AI News of the Week - August 11, 2023

Welcome to AI Decodes the System, the podcast where we delve into the intricate worlds of data, policy, technology (artificial intelligence), and law for everyone to comprehend. I am your host, Amber Ivey, affectionately known as AI. In today's episode, we are thrilled to have a discussion with Marcy Nuelle, a brilliant researcher and renowned YouTuber, leading major faceless YouTube channels and emerging as your definitive source for AI insights. We're positioning ourselves as the go-to authorities in AI. Call us the 'go-to girls of AI.' Yeah, we like that!

Marcy provides valuable insights into the use of AI in creative content, recent updates to popular design tools, and the introduction of new AI models by various companies. She also shares concerns about AI's implications in finance and highlights the disparity in AI accessibility between individuals who use desktops and those who rely on smartphones and laptops.

AI News Of The Week:
There are a lot of updates today so we barely needed the entire week. Topics in order.

AI News of The Day
Stability AI StableCode
StackOverflow Overflow AI
Leonardo AI IOS App
Nvidia AI Nemo
No Code AI
API for Open Source AI
Nvidia DGX Cloud & HuggingFace
Meta Llama 2 Update
Stable Diffusion Audio
AI Voice Cloning
Google IDX
Google Adatape
ZOOM AI Training
Dating in AI World
AI Training Models GPU Usage
S.E.C Gary Gensler AI Will cause Financial Crisis
AI Laws
Cornell AI Keyboard Tracking
AI Educational Divide
Midjourney Vs Leonardo AI PhotoReal

Don't forget to follow Marcy Neulle and AI Decodes the System on YouTube, and stay tuned for more enlightening discussions on the intersection of AI, data, technology, policy, and law.

Guest: Marcy Nuelle

Connect with us:

• AI: Decodes the System: Website, Instagram, Twitter, Facebook, and TikTok
• Marcy Nuelle: YouTube Channel

Support the show:
https://www.patreon.com/Aidecodes

Transcript

[00:00:00] Welcome to AI News of the Week. We are your go-to girls for AI and all topics related to ai. I am here with Marcy Neulle, and I'm Amber Ivey, a k a ai, and we have another crazy week for you in the world of ai. I read a few things. I'm like, what is happening? It's happening so quickly, like when we look at this moment a year from now, and like, look back, it's gonna be wild.

[00:00:25] But Marcy, what do we have on deck to talk about today? Oh man. We have a lot of updates just from today. Not even like the whole week, like today. I have to like shut it down because it, there was just too many. It's too much. Yeah. Alright, so we're gonna start with the fun stuff, like always. So you guys know I love stable Diffusions, one of my favorites.

[00:00:45] And stability ai released stable code, right? So now we're getting back to task related. AI software. I've been talking about this from the very beginning, like months ago. I try to tell [00:01:00] people it's gonna get task related because now we're gonna start going to the professional level and at professional levels.

[00:01:06] It's not generic it's always going to start going towards very specific task related software. Like what does that mean for the normal person? Okay, so for example, like with chat, G B T, . You know, people are trying to use plugins to make pictures and mind maps and ask questions, and it's like, no, once you get to like the university and then corporate level .

[00:01:25] It's going to be task specific because then the software can optimize for better results and accuracy. You're not gonna have as many mistakes, so like that's the purpose of it. Yeah, of course. I don't know what happened today, but it seems everybody wanted to drop their code today because there's a few of them.

[00:01:41] Okay. Like even just a few days ago with Stack Overflow, they had overflow ai. I just posted about that today too. . And it's like, it's, it's gonna start happening people. So next step we have with Leonardo ai, which is like a text to image app and they drop their iOS [00:02:00] app. It again, talked about them last week.

[00:02:04] Yeah. And now we're going back to more companies trying to do what I've told you guys. What would be the big business that's coming out? Server side hosting for individuals who don't have desktops. . Prepare to start paying a lot for different subscriptions. People I've been trying to tell you, you're gonna be able, we have 20 streaming services all on our iOS if you don't have the exactly.

[00:02:30] So they need to. So please also make sure it's the right ai tool, because a lot of folks will drop things that use language when you search. So please make sure you're downloading the right things if you are on iOS or a Android. Yeah, that's a great tip. Then next up we have Nvidia with their ai, it's called nemo.

[00:02:50] And it's to build the l l m, which is, you know, large language model, . At scale from anywhere. . I think they were focusing on enterprise companies. [00:03:00] So once again, I'm trying to tell you guys now they're moving away just from the general population. They're moving towards corporate, they're moving towards large corporations to implement that within their business.

[00:03:11] Why does that matter? At a level of first we OpenAI, like not first, but like in a way where people could understand OpenAI released something to the masses.

[00:03:21] Anyone can get access to it at a point. Granted, before that, researchers were having it and some limited folks, they released it to everyone. Now all these companies are going trying to figure out their play. Why does that matter? . Alright, so the reason that it's so important, let's say your grandmother doesn't know ai. . She's going to have to start understanding how to then update different systems. Yeah, update different process on their app or whatever company it is that they're using.

[00:03:52] Once companies at scale start implementing these AI softwares, it's gonna be the trickle down effect, meaning. [00:04:00] Your grandmother, your aunts, your non-tech people, they're going to have to adapt. They're gonna have to learn. And of course for me, because education is extremely important, right? The educational divide is gonna get bigger because individuals who are not aware of how to use D softwares when they're gonna have no choice but to use them.

[00:04:20] . It is gonna be an issue. And by the way, That's another opportunity for a business that's going to also happen. I was gonna talk about that another day, but I might as well say it now. Go for it. Is that now we're gonna have, instead of marketing agencies, you know, with all these fake experts anyway, we're gonna have these companies that are gonna say, oh, we are AI implementation companies.

[00:04:41] We're gonna help you adapt to these so different. . And we're gonna see more and more of that because people are gonna be confused, right? Like, how do I all this information. So now we're going to see the trickle down effect. So that's why it's important and then we're gonna start seeing all the other companies, you know, on even entrepreneurs.

[00:04:59] Yeah, [00:05:00] and I'm sorry, you need to know how to code you need to know how to code because you're gonna get left behind. Point blank period. You're gonna get left behind because guess what? You are waiting for people to tell you. Oh, okay. You can use this point and click software. Oh, I wish somebody would create this.

[00:05:15] You mean what we did six months ago? It's already there. If you would pick up a book, if you will pick up a tutorial, a free course by m i t, by Google, or hire somebody who's educated in this field. All your wildest dreams about these creation that you wanna add to your company that already exist. You could have it right now.

[00:05:37] You don't have to wait. And there's enough libraries and people have done a lot of work. So learn the basics. Understand how to access all this stuff is very important and you're not gonna be a market leader. I'm gonna put it bluntly. Women are going to be the most affected. If you look up research, we're talking about between 40 to 80% of women who are going to be impacted as far as their job role. If you are going to [00:06:00] believe these so-called experts are tech girl, you don't need to know how to code. It's all good.

[00:06:04] Just point and click. . You're not gonna have a job. Your business is not gonna thrive. You're not gonna be a market leader, and sorry, you're not gonna be able to compete against who already use AI to automate everything. You're gonna get left behind 

[00:06:15] It's not too hard. You don't need to everything. APIs. . 'cause if you get a key, then you can download certain software.

[00:06:24] It's literally like put in a password. You upload a little code to your website, boom. It gives you access. Different things. You can copy and paste certain code. It's very simple. I probably need to make some tutorials for you guys

[00:06:34] I'm hearing this as a person who even likes tech and I'm overwhelmed, like, oh my God, I gotta do this. What, what am I supposed to do? I'm sorry. Walk me through step one, what are like the basics? Someone should know? Are you talking about different coding languages? Should I know Python? Should I go to C?

[00:06:50] Like, what does it look like? Like, where should I start today? I've never touched a code other than I may know how to change the color of my text. What did the person do? I would say [00:07:00] three very basic things. Let's say you have. A basic WordPress website, okay? . . That's like step one.

[00:07:06] You need to have some sort of website where you have like F T P access. Okay. And you can actually do, you're not gonna be able to use some sites like Wix, some sites, what's the other one? One. None of those they don't have. You can get developer keys for them, but they're not built that way. You need something where you can have F T P file transfer protocol.

[00:07:24] That's the backend where you can actually have some access to how things are run on the backend. You need. That's like number one, because people are like, well, I'm just gonna go get the, sorry, it ain't gonna work there. They don't even give you access. . So you just need something with a website where you have access on the backend.

[00:07:40] Okay. And then you can just say, Hey, I'm gonna learn how to use an a p I. They have free courses on like Udemy. . You can go to Coursera, I mean, there's people on YouTube for free. Like, yeah, just learn how to use an api, basically. Like download a piece of code. What is an api? I'll quick version it. It allows you to transfer data or information from one place to [00:08:00] another. Using an integration. Like, that's the simplest way to explain it. Way better, because I was gonna go a different direction. No, that's why I was, lemme do the simple, simple version go a little bit deeper.

[00:08:12] You was bringing that your tech background. I'm like, you know, I'm playing language real quick. Okay. I appreciate that because I was about to say, well, yeah, this is what it's, but yeah, yeah. So, okay. So yeah, it's just basically connecting, like she said. So once we do that, then you just learn a few little python scripts.

[00:08:30] You can use p h P code to start doing like little copy and paste codes and start that way. It doesn't have to be super complicated, but for example, let's say like with an open source software, right? And they say, Hey, you can show people, you can copy this code and then people can make text to image on your website.

[00:08:51] Yeah. You see, it's something simple like that. It's not like it's, oh, we're doing crazy, weird things. Let's say you are running a t-shirt company. Let's, let's put a real [00:09:00] example, right? And now you want people to create the designs on your site, right?

[00:09:04] And then you can use like an a p i for a company that has open source for text to image file. And then now you can offer that service on your, on your website as long as it has commercial use. And then you can go ahead and offer that service and then people can say, Hey, sign up for a membership. Or you can use that, right?

[00:09:22] You commoditized be part of my community, or . Or things like that. And that's another way to increase your revenue. That's another way to make people say, Hey if I'm selling a course, if I'm selling an ebook, I also have the ability to create the images directly from you. That's how you add value to people.

[00:09:39] . So when people say, oh, well no, you have to go to chat gpt. It's like, no, you can come to my website. Even with the same thing the they offer. Exactly. So it's like, that's what, that's how you offer value, that's how you're gonna compete. . And people are like, because they think they're have, they have to make things from scratch, and it's like, no.

[00:09:57] So teaching their [00:10:00] software that they have and then you're using it on your site, it's, and then adding it, making it work for you. Exactly. You don't. Exactly. And if you don't know how to code, you can hire people. You know, you can, but you have to realize, you have to prepare now for it. You know, let's say someone doesn't know how to code in your company, so even though you, and I'm glad you said that, even though you may never touch a code having a expert on your team who knows how to do that, who can do it for you, that's akin to the same thing.

[00:10:30] . Because it's gonna get more competitive, right? Like you may be at the top of your business right now. What Marcy's saying that in order for you to keep competing in the new world, you gotta take it to the next level. Before no code could work up to a point. But now with these advanced learning models and AI models, you gotta be able to integrate that stuff and they're giving a lot of this stuff away for free and giving you the a p I or if you're paying a certain amount to use a certain amount of A certain amount of information that comes across, you have to pay for a certain amount of it, but they are giving you access to all the stuff that they've done [00:11:00] before.

[00:11:00] Yeah. So why not use it to make your company go to the next level? Think about if you owned a, like Marcy was saying, around a t-shirt company. If you own companies that create furniture or text off or the furniture, you could have people create all their stuff on your site and take your stuff to another level where you literally do on demand printing for folks and all that cool stuff using these ai.

[00:11:22] Tools. So we're trying to help you understand. I made a video about that, about the like you said, textile. Yeah. I made a video about that like three, four months ago. On mid Journey you create patterns. Yeah. In one minute I went and created an entire leggings company. Yep. I can make 3000 leggings in one month.

[00:11:43] It's like, and then people buy the stuff. Now that's a big thing. Like I was. I'm bringing my book out another week. But no, this picture is from Mid Journey. Like that's what inspired my book. That's not, I didn't create this. I played around with Mid Journey to see like how I wanted the characters to [00:12:00] look. And that will be like a future version of the character. But at first the reason, honestly why I got on, 'cause I was gonna do that with t-shirts and I was like, oh, I could just set it up and have these t-shirts that are customized and then get them printed and then just make some quick money on the side.

[00:12:15] Like I was thinking about that as well, like AI generated t-shirt. So there's ways to do this for sure.

[00:12:19] So make sure you check out her YouTube 'cause a lot of stuff is there and she'll go deeper in it even though AI news the week will hit the topics very quickly. Check out her too for all the details. Yep. And then the next one, which is kind of connected to this is Nvidia also. They dropped the the D G X Cloud and they partnered with hugging face which the found the foundation of that is to customize generative AI models.

[00:12:45] You don't have to customize it yourself. They're literally like, oh, you want the software? And then you can also have it customized. So it's like literally everything that you want and you desire is already created by the biggest [00:13:00] companies on the planet. Yeah. And they're putting out there.

[00:13:02] Let me say one thing before people get on because no code does matter. No, I would say this. Look, you can make millions of dollars with no cost. You sure can if you already are aware of how to use certain things. 'cause I'm not one of those people that are like, you're gonna make 10 grand in one month.

[00:13:19] Okay, liars. That's not the truth. Because the truth is that you still have to have some sort of understanding of how things function, okay? . You have to have an audience or money to run ads in order to then sell.

[00:13:34] There's always going to be a pro and a con, and you have to be doing it right now, I'm sorry to tell you, you think you're gonna be doing this in a year? This is one day people. Okay. This is one day update. There's gonna be like the whole list. There's 4,000 updates. Every day is gonna be coming out.

[00:13:55] Yeah. Every day, if not every few minutes. [00:14:00] And then companies like Nvidia, 'cause they create these chips that allow all this stuff to happen. They're producing, also producing stuff. There's also other companies that are putting things out that are in the space. So if, even if you just search like AI companies and how quick a lot of them started in the last three less than three years.

[00:14:17] Like they're quick. A lot of them even started last year. Many of them also started after their chat. G P T was released, right? . Because a lot of people even created companies off of chat. GPTs a p I like. Thousands of thousands, if not tens, a thousands of companies have been created doing what Marcy just said, like using their a p I to create another 

[00:14:35] use case, and that's gonna be even more. Shoot, there's AI already running models for finances. I was reading an article that said AI had a 500% return on stock trading, 500%. Someone took the, took the model and was able to apply it in a financial setting. So imagine that like I was like, lemme go ahead and get this set up that that connects to what we're gonna talk [00:15:00] about later.

[00:15:01] Yeah. And before, before, honestly, before government starts regulating a lot of this stuff, a lot of people are making money off these tools in very unique ways. 'cause there's billions of people that live on the planets of customers that you can always reach out to. But what is on our list?

[00:15:19] All right. Then with the next update is for, oh, for meta, they updated lama to reduce false refusal rates. And that just means that they're trying to manage the fake users, right? So they're trying to see who is legit and who isn't. . And why is that important? Because they're using bots in order to get information, in order to scrape more data and.

[00:15:44] That actually can reduce the system as well, but that's another issue. But so trying to manage legit users. . Actually for, it has something to do with privacy too, but that's, that's a topic for another day. Okay. So that's like a, a really important issue and I'm glad they do have that [00:16:00] update.

[00:16:00] Next up we have stable diffusion now has audio. So, I mean, shoot. Nice. I gotta go ahead and get That's, it's a lot. Yeah. I mean it's great, but I'm excited about that. It's great. It's a lot. We still need more laws, I mean, for that because Yeah. We need regulations and I'm like, I'm the type of person that's all about companies growing and us being innovative.

[00:16:21] But I'm also like, it's so much happening. We need copyright laws and protection laws that match the current state and the future state of where this is going. Because a lot people are gonna get messed up even So that picture I showed, if you actually watched this on YouTube and Marcy's channel the picture I showed, the funny thing was I did an interview for N P R a couple weeks ago for the book or whatever.

[00:16:40] . And the lady was like, oh, that image looks like an image I've seen before. And I was like, it probably is. And I told her, I was like, it probably is 'cause they're probably using the metadata and my words are gonna be similar and it's scraped from that image to build the image. And like legit, she knew the author.

[00:16:54] So I told the author and we, she was like, oh my God, it does look like, and it could have easily trained off of her, [00:17:00] but how do you then protect someone who has their image out there, the metadata matches and the AI scraped from it and create it using that image. 'cause that's what it's doing, right? So how do you protect people when we don't have laws that talk about this, you have laws about copyrights, but then does it cover an AI taking your image and changing it?

[00:17:19] Yeah, where it's a different character, but there's a slight resemblance, so it's very interesting. And then even with music, like you're saying, audio, You know what it reminds me of, right? It reminds me of the Little Mermaid in Ursula. Oh. When she stole her voice and she was her. Sometimes it be like that.

[00:17:39] Yeah. I mean, she had her sign the contract and it's like, I always say like, what is it gonna cost you? Is gonna cost you your voice? Yeah. It's your likeness. It i's gonna, it's, it's like, it's like we don't have something to protect our face, to protect our voice. There's nothing that protects our likeness.

[00:17:55] Unless you're famous and you've gotten things protected and you're brand or whatever, it's totally [00:18:00] different. But most of us do not have anything to protect our license, and we always throw our faces in those ai apps and all that stuff to train this stuff. Yeah. To Facebook alone. And your, your phone alone trains on your face every day.

[00:18:12] So, yeah, I said I should create an ai, call it Ursula ai, all these contracts, everybody's signing their souls and lives away. I'd be like, it's gonna be for lawyers. At least she got, at least she got the prints at the end. So it all worked out for her. But it don't work out. We are not fairytales. It's not gonna work out that well for us.

[00:18:32] So I protect your data, protect your your information. Can I say something really quickly about kids in this? I saw a really cool video that was basically showing, and sorry parents, I'm not trying to knock y'all 'cause those are your kids, but it was a video that was basically showing it, it's like a parents come into a movie theater.

[00:18:49] . And they see their daughter as a kid and they flash pictures. Then they show their daughter as a future ai and she's like, mom and dad, I love you, but all those images you put [00:19:00] online, someone has now taken my identity and created likenesses of me. And have created fake images of me because all these pictures were posted when I was a kid.

[00:19:09] Yeah. That is a real threat. It's a threat for adults, but for kids, yeah. We just definitely gotta be careful on what we share because voices can be picked up very quickly. I think we talked about last time, but I've trained my voice on my ai The last time it was trained, I think it was trained on literally five minutes before that was trained on four four hours.

[00:19:26] Wow. I think even now you could train on less than 30 seconds. So, If you're sharing clips of things, please be aware that the future of this could be, and folks, we've seen folks doing this calling folks using your voice. . To get to your parents, to have them send money or say you're kidnapped. Like there's so much stuff that's gonna happen that we don't even know yet.

[00:19:42] So I just wanna remind people, especially if you have kids, we talking about a Disney movie, a kids movie. So especially if you have kids, just be careful. 'cause that is a reality that they're gonna have to face and something else they have to worry about that we. You're getting older and we're gonna die off, so we don't have to worry about it as soon as they do.

[00:19:56] Sorry. Anyways, what else? What else is on the list?[00:20:00] . Okay, so the next one is that Google released I D X, which is also a browser-based coding environment. . So I told you guys this week, they're all dropping.

[00:20:12] The coding tools that are specific for that, because now we're moving towards accuracy, right? Like the novelty of Yeah, it got go an answer like, okay, great. You know, like when I try, like if you're a professional, most of the time you, you really can't use a lot of these thought because they're not, they're gonna be screwed up and it's like, review everything and then research it again and verify it's extra time.

[00:20:35] I might as well do it directly from my database. Yeah, point and just have it, maybe edit the language or clean it up. But you, you need to do the work. It's like that lawyer who got caught up using that AI to give it fake case law. It's like, I saw that that was of all things. You could have searched that.

[00:20:53] But anyways, I'm like, you don't have an assistant said, let's verify. Yeah. All right, so [00:21:00] next up we have Google Adapt which is an adaptive computation model. For input sequences, basically it's for machine learning.

[00:21:09] . And it handles new data, so that's important. So for example, you know let's use chat, g p t in their plugins. The reason that they need plugins is because maybe different software can handle different information in a different way. You know, it's like multimodal, like a multimodal system. But with an adaptive computation model, it's able to say, okay, this information is this, and now we're able to create different feedback for you.

[00:21:31] Okay. So that's a really important feature that they created. Then the next topic it happened like what I think, was it yesterday or day before yesterday? With Zoom. I also, was that yesterday? I can't remember this week. Yeah, yesterday. Was the week. Yesterday or day before yesterday. It was. We just thought about it, but I can't remember what day that was.

[00:21:52] Okay. Yeah. So just to give a quick recap, so basically everybody was kind of going crazy because Zoom updated their [00:22:00] terms. . And they. People were saying, okay, zoom is training their AI system with your data, right? Yep. And then everybody went crazy immediately. Like, what do you mean you're training?

[00:22:10] Right? So then Zoom immediately created this blog post. They were like, no, we're not training the ai. What are you training? We're not training. Don't sue me. Zoom. Okay.

[00:22:25] Without your consent. And I'm like, wait a minute. If they sent new TSDs, were in there. It's, you can't do anything about that. If you keep using the technology, you've signed up for it. This is the issue for the people, right? We're not talking about them. The companies are, I do it through my company and I've done it through my own, I have my own license, but I also have company licenses like, But the issue is, is that how that the article is not always as clear as, okay, well, How are you acquiring the consent?

[00:22:56] Are you using a popup to make people more aware that [00:23:00] you are training? No, that's not the software versus using the feature. Because I think that's the point of confusion of why so many people were upset. . Because they're thinking, okay, if I'm saying I'm consenting to using your AI software, I'm thinking I'm just using the feature.

[00:23:14] Oh, it's a cool, right. Not the training part doesn't, they don't think, they don't even know what training means. They might not even no ones conditions. I, I really believe in active and like known consent because even in the legal field, this is something that comes up is that people are saying, oh, they can sense it.

[00:23:33] You can't have true consent if you don't understand it. And I know there's been a movement to get to plain language. Yes. If you do this, I will take your face. Yeah. And your information and train this tool on that and therefore, use your image, use your words, use your voice, use everything. They have to be very clear.

[00:23:52] Companies will not do that though. Yeah. And then the third part of it that actually didn't see many people even mentioning this . Right, [00:24:00] is actually the G P U. Usage of training the AI model. So I talk about this last week too. This frustrates me with people not understanding why you need desktops or why you need different computers.

[00:24:14] Explain why do we need desktops? Okay? Because a lot of this software, I know with server side or cloud hosted AI software, those companies are typically hosting that software on their computer, right? Right. But guess what, once you're using, let's say on Zoom, you're on your laptop. And then all of a sudden, It slows down and it starts flickering now, and you're like, what's going on?

[00:24:38] It's a bad day. It's not raining. What's going? And then in the back end it's like, oh, we're training. We're taking all your power. You know? It's like, oh, okay. Like what does that mean? It's like, oh, we're taking all your power. You thought you're gonna get good connection today. No, you're not. You know, so it's like, 'cause a laptop does not have the same [00:25:00] capacity as a desktop when it comes to .

[00:25:02] The amount of data captured that can be stored and transferred across the internet. So yeah, that's what Marcy always talks about. . This stuff, some of this stuff actually requires you to have a desktop by virtue of like having to have a cart or, or what do you call 'em, things. Graphic card is big enough to .

[00:25:21] Even be able to take any of this stuff, even the quality of your monitor is different because of the internal components in your computer. . So that's something just to think about and know. And then most of us have laptops and smartphones. We don't, yeah. We don't have the computing power of a desktop.

[00:25:37] A lot of us went away from desktop 'cause everyone told us. You could take your life everywhere, but those who are actually doing this work do have a desktop. Marcy's on a desktop. I'm not on a desktop, I'm on. You better fall in love with your your gaming boyfriend now because he has to hook up and he don't.

[00:25:54] Yes. My boyfriend, he just, he bought a desktop and he just updated it 'cause it wasn't the [00:26:00] power he wanted and he also trained it and like got it to the point of like, you can train and tinker with these models to get it to the high quality. So, but people don't know that stuff. So like the everyday person is me, 90% of humans do you see smartphones, right?

[00:26:16] And then the rest of folks may have a laptop, then I then a tablet of some sort. But to do the type of stuff you wanna do, you have to get a desktop. And some people can build their own desktops. I know it sound weird and sounds crazy, but a lot of people do build their own desktops or buy a desktop and then edit the parts out.

[00:26:34] You can literally sub the parts out and buy parts. It's like, Hey, tech bros. Hey, I cannot, oh, I think that's gonna be like the future flirtation thing. Like, you got your own test, you got your own test. Holler at me. Did you build it? You have a cool system. System. It's like, okay, cooling system. Okay. Right.

[00:26:55] Literally, I got three fans. What?[00:27:00] 

[00:27:03] Right. I cannot, you have money with Nvidia? You can. No, but it's a lot of money. But lemme find. But this is another point. Talk about the side of, like, back to what Marcy said earlier about the gap, if you can't afford that, That's a whole, and most people, honestly, I think the average salary right now for most Americans is around 60 something thousand dollars.

[00:27:26] . Most people who make $60,000 and also have a family are not buying and building a computer that costs 2000. They're just not doing that, and because of covid, the price of technology. Components went up. They're down now because stuff is caught up. But during covid to build a computer was ridiculous.

[00:27:42] And right now I think the cost to build or buy is around the same based on what I know when Louis was doing his looking at his stuff, that's what he said, but yeah. So it's hard because folks do not have a way to do that. So in a future episode, fix Your Face. In a future episode, I do wanna [00:28:00] talk about how does a person who don't have the means.

[00:28:04] Figure out how to make beans and we can actually, we may talk about that tonight. I don't know. Keep going and we'll come back to it. I was, I was gonna talk about dating the tech Bross again, like, oh God. That's why I stopped. It's not the only way. Stop yourself, tech bros. That's her. That's not ai. Live y'all life.

[00:28:23] I got a tech. Put it in your guideline. Put I can, I can afford it. A desktop, right? You're gonna get more girls. Put in your tender. Can you afford a desktop for the future of our lives? Okay. And then the next part to this but imagine if now you have 10 different companies training. Laptop. I got room for that.

[00:28:51] Just being honest, like you have, you don't have room for all that on your computer. Like you can't do who, who's gonna get, who's gonna get a clear per a clear vision on their video? [00:29:00] Nobody. No. It's gonna be like, like, it's gonna be like my video right now. I mean, Android people, low level. I mean, I don't know about you iPhone people.

[00:29:10] You know, please don't do that. That's not appropriate. Android people, you know, will be good. I mean, y'all will probably have, I do admit that y'all will get us. We will not. We'll be good. We're okay with that. We've settled. We've settled that. We're good. Yeah. Yeah. But it's a lot because, and then you're not gonna know why, and imagine if you're not even a tech person, you don't know how to check for that.

[00:29:31] . Like most people don't even know what the task manager is. They don't know how to check the G P U usage. That's the thing I learned about 20 some years ago. Yeah. That like could throw delete or is that it? I forgot what it was. I forgot the commands. It's like I always have to clean mine up.

[00:29:47] Can you imagine your grandma, like, I don't know why this is not working. I mean, even now, but that is also a business or industry even now, like I've known people who have taught and taught people who are in a different generation, like gen, [00:30:00] more black baby boomers than anything. They have courses to teach them how to use smartphones.

[00:30:04] And like in-person experiences because all the phones are moving as smart, so they had to learn how to use it. So people were actually teaching people that and getting paid for that. Yeah. Or setting up their computers. People will still buy desktops, especially depending on generation. We're still buying desktops.

[00:30:18] They need someone to help set it up with all the new stuff. Yeah. They can hire someone to bring or to come in and set that up for them. So that is an actual thing that will need to still happen. Like populations will age new technologies will come out. So people who train people on those things and who are 

[00:30:32] technically savvy can make money in that. Yeah, but I mean, but this is important as far as laws. That's why I said that the thing with Zoom is so important because having it in your terms is not gonna be good enough. Because now when all these companies are trying to train on your system, and that's a lot, you can't even do your work.

[00:30:51] It's, and people don't know why it's gonna be a problem. That's a violation of my own system now at that point. Right. And the amount of data that's [00:31:00] coming through there, exactly what we were talking about the other day, like around the data paying people for their data and stuff like they're, that even makes me know more.

[00:31:08] They're gonna have to pay people for that. Yeah. So it's, it's, that's a serious issue and I think people don't realize how big it is, you know as far as that in the future. Yeah. And then one of the it's gonna be everywhere. Everything. We try to do something. I mean, think about the places we know now that are training our data that we already kind of agreed to.

[00:31:25] Like, yeah. Facebook to the world, everything dealing with marketing, Instagram, Twitter, TikTok, like all them. We've given access. Now imagine everything you sign up for is training on your data. I'm gonna need to cut government. So y'all need to figure that out. I'm percentage, these companies are making trillions of dollars .

[00:31:43] Off of my, my me giving them something, even if it's a few cents, a few whatever. That counts and adds up. Like they're gonna have to figure that out, especially when we think about jobs in the future. But I'm gonna stop. Go ahead. It was, it was kind of like, like Amazon with the Ring app where they [00:32:00] were interconnected in order to I think provide the, the internet service or something.

[00:32:05] Was it the internet service or they, they wanted to provide a, if you're talking about this, they wanted to provide. Connecting all the cameras so they can get more view. I did not do that 'cause that's tutoring darn much. And I already felt like it was too much already in my house. So, but they wanted you to give access so you could share data across different Right different accounts so my neighbor could see, or police could see.

[00:32:28] My Camera. I'm not doing that. I'm not dating. Yeah. It's a lot. I'm a little bit in the past when it comes to stuff like that. Alright, so the next up we're gonna talk about the s e c Oh yeah. Which is the, which is the US Securities and Exchange Commission. . And we're gonna talk about the chair Gary Gensler.

[00:32:45] And he recently made a statement that he believes that AI is going to cause a financial crisis. Once again, I've talked about this four months ago, but y'all don't keep watching my videos. [00:33:00] Clearly. She wants y'all to watch her YouTube. Please watch this girl YouTube so we can stop saying this, but go ahead.

[00:33:05] I'm just saying because I've talked about it seems like I, I. I have a good sense of foreshadowing. Don't even know to look for. So that's why it's important to come on or to do this, to podcast it, to put it on Instagram. Like people gotta know what's happening. Most folks, you live the world. I kind of live in it.

[00:33:21] I'm like I'm on a street next door to it. Like I don't even live as much as you do. But it's good to have these conversations. So now people will know like, what did you just say? Financial? 

[00:33:29] So his quote was, well, obviously he said that there could be a financial crisis due to a, due to ai.

[00:33:37] But he also mentioned He said that he believes that it could promote hurting with individual actors making similar decisions. And I'm like, let's do a throwback to GameStop. What's the word you said? You said hurting or hurting? No. Hurting. Okay. Making sure you're hurt you. Right, right. . So I'm like, think about GameStop.

[00:33:57] Right. But now [00:34:00] what happened was GameStop Yeah. And then shorting the stop. Imagine using ai. I mean, right now they're making 501 AI made 500%. Exactly. Imagine if you were to able to do that and, and like some people can make a whole bunch of money, or you can use that AI to manipulate the system when it comes to stock.

[00:34:18] . You can send out either a fake. Elon a fake, someone who can say something and change the world. You can steal their voice. You can do, you can do a whole bunch of stuff to literally make the market go. I totally see that happening. Yeah, and I think the key word that he used is individual actors, and this is where we go back to education, the, the divide as far as resources.

[00:34:41] . Meaning if you have the desktops, if you have the knowledge, if you have the coding abilities, if you have, . The community, you know, yeah. People around you who can all work together as well. You're gonna have people who know how to advance [00:35:00] beyond what most people have the capacity to do with ai.

[00:35:02] And that's why the next quote he said, he says he wants to make sure that the s e C also promotes competitive efficient market markets and transparency and fair access. But the truth is, There is no fair access because majority of people have no idea what's going on right now. They don't have even the proper resources.

[00:35:26] Like I think a lot of people online who are H to a I we're kind of like stuck in a bubble because we're thinking, oh, we know all this information. Everybody must know. We're like in a very 1%. No. You know, most people are not thinking about this stuff. Don't gt. Yeah, they got a hundred million plus users, but I think.

[00:35:45] Tech people have to understand that the average person is just living their life. Yeah. And they see these things as novelties, as fun things to interact with. They're not seeing, I mean, I think more people are awakening to it [00:36:00] because of chat gcpt, but been talking about this for a few years as well. And it's just something where people to this day are still like, oh, I didn't realize.

[00:36:08] Right. And they're still gonna be that way because there is a small percentage of y'all who truly understand what's happening. So I, I think that's very important for us all to remember. Yeah, and I think the last thing he said, the foundational thing that the s e C is focusing on is how to solve the problem.

[00:36:23] Primarily where he said brokers would put their interest over the investor's interest because they're going to focus on making, make percent, making the most amount of money. Yeah. So that has to come with laws, policies, regulations, but the truth is all this stuff is gonna be. Implement it already.

[00:36:43] Yeah. Then it's like the laws, they're gonna have to try to catch up. Look how long it takes for people to vote, and even when they are speaking about this stuff, I mean, I don't even wanna talk about the hearings because these people sound like they have no idea what they're talking about. They don't, so how are they gonna [00:37:00] properly even.

[00:37:01] Vote or even come to a helpful conclusion, you hope that the people who are working with them and doing the research and the people who are testifying bring enough information, but I still feel like they're way too late on this. I think by the time of the year, if they don't have something in place, it's, it's already, I already felt like we're already out the back.

[00:37:19] But if they don't put some laws in place by the end of the year, I'm really concerned. I'm not saying very, I'm not saying be crazy about it. I'm saying like there should be guidelines. I'm like, if China did it, and China's pretty like tech, like their advance when it comes to technology. If you look at what they're doing with their kids and AI and learning and teaching and all that stuff, and what they're doing with their social systems and getting tickets, their AI softwares.

[00:37:44] It's far more advanced than ours. I'm just gonna call ip. Right. I'm saying if they're doing it, we have to also look at, if they're doing it as someone, a group or country that's very advanced and wants to push the limits to it, we need something too. Even if it's just [00:38:00] guardrails and guidelines around how to use it.

[00:38:02] Like even I'm, I'm still upset that we don't have data laws in this country like in the uk they have laws around how you use data for people. 'cause even in that article you're talking about, they were talking about data ownership. And what happens then when these major platforms benefit?

[00:38:16] 'cause they know how to use ai, they can manipulate systems to use ai. If you come the top, then what, what do you do? Like, who owns that data? Do we get a percentage? I'm gonna keep saying that. I'm gonna yang that to death because.

[00:38:32] And you're gonna have to keep a population not fighting each other and, and not hungry. Like you cannot take it. Make people have to be able to contribute to life, whether it's through ai, content creation, wherever our jobs are in the future, they have to figure that out. And if we don't do that by end of the year, I would be very concerned.

[00:38:50] But like to say two things Like I also, yesterday I also posted about the proposed AI Act by the European Union. . Where they want to [00:39:00] control, you know, how AI is actually implemented. Right. But in the video that I posted, it's like, okay, now I imagine running your business and now you have to deal with 10 different unions with all different types of laws law.

[00:39:12] Are you even gonna be able to match up? Because I think one of the laws they proposed is like, oh, you have to. Have certain disclaimers when you use AI software and it's like, or you may not be able to, you know, use certain AI models or may, you know, because cultural acceptance of different behaviors.

[00:39:33] They may say, well, we don't want, but I appreciate, I appreciate that because even remember when they did the privacy updates with the cookies and all that in the uk? . The, we had to adapt and US countries had to adapt because they operate also in the uk. So we also got the benefit in the US from what they were doing.

[00:39:49] But that is a little bit different when that to me should be a global, and I know some people have been talking about what it means to do a global conversation around AI limits and ai. I got roles. But when they did that, I [00:40:00] was happy that I at least got some data and I had active, active prompt that said, do I want to be tracked?

[00:40:07] Always whatever cookies, and you actually have to do it on every single website. So that came because the UK put that in place, not because of us companies wanted to do it. They want, they did it because it's too much work to try to do two things for two different places. Yeah. The issue is that it can limit innovation though.

[00:40:26] No, that's my only concern. Why I think it needs to be at a global level when it comes to that, because I know we're global. Besides that, Who's gonna decide to say, Hey, no, they all come together. No, I'm saying like, kind of like think about they already have groups that come together around this stuff.

[00:40:39] Like they talk about world issues. Granted, people are missing from those world groups. A lot of people are missing from those groups. Only mainly countries that are Western that sit there. But within that, They need to figure that out at a, at a different level. And it shouldn't leave out countries 

[00:40:57] but the truth is also how [00:41:00] many months, like, let me ask you that, since that's your industry, how many months does it usually take to? So why some, it's not happening by the end of this year. No, no, no. That's what I'm saying. But what is a usual time period to create a law on like one subject? There is no one subject law.

[00:41:18] No. Like, because if you're looking at, if you're looking at the federal level most, that's one thing I actually, when I was like interested in, in doing that . I'm like, I believe law should be on one subject. . Because then you can pass a one subject law. Most laws at the federal level, it's.

[00:41:32] Hundreds, if not thousands of things in that law. . So you literally are saying Yes, if I say yes to someone, AI people wanna put their stuff in there because it's leverage. Yeah, exactly. So if you say yes, then all this stuff has happened. That takes a very long time. A lot of laws, if you look at like how long, no, I'm saying it depends.

[00:41:50] It could be a law could take 20 years to pass. Like I'm not trying be funny. Depending on the law, it may not make it through. What would be like. The shortest and like during, say this, [00:42:00] during 26, and I may get my numbers wrong, but during around the 2020 time period, I think the success of loss in Congress was under 10% of them getting passed.

[00:42:12] And I think when I looked at the numbers, I could be wrong because this has been three years ago, but it was very low. . I actually did ran the data or ran the numbers on it because I was really trying to figure out, I think it was like a single digit number, but I don't wanna be wrong, but I think it was, I pretty sure it's under 20%, but it was very low to amount of loss that passed.

[00:42:29] Okay. So that's what I'm saying. And that's every, so now that's just at a federal level. Now you go down to state level, they have laws that they govern, they have. County level, all that stuff. And depending on where you are, like this should be a federal law because it's dealing with, in my mind, commerce and all that stuff.

[00:42:45] But it's so many layers. It is the feds have to do something about it. And honestly, at global level we have to talk about it, but it's hard to pass laws so then and pass a law about something people dunno about. So you have to educate. Let's [00:43:00] do a fake case study, right?

[00:43:02] Let's say a year. I just wanna get, it's not gonna be, no, I'm just making up a number. Oh, okay. Let's say it was a year to pass a law, right? Executive orders can happen quicker if there's something, but then that's hard because you gotta see if it's something. He even has powers to do, like at the presidential level, but it's hard.

[00:43:20] But yeah, the issue is we have like 20 updates per day. That's what I said. Even if something get done, the end of the year is too late and they're not gonna get something done the end of the year, which is sad. So it's like if we have me, I'm open to be surprised. So it's like, it's a lot of updates, a lot of changes.

[00:43:39] The laws are not gonna be able to catch up right now. It's insane. Oh no, no. They're already behind. Last week they were behind December, 2022. They were behind December 20. So we're literally talking what we know is that talk about two other things. Right. I know you wanted to talk about Oh yeah. I wanted to talk about [00:44:00] Cornell release a study.

[00:44:01] I did a an Instagram, TikTok or whatever. I did a few videos on as well, on YouTube as well. Cornell used release a study where they basically took different types of keyboards. . And they trained the AI on the keyboard sounds for each key. . So it could identify what you were typing into the computer.

[00:44:20] So for example if I have a right, right here, I have a Logitech. They took a Logitech, had someone type all the different, and trained it on all the different numbers, keys and everything that were there. And they were basically able to, through you typing, get your password and your data and your information just by you typing it out by training the model on your keyboard.

[00:44:41] Through a and then this is the part, it, it was through a smartphone microphone that they trained it through and it could pick it up. So that made me very nervous in that you could be in a cafeteria at work somewhere open using using someone's wifi or whatever, just sitting there. Someone could have their [00:45:00] phone and literally, Look over to where you not even look over.

[00:45:03] 'cause they can hear you typing in what website you're going into. Yeah. Bank of america.com. She just typed into her username is da da da da. Her password is this, without them ever having to move, they just need to know what type of laptop you're working on or what type of keyboard you're working on.

[00:45:17] And then basically clicking the thing and say, oh, I'm, she has a MacBook Pro. All right, let's go and can do that. So that was crazy to me. So that was the first one. The second one was, there was a, I can't remember which. I feel like this was also a research one, but there was one where they were using chat, G P T or G P T three G, PT four, and actually using AI to figure out your passwords, and it had a 50% success rate.

[00:45:44] So I'm like, I'm done. Like I'm in space. Y'all can stay down there in earth, but my backdrop is a space backdrop there. There's no way we are gonna be able to keep up with the craziness that's gonna happen. You can train AI [00:46:00] on darn near anything y'all. I think you need to say that again on anything.

[00:46:04] Anything that has a pattern that can be repeated. Humans too, because we are, we're not as. Random as we think we are. We stick to patterns. Our brain works in a certain way. It can talk like you can write, like you can type like you. Yeah. If it's now guessing your password because it's that good and has that much information on you and it can listen to your key keyword types, like it's just so much level of privacy that's gonna have to happen for humans.

[00:46:29] And again, the government, I'm sad to say, is not where I needed to be on this topic. And the tip is don't use open networks. Say that again? 'cause a lot of people do not listen to that. Don't use open networks because what is the number one thing? Okay. Don't use public wifi ever. Exactly. Whether you're on a plane, train automobile at your mama house if it's open or accessible.

[00:46:54] So everyone don't use it because literally I've seen like one guy, he actually does his own purpose. He sits on planes. [00:47:00] He gets on the wifi, then he sends people and lets them know, I see all your stuff to stop them from doing it. Yeah. You have a phone with private hotspot or what have you, or use a network because you can basically give all your data away.

[00:47:14] And that's just, that was, that was from the beginning. Right. That's not even a new stuff. That's just like common privacy things you should be thinking about. But many people are not aware of that because once again, oh no, they're not. We go back to the educational system. We go back. I know you always tell me about the literacy rate.

[00:47:30] Give me that number again, because it's unbelievable. Most people speak or most people can read on average at the eighth grade reading level. So you write things at that level. 'cause most people are there. We have a lot of, we have a huge problem with literacy in this country. If you don't believe me, search literacy rates in different cities you're near.

[00:47:47] That is a big problem. But on average eighth grade reading levels where they tell us to focus. So for me, because that issue is so important, I always go back to the educational divide. I always [00:48:00] go back to the financial divide as far as resources that are not put into, for example, a low income community.

[00:48:08] And how that is going to create an issue for people who cannot afford the tech. They can't afford the education. They're attending schools that only don't even have books for every student. They don't have, you know, laptops. They don't even have the ability to pay their teachers, right. Who are not even trained in ai.

[00:48:29] How are the teachers who AI gonna teach the students? They can't, you know, so, For me, this issue is extremely important because we always have to be aware of how our population, and by the way, this is not about low income or middle class. The truth is as we're living in the US, we all have to function together.

[00:48:50] So when we start creating this divide because of AI and we don't take this serious . That's going to affect the economy as a whole, because then individuals are not gonna have the [00:49:00] opportunity to be as productive as they can be without the resources. So we really do have to have an entirely new restructuring of the educational system in order to try to bring all citizens back to at least a level.

[00:49:14] Of higher than an eighth grade reading level and have the competence and ability to use AI software to improve the economy for the next generation, because we cannot continue with this, you know, caste system of finances. It's, it's not gonna be beneficial for anybody in the long run. Yeah, that's the hard part.

[00:49:34] Anything else you're excited about this week in ai? I'm gonna do a review, so I'll do my, my fate. You guys know I love Mid Journey. Yes. So with, I guess, I don't know, I might do it tonight. I might do it tomorrow morning. Okay. Now with Leonardo, they have our, I forgot what it's called, photo reel, I think.

[00:49:54] Oh. So now we are starting to see like more models now having like better skin [00:50:00] texture, you know, more realistic features. Okay. So I do wanna do a comparison with Mid Journey and then Leonardo, I'll probably be doing that, I'll probably do it tomorrow morning. Okay. So that'll be something fun for us to try out and test out.

[00:50:15] Lemme think. Oh, that's great. Yeah, I'm trying to think of anything else like that's, it's also fine to not have anything happy about AI this week. Other than me, I'm ai. But no, this has been, this is a so the reality is each week on AI may not be happy. Some stuff is a little bit scary and a little bit sad, especially if you think about it from a human point of view.

[00:50:34] And my thing about AI is that we have to do this with humans, or we shouldn't do this at all. And we have to remember, when I say humans, I mean human emotion, compassion, understanding that there are gaps in this country. Whether you agree on why they exist or not, they exist. They're clear as day.

[00:50:49] . And if we wanna have a productive, other countries are gonna figure out how to come together quicker than us, and we gotta figure out how to come together. So let's all like rest on that as a good closing thing is that at the [00:51:00] end of the day, we all can be very successful with this thing or we all can have a bigger gap and only a few are successful than the rest of us are trying to take your house.

[00:51:07] So we don't want that. Just joking, but seriously, you started the upcycle and that didn't work. I did. I had the joke. I had to crack that joke for a second. But no, seriously, let's all figure out how to do this together. We'll be here every single week talking about all the news and ai. Thank you Marcy, for bringing all the news.

[00:51:25] Marcy's. Channel is Marcy Neulle. It will also be in the show notes on the podcast. On YouTube, I'm AI decodes at everything at AI Decode. So please make sure y'all stay upbeat and stay listening to us.

[00:51:38] We're saying this so you can know. I know you may not follow other people because you may not even know to follow those other people. So we're trying to be your go-to girls for ai. We're trying to bring this stuff and break it down in a way that's easy for you to understand. Yeah. So make sure you subscribe we'll be happier next week.

[00:51:53] This week. No. Alright, y'all.