SPEAKER_00: Or if you're a broker, ask the owner, and so on.
• 48:25
  SPEAKER_00: So just keep yourself safe out there.
• 48:27
  SPEAKER_00: A lot of people, well, suing is fun, apparently.
• 48:30
  SPEAKER_00: Image modification techniques, in painting, outpainting style.
• 48:35
  SPEAKER_00: This prompting thing, I think that's probably gonna be something that's gonna live in legend and lore at some point.
• 48:42
  SPEAKER_00: And the systems themselves are gonna have a prompt overlay where you ask it something, and then the system's gonna reprompt itself with the correct way it needs to hear the prompt, understanding from your history on the on the system actually like what you're looking for.
• 48:59
  SPEAKER_00: Because whenever you're um iterating, discussing, talking to your own AI, it's gonna learn a lot about you.
• 49:06
  SPEAKER_00: And human beings, I understand conscious and spiritual and all those fancy things, but we're not complicated for something that has, you know, at some point thousands of times exponentially more intelligent than us, it's not gonna be difficult.
• 49:22
  SPEAKER_00: And right now, these systems with all the data they contain, their alignment, their biases, however they're put together, the compute and the power, they they have a good almost everybody.
• 49:33
  SPEAKER_00: And right now, the in the a the you'd have to look into the history between about the IQ test, but that's really the basis that we use, and maybe there'll be something better at some point.
• 49:43
  SPEAKER_00: But on a human-level intelligence IQ scale, um, they're past it.
• 49:47
  SPEAKER_00: They're where most of us happen to be.
• 49:49
  SPEAKER_00: There might be a few outlying humans that are smarter, but you know, you throw enough, enough power and enough chips at these things, and they uh they continue to grow.
• 49:58
  SPEAKER_00: That's not gonna change.
• 49:59
  SPEAKER_00: It's just gonna continually get smarter.
• 50:01
  SPEAKER_00: And then at some point, right now, for the last several years, they've been using AI overlays and all your social media.
• 50:07
  SPEAKER_00: So when you're looking online, there's a reason why you're seeing things that potentially you really like or they they anger you, but like to we like to be angry.
• 50:18
  SPEAKER_00: So we're gonna see this stuff.
• 50:20
  SPEAKER_00: The thumbnail stuff when you see the distressed faces on most of the YouTube channels, that's kind of where um a lot of people like.
• 50:29
  SPEAKER_00: That's why these systems are done in that way.
• 50:32
  SPEAKER_00: That's why these videos have those expressive looks on them.
• 50:37
  SPEAKER_00: So, you know, you know, the shocking and uh, you know, warning and nightmare.
• 50:42
  SPEAKER_00: No, this guy here, great guy, and then so on and so forth.
• 50:45
  SPEAKER_00: But you know, this is the you know, John Wick, of course.
• 50:48
  SPEAKER_00: Um, why wouldn't you have John Wick and a massive Doverman with a cool muscle car on the home page here, and so on.
• 50:56
  SPEAKER_00: So you see this, you know, everybody knows Tony Roy, diary of a CEO guy, if you do, but shock and awe.
• 51:01
  SPEAKER_00: Shock and awe.
• 51:02
  SPEAKER_00: Maybe in some case, that should be your method, depending on your particular service, but these want clicks.
• 51:10
  SPEAKER_00: We want clicks, and that's what people are going for.
• 51:12
  SPEAKER_00: All right, I'm gonna close out.
• 51:14
  SPEAKER_00: If anybody has any questions, please let me know.
• 51:16
  SPEAKER_00: There's just so much with images, but I think I've given you a few good ideas.
• 51:20
  SPEAKER_00: It's really the language that's gonna matter.
• 51:23
  SPEAKER_00: I mentioned prompt overlays when the system itself is going to create a better prompt than what you put in, but you can always ask it at the very beginning.
• 51:32
  SPEAKER_00: You can ask it for that step.
• 51:34
  SPEAKER_00: I want you to create for me the correct prompt, high-ranking YouTube video based on real estate, based on me, Connor MacGyver, real estate agent, Connor McIver, AI Growth Architect, Connor MacIver, Facto Fitness, whatever the hell.
• 51:48
  SPEAKER_00: But put that in there for me, please.
• 51:50
  SPEAKER_00: And then I want that best prompt to be able to take my particular image, generate a thumbnail that's gonna be ranked high using best practices at YouTube, and give me that prompt that would best do that, and then it will spit out the prompt that you should potentially use.
• 52:05
  SPEAKER_00: You look at it, read it, you do that a while in your business, and you're gonna then learn the best practices in making prompts, and you'll just go from there.
• 52:15
  SPEAKER_00: All right, so if you have any questions, let me know.
• 52:18
  SPEAKER_00: Oh, here's the keyboard shortcuts.
• 52:20
  SPEAKER_00: I'll take this particular show, I'll convert it to a web page, and I'll stick it up uh today.
• 52:26
  SPEAKER_00: We'll make this show 01192026.santaclaritaai.com.
• 52:34
  SPEAKER_00: How about that?
• 52:35
  SPEAKER_00: 01192026.santaclaritaai.com.
• 52:40
  SPEAKER_00: I'll put that in there and then you'll be able to go and uh look at this in the future.
• 52:45
  SPEAKER_00: But I'll dump this entire thing in there so you get access to it, and then you know you can look at it.
• 52:51
  SPEAKER_00: We'll go through the other bazillion slides I put up that I never got to because just talk too much.
• 52:56
  SPEAKER_00: Hope everybody's well.
• 52:57
  SPEAKER_00: Thank you for watching.
• 52:58
  SPEAKER_00: If there are no questions, I will bid you good day.
• 53:01
  SPEAKER_00: Good to see you, Andrew, as always.
• 53:03
  SPEAKER_00: And thank you all for watching, whether it's during the show now or later on.
• 53:07
  SPEAKER_00: I appreciate it.
• 53:08
  SPEAKER_00: Let me know if you have any questions.
• 53:09
  SPEAKER_00: If not, I'll kick you out of here.
• 53:11
  SPEAKER_00: Very good, everybody.
• 53:12
  SPEAKER_00: I will see you very soon.
• 53:13
  SPEAKER_00: Thank you for watching and uh be well.
• 53:15
  SPEAKER_00: We'll talk to you in the next one.
• 53:16
  SPEAKER_00: Take care.