ChatGPT Can Talk & See, DALL-E 3 Blows Us Away & We Learn About AI "Monster Milk" | AI For Humans

AI4H EP025 TEST
===

Gavin: [00:00:00] Welcome. Welcome. Welcome everybody to AI for humans. We are here to be your guides and make sense of the world of AI. We are a show that likes to have fun and make sense of these technologies that are changing our present and our future.

Kevin: It's episode 25 and this thing is going to be packed. It is overstuffed with AI goodness the major players in AI are spending tens of billions of dollars and it feels like the AI wars have actually just begun. 

Kevin: We're gonna get you up to speed on everything and perhaps more importantly, we'll tell you how all of these announcements will affect your personal and your professional lives

Kevin: plus... I've got a new way for you to generate some really surreal looking AI video in seconds, and you can do it with the computer you're watching this on, or the phone in your pocket, it doesn't take much. 

Kevin: And a major cell phone maker, Gavin, is marketing a way to use AI to replace your family. If you don't like them, you can finally get rid of them.

Gavin: Thank you so much .

Gavin: Thank you. Just kidding. [00:01:00] Just kidding in case my kids and wife listen to this I'm just kidding. That was a joke. Okay, everybody. 

Kevin: They can't hear you over the tears, Gavin. Sorry.

Yeah. Yeah. Yeah.

Gavin: Everybody please please it take a second right now like subscribe Do all the things you need to do to follow our show anytime you engage with our content it helps us helps the show so go and like and subscribe do all the things that will allow you to get our feed in your face because we are here to bring you good stuff and we promise we are not going to insult you.

Kevin: Now, our AI co host might, because 

Kevin: each and every episode on AI for Humans, we use AI to bring a co host to life. We create a little prompt, we give a little backstory, and then we let the AI take the wheel. So, should we meet our co host today?

Gavin: I cannot wait. I cannot wait. Bring her forth.

SheilaPR: Hey 

SheilaPR: there, Gavin and Kevin! I'm Sheila, the PR maven who's all about that razzle dazzle and getting folks noticed. I'm here, sipping [00:02:00] on some delicious monster milk, ready to dive into the world of AI with you all.

SheilaPR: Whoop, whoop, let's make some magic happen together. 

Kevin: Okay.

Gavin: so, let's first, okay, two things. One, just so everybody knows, our AI co host, as Kevin mentioned, is AI generated. We bring their voice to life with 11 labs, now, Sheila is a PR maven, and I had heard that Sheila was on our show to help PR person will appear, and they'll be like, Hey!

Gavin: I can help you out. There's something else going on with Sheila right now. Is that, what did she say? Monster milk? Is that what she's drinking?

Kevin: I don't know if that was a typo. I haven't heard of that particular brand of whey protein or energy drink, or maybe an awkward in between. But, whoop whoop, maybe it's Juggalo certified? Maybe it's got some Faygo sprinkled into it?

Gavin: a Faygo energy drink.

Kevin: yeah, I don't know what's going on there, but I love the notion of growing our podcast, Gavin.

Kevin: We started it for funsies, but it's becoming very serious, and we could use some help growing it. [00:03:00] That's why we beg all of the listeners and viewers each week to help us out with that. So I love the notion that Sheila is gonna bring some expertise to this. Should we see what her plan would be to grow the podcast?

Gavin: percent. This is what we should be asking. We have a PR expert on the show. 

SheilaPR: Oh, honey for Kevin and Gavin, we're going to create a buzz. That's louder than a beehive. Oh, first off, we're going to leverage social media platforms to amplify the voice of AI for humans. . Maybe even throw in a giveaway or two who doesn't love free stuff.

SheilaPR: I love it,

Gavin: we do.

SheilaPR: Let's collaborate with influencers and other podcasts in the AI and tech space. They'll help us reach a wider audience and we'll do the same for them. . 

SheilaPR: Let's get you guys some airtime on radio shows and features in tech magazines and blogs. And hey, throughout it all, I'll be sipping and showing off this Monster Milk. It's the drink that keeps me energized and ready to conquer the PR world. Another sip for the road [00:04:00] to success. Sips, Monster Milk. Whoop!

SheilaPR: Whoop!

Gavin: Whoa. What's going on with Sheila? Okay. Hold on. Now we need to directly ask what's going on here. Like what have her explain why, what does she have a connection to monster milk? 

SheilaPR: Oh, absolutely! I've got a connection that's tighter than skinny jeans on a hot day with Monster Milk. They're one of my top tier clients, and let me tell you, this drink is the bee's knees.

SheilaPR: It's brewed from the essence of ancient kombucha, giving you that zing and zest to tackle the day. Now, between you and me and the fence post, Monster Milk is on the fast track to becoming the next big thing in energy drinks. It's like if a lightning bolt and a dance party had a baby. It's all about that raw, unbridled energy, baby.

SheilaPR: And you know, the more I feel like I can see sounds and hear colors. It's a trip to Flavortown and beyond. No tickets needed. [00:05:00] Whoop! Whoop! Let's ride the monster milk wave together to the land of endless energy and exuberant excitement. 

Kevin: Whoop! Whoop! Oh no, Sheila, I think you need to... I think you need to put the monster milk down until it's, I don't know, FDA approved or something.

Gavin: she say a trip to a trip to flapper town

Kevin: A trip to Flavortown because of the, uh, unbridled energy that it gives you. 

Kevin: She's getting synesthesia, I think, from it, which is

Kevin: pretty 

Gavin: that's an incredible thing if it can happen. Um, alright, well we'll check in with Sheila again later. 

Kevin: Each and every week, Gavin, you and I mess around with AI. 

Kevin: What did you do with AI this week?

Gavin: Okay, so this week, as we said at the top of the show, is probably, I don't think this is hyperbole, Kev. This may be the biggest week we've seen in AI since we've been doing this show if you're out there and you're kind of curious about AI, Now is a time to lean in because the next iteration of these big tools are coming out.

Gavin: I think the biggest announcement from a perspective of the kind of [00:06:00] stuff that we spend time with and I spend time with, , Is Dolly three. So Dolly three. If you're not familiar, DALL-E is OpenAI's new image model.

Gavin: They've had Dolly one Dolly two and Dolly two was kind of the model that kicked off the image model boom last summer. Really, I think is when it came out because you had Dolly to come out and then you had mid journey come out. Mid journey was around, but then you had stable diffusion. The three big image models 

Gavin: What is amazing about dolly three and it's not out yet But the news is that it's gonna be coming in the next couple weeks Dolly three is remarkable in terms of what it can do and what it's generating with the same sort of prompts you're generating with now versus other image software out there.

Gavin: And again, we haven't been a hands on with this, we're going on what open AI has shown us. And a really great Reddit user who somehow has access to this software, I don't know exactly how, maybe he works at OpenAI.

Gavin: His name is u slash , derpgeek on Reddit. And he's been sharing these incredible visuals. He's asking people to send him prompts and he'll reply with stuff. I [00:07:00] saw. Really the most stunning artwork I've ever seen created by an AI engine from a prompt that was about this rabbit in the 1890s. It was like this kind of hand shaped thing.

Gavin: This is like the leveling up scenario of what AI image art can do.

Kevin: One of the important differentiators here is that open AI. If you look at ChatGPT, they were the ones that really mastered having a natural language conversation with an AI , and that's what they're applying to DALI 3, because it looks like you can just naturally converse with an AI about the type of image that you want to see, and it will remember the history of that conversation, it will keep it in context, and if you don't know specifically what sort of lens or lighting tools or illustration techniques, if you don't know the exact terminology, you can just describe in natural language what you want to see from the image, and it interprets it.

Kevin: And so, Whereas a mid journey or stable diffusion, these other AI image generating softwares, you might need to become a prompt expert and look at guides to figure out how to fine tune [00:08:00] everything. The DALI 3 approach, at least as advertised, seems to be, just ask it, and it'll figure it out.

Gavin: Yep, a hundred percent. So let's take a look at a few of these examples and we can go and show how mid journey and ideogram works on them to 

Gavin: look at the potato one first. The prompt for the potato image was tiny potato kings wearing majestic crowns sitting on tiny thrones.

Gavin: overseeing their vast potato kingdom filled with potato subjects and potato castles. So you look at the Dali prompt and it's, you know, it's very cute, right? 

Kevin: It's a large potato king, , in the center of an image. It's tilt shifted, which means you're kind of looking down to give it that forced perspective, so you get the sense that they are miniatures on some sort of wooden table.

Kevin: , it gave the king cute little black dot eyes and a little upturned smile. And then it's got, he's got his little spud companions there, right? His tiny little fingerlings? His tiny, they're not underlings, they're

Gavin: yeah, their finger legs. Of course, yes.

Kevin: He's got his tiny little potato family, and the whole squad has his same kind of eyes and [00:09:00] mouth.

Kevin: The focus, the lighting is here. Every little potato is casting a realistic shadow. There seems to be one major source of lighting, which is sort of blowing out a little bit of the main king potato. It's got a beautiful royal, , cape on with some gold inlay.

Kevin: There's, like, beautiful details on this image. And the prompt didn't say... Shoot with a Leica lens, 35mm f stop, blah, blah, blah. Dolly 3 just made a bunch of... Decisions, basically, and out came this beautiful looking photo.

Gavin: And I think what you'll see is we compared to the other ones is it's also parsing the language really smartly. So all I did was do a slash. Imagine with mid journey. Same prompt and you can see it still gets little kind of potato creatures, but They kind of made guesses on what they were.

Gavin: First of all, they have toupees on. These are, they're almost like eggs with toupees in some way, this one. Like, it's a, it's a strange choice,

Gavin: . One thing I want to point out about this is it's like the majority picture is still really a cool picture, 

Gavin: right? Like it's lit well, like it's beautiful. Like if you showed this [00:10:00] to somebody, you'd be like, Oh my God, a computer made that. But it's not what you asked for per se, right?

Gavin: It's not exactly what you asked for. 

Gavin: , now go to ideogram. So if you're not familiar with ideogram, we feature this on our show. Ideogram is a new model created by ex Google employees. So this gives a kind of a unique kind of take on it. It's a little more cartoony in some ways, but again, I'm not sure that it really got the prompt in the same way.

Kevin: to be fair, the prompt doesn't say, hyper realistic or photo at all, so the fact that the model is choosing to go in a cartoonish route, I, you know, I won't shade it for that, but,

Gavin: But what give me your give me your own unvarnished opinion. What is your first thought looking at this one?

Kevin: It looks kind of cheap, it looks like something, it looks like a draft. The other renders look like a fully featured something. This, stylistically, is an interesting choice. I don't know, I guess it does, they read potato a little bit more because it's cartoon and there's potatoes at their feet.

Kevin: Which, I don't know if [00:11:00] they're stepping on their citizens as they walk through.

Kevin: there's also potatoes floating in the sky, there's like weird Pringle clouds. So, uh, it's, it's not, it's not bad. And I love that we have choices, I love that there are multiple AI image generators.

Gavin: Okay, so that's the start of this. We're going to go through a couple other ones. This one, look at the meat landscape. So this one is... Remarkable. When I saw this one, I was like, this is insane. If people are going to be able to generate information and images like this, I will not believe it. So the prompt for this one was a vast landscape made entirely of various meats spread out before the viewer, tender succulent Hills of roast beef, chicken, drumstick trees, bacon rivers, and ham boulders create a surreal and appetizing scene.

Gavin: The sky is adorned with a pepperoni sun

Gavin: and salami clouds. 

Gavin: Yeah. 

Kevin: I gotta check my wife for a Neuralink because she had a nightmare the other night and popped awake at about 4am in a flop sweat about, yeah, this is, this, this torturous animal valley. How sadistic that a [00:12:00] pig is having to

Gavin: I know!

Kevin: on slices of itself.

Gavin: ham! On it's own

Kevin: it's on its own bacon, what is going on?

Gavin: Okay. So now go to the mid journey version of this go to meet mid journey meets and this is exact same prompt

Kevin: Missed opportunity for meat journey, by the way. If we just, if we just make a mid journey that only does cold cut

Gavin: should, that should be next week. . Okay. So exact same prompt. And this is a lot of what like you normally see, 

Gavin: so hot dog city, if you're not familiar, which you probably aren't hot dog city was an adventure of Kevin and I went on in a much earlier podcast where we try to create visuals of a city made out of hot dogs.

Gavin: And it was very difficult because weird, surreal images are really hard to make a lot of the time. And what you get in this image, if you're looking at it, is a vast landscape, almost like a canyon, in what looks like, I guess, a bunch of meats, but could also be kind of like cheeses or something on the ground.

Gavin: But it isn't made of meat in any sort of way, correct?

Kevin: Big fail. These look like macaroons!

Kevin: Who wants this?! [00:13:00] this?!

Kevin: isn't Meat Mountain! There's like a weird atomic blast in the background. There's like, what look like cotton candy flamingo clouds. Like, is it an interesting or cool image? Sure. Absolutely. Is it at all related to the prompt? If I saw this, I would say nay.

Gavin: so move on to the Enneagram meets. At least this feels like it's a little bit closer But it's so surreal in that it didn't really understand it in any specific way And I think what's interesting about this when you compare What dolly three is doing versus mid journey and ideogram here.

Gavin: Dolly three is just doing a much better job of parsing the words in the prompt itself. And this is something that they've talked a lot about that the way it reads the prompt and how it's grabbing the images out of the prompt and putting it together is their secret sauce. And, and Kev, as you mentioned earlier, one of the things I'm really excited about that they've showed off with dolly three is the ability to take an image.

Gavin: And then change it based on a new prompt. And that's always tricky to do with mid journey or with stable diffusion. You can in [00:14:00] paint, but this is the idea that by changing the words in the prompt, you can actually change the image itself.

Kevin: On that point, a slight detour, and then we can get back to the final example. , there's , a tweet. It's linked in our show notes. Nathan Shipley, , a. k. a. Citizen Plane on X, did a Dolly 3 test. And someone on Reddit suggested, create a list of 50 different objects and see if it can generate that. So he used chat GPT generated a list of 50 objects, like a teacup, a basketball, a grandfather clock, a typewriter.

Kevin: I mean, random objects. Dali 3 nailed it and gave him an image of all 50 objects neatly arranged in one scene. And then, He supercharged it by saying, Give me an image of a professional surfer on a massive wave. Struggling under the weight of these 50 objects and it did it and it remembered the 50 objects and it put him as a Surfer is there with his mouth open clearly straining getting the emotion of the character on the surfboard Gavin [00:15:00] And then he swapped it to being an old grandmama and you have grandmama Carving a pipe with all these objects on her surfboard.

Kevin: That is incredible.

Gavin: And again, I think this goes to the point of we are entering to the next stage, right? So I, the last example, look at the Venus picture. So there's a picture. One of the things obviously we talked about with mid journey, 

Gavin: is text, right? And when you have text, mid journey just fails completely at it. There's this beautiful poster of Venus, right? Almost like a tourism poster of Venus. The prompt for the Venus poster was a vintage travel poster in Venus in portrait orientation.

Gavin: The scene portrays a thick yellowish clouds of Venus with a silhouette of a vintage rocket ship approaching mysterious shapes and mountains and valleys beyond the clouds. The bottom text reads, explore Venus, the beauty behind the mist. So if you look at the image from this particular, , from Dali, you can see not only does it get the image right, but it actually gets, if you zoom in on the bottom, which is kind of crazy.

Gavin: It's not, you 

Kevin: you can 

Kevin: see Tiny little [00:16:00] weird, Klingon font to the bottom left, that weird AI hieroglyphic font. But it does say the mist with two S's for a double dose of the secrecy.

Gavin: so something to point out with this stuff is if, and we can go, in fact, look at the mid journey version of this same image prompt. So if you go to the mid journey version of this prompt, you can tell immediately the word Venus itself is not spelled, right? So you can't see Venus. And the interesting thing to me from somebody who's done graphic design stuff or tried to do a lot of AI art is , okay, we've now got to the level where like the big words it's gonna get.

Gavin: It's already starting to get the smaller subtext right and that is another step in a crazy direction where you could like really think about a world where you could make a graphic image with like multiple lines of text and you don't even have to do any Photoshop work to it 

Kevin: I'm so excited to get my hands on this because I bet you could with natural language say, Hey, the bottom text on this poster is incorrect. Make it say this. And then when Dolly can focus on that part of the [00:17:00] task, the regeneration's gonna be so much better.

Gavin: I can't wait. I mean, listen, who knows? You never know with what's real and what's not. But like we said, we've seen people out there using this and promoting it and showing what they can do. So hopefully access will come for everybody who has chat GPT plus very soon. That was the thing that I did with AI this week. Kev, what did you do with AI this week?

Kevin: I checked out a new service called genmo. ai hashtag, not an ad.

Kevin: We will get to one of those, but genmo is an AI platform, Gavin, that it gives you a bunch of different tools and, AI widgets to play with. So if you want to make 3d assets, you can go in there, type in what you want, and it will make a 3d model of the thing. You can create animations. You can do image generation by chatting with their bot, but what they unveiled.

Kevin: was their new video creation capabilities. And you and I always love AI video generation. I especially love the broken stuff

Gavin: Oh, it's, uh, it's so much fun to watch things fail. , I think in general people are either horrified or filled with joy. I think you and I are both filled with joy by that.

Kevin: The [00:18:00] key is insertion, by the way. I've maintained that my entire life. Whether it's food into mouth, or, like, a humanoid into a car, the models just don't get that concept yet. There's a lot of broken video of, , weird AIs walking and then being sucked into a vehicle. 

Kevin: I love that stuff, but Genmo is actually pretty competent already. , it is in a weird in between. When we've mentioned AI video in the past, we've mentioned, , Runway Gen2, which is the gold standard right now. There's PicaLabs, which is another company that's making some quick strides and starting to catch up to Gen2.

Kevin: There's some open source stuff genmo is this weird in between where it wants to make prompting as simple as a single word, Gavin. It wants you to be able to give it the term hamster, or spaceship, or hot dog, and let it run with it. And it does a pretty good job.

Kevin: of figuring out an interesting color palette, an interesting look. You can even control the camera if you want it to zoom in 

Gavin: Oh, that's cool.

Kevin: rotate. . But they want to be this simple, just give us a word and we'll give you a [00:19:00] video. Issue with that is that, if I say hot dog, what are we talking about?

Kevin: Is it a hot dog on a plate? Is it someone eating a hot dog? Is it a cartoon of a hot dog? Is it a hot dog soaring through the air? Is it Hot Dog City?

Kevin: So when you try to give it a more detailed prompt, , it doesn't do a good job of capturing all of the things that you ask for. , for example, I gave it Hot Dog City, and it kind of evokes a hot dog like palette. There's even a, a few frames from the start of the video where it looks like they're miniature hot dogs, like hot dog gumdrops

Kevin: Turkey jousting was an interesting one because I didn't specify should there be a human on top of a Turkey, should it be turkeys on top of horses jousting with poles? It gave me, Two bird looking creatures.

Gavin: ha

Kevin: that I would call them

Gavin: No, I, yeah.

Kevin: but it put them in an arena of sand or dirt, and if you look in the background, it kind of looks like a renaissance fair.

Gavin: Yeah, totally,

Kevin: banners and an audience. So it did a good job with that, and then, all I wanted, Gavin, was a cannon that fires kittens.[00:20:00] 

Kevin: I want a kitten flying through the air, like, as if it were shot out of a cannon , like, a circus performer. I don't want this to be 

Gavin: like, cartoony in a fun

Kevin: Yeah, I was like, give it a helmet and goggles and a cape if you want.

Kevin: I don't care, it's not supposed to be violent, it's supposed to be adorable. 

Kevin: So if you look at some of the renderings, I got a kitten sitting next to a cannon, with no ear protection, very dangerous. But it gave me a It gave me a kinda lookin cannon thing and a kinda lookin cat thing, , and then there's another thing labeled Kitten Cannon, which is not at all.

Kevin: It's just a weird psychedelic cannon floating through space. 

Kevin: April wanted to see a UFO shaped like a hamburger abducting something, and I tried a couple different prompts of it. Yeah, it was a good idea, and

Gavin: It didn't really work out

Kevin: Genmo didn't quite get the task. But then when I just gave up and said, hey, make me cool spaceships, you seem to be pretty good at that.

Kevin: It is! It's pretty good at rendering some spaceships. There's a weird cyberpunk warrior looking thing, which I thought was cool. 

Kevin: one of the most beautiful ones is just a close up of a psychedelic eyeball. , and [00:21:00] the prompt was, Melting Influencer.

Kevin: I didn't know exactly what I would get, but it pushes into the eyeball, and it's really a beautiful, coherent video. 

Kevin: whenever you generate an image, fine, when you start moving and the AI needs to, , keep the coherency, it needs to understand what came before it, and make the model or the person or whatever's in the image.

Kevin: track and look the same as it moves, a lot of times these things fall apart. And here, it's a subtle little move, but it really keeps the coherency, and I think looks really nice.

Gavin: What's going to be interesting to me with this video stuff is I'm really curious to know. It's going to be about the teams behind them, right? You see runway putting a lot of effort and a lot, I've assumed a lot of money and engineering talent towards making this a much better product towards making this more coherent towards giving you more tools and controls over the camera.

Gavin: I just am not convinced that companies will have to see where the scale gets to for something like this, right? Like , this is to me. Below the level of runway ML for sure, like it feels like it's, it's significantly [00:22:00] lower than gen two video.

Kevin: It is, but I think where they hope to excel is if you look at old mid journey stuff, I think the reason mid journey was able to pull away and excel so quickly was they had so much human learning. Right? They got the reinforcement learning, human feedback, they got all the stuff of people saying thumbs up, thumbs down, and they were able to fine tune their model and go, Oh, okay, people like this style more.

Kevin: People like these features more. And so if a Pika Labs or a Genmo or anybody else playing in this space can get enough users and has the team. The proper team in place to gather those learnings, then we could see some real challengers emerging.

Kevin: by the way. . Just to round this out, go to genmo. ai. It's completely free.

Kevin: It's really fast. They give you a bunch of credits per day. You can make a ton of videos. You don't need a powerful computer. You can do it on your phone. Hashtag not an ad. Just something that people should play with. Go and try to psychedelic hamburger or melting influencer. Or see if you can make a better Hot Dog City.

Gavin: Please and then send it to us. Yeah. Send it to us. 

Gavin: Well, speaking of not an ad, it is now [00:23:00] time for a very special message from a friend of ours who has decided to sponsor our show. Ladies and gentlemen, supermanaged. ai. Slash humans is the URL you go to to learn about supermanage.

Gavin: Supermanage is a really cool new program that allows you to use AI to get to know your employees better. It actually takes a bunch of their messages in and can give you sentiments about what they're expecting and what they want out of their job. I think we have our very special guest, Sheila, to say some wonderful words about the people at supermanage AI.

SheilaPR: Oh, 

SheilaPR: Kevin, Gavin! Gather round for a tale of triumph and 

Kevin: teamwork.

Kevin: Okay, we're here, we don't have to gather, Sheila, we're all here on the podcast, I, but, great. 

SheilaPR: Imagine strolling into your Wump Ones armed with the power of a 

Kevin: thousand suns. Okay, by Wump Ones, by the way, she means one on ones.

Gavin: Wump ones. We're gonna have to have a wump one this afternoon. What are you talking

Kevin: That sounds like an HR violation, I don't know if I signed up for a Wump 

Gavin: Ha. Ha. Ha. 

SheilaPR: to SuperManage [00:24:00] AI. In just two minutes, get a snapshot of all the chit chat from your public Slack channels. No more catch up, only leveling up. Unleash the potential of every meeting. Dive deep into contributions, challenges, and engagement.

SheilaPR: And watch your team bloom like spring flowers after a rainstorm. That's nice. And guess what? It's free for one manager and up to 15 team members. Whoop, whoop. Whoop, whoop, indeed. Now, let's toast to thriving teams and seamless 1. 1s with a hearty chug of Monster Milk. No. Chugs Monster Milk. No, don't, and don't say it when you Oh, I can feel the cosmos aligning and the stars spelling out success.

SheilaPR: Onward to victory with SuperManage, AI, and Monster Milk. No! 

Kevin: SuperManage.

Gavin: is not sponsoring us. Yeah. Monster Milk is not sponsoring us. Let's just be clear. Supermanaged. ai slash humans. Supermanaged. We don't know what Monster Milk is. Sheila has brought Monster Milk into our world. But,

Kevin: She can feel the cosmos aligning? [00:25:00] I screamed that once on Salvia, I don't think monster milk is a kombucha based drink, Gavin. I think we need to get to the bottom of that later.

Gavin: Okay, let's get into that. Alright, you know what time it is, Kev? 

Gavin: It's time for the next energy.

Gavin: Why say news?

Gavin: Zero sense, but it is time for the news. Everybody

News, eh, news, it's the A. I. for humans, news, 

Gavin: Let's start off, I think one of the things that you and I had said we want to do this week is just catch everybody up on where everything is, because it's almost too much. So if you haven't been paying attention, there are, three to five major players in this space.

Gavin: You have the open AIs, who's our partner with the Microsoft. You've got the Googles, who are kind of on their own. You've got the Metas, you've got now Amazon involved. What we want to do is take a quick [00:26:00] step through each of those companies and say what's going on and tell everybody what's important to pay attention to.

Gavin: And I think Kev, why don't we start, we talked a little bit about dolly three already, but there's some big, big news coming out of open AI this week and it involves them basically, I think trying to get ahead of what Google Gemini is big announcement is going to be. What did, what did they have to say?

Kevin: Yeah, they had some massive announcements, Gavin. And you might be aware, because I know everybody listening follows AI for Humans show on TikTok and on YouTube. OpenAI demoed multi modality. 

Kevin: The ability to interact with these language models, not just through text, but through sound and through imagery, and that could be stills or video.

Kevin: And OpenAI, demoed the next version of the app, which should be rolling out in the next week or two, and you will be able to do a few things. That's one. You'll be able to have a natural language conversation with ChatGPT.

Kevin: So you can just speak with it. Yeah, with your own voice, ask a question and you get to choose the voice that you want for OpenAI or ChatGPT to speak back to you. So now you can get your kids that bedtime story and [00:27:00] let them interact with it. You can ask for help figuring out a complex legal form. You can have it generate 50 buzz worthy, must click YouTube titles.

Kevin: That probably won't work. But, you can have a natural language conversation with it, which is big, and Gavin, you think , this is going to get AI into the hands, or I should say, into the heads of more users.

Gavin: You and I have both talked about how, , incredible it has been to use PI eight eyes. So if you're not familiar right now, today, you can download PI's, , iPhone app, or I think they have an Android app now, but you can download their app and you can interact with it via voice. And there's a magical moment when you can just talk to something.

Gavin: And I had this moment the other day, I was trying to get something out of Alexa and Alexa is so dumb. I'm sorry, Alexa, it's true. You and Siri are both dumb. And it almost is like prompting, right? One of the things I've thought about is. To get anything out of Alexa, you kind of have to learn a language to understand what you're asking it.

Gavin: What this changes is you're going to be able to say to the chat GPT app anything you want and it will respond [00:28:00] back. 

We want to hear a bedtime story. Tell us a story about the super duper sunflower hedgehog named Larry. 

Larry was a unique hedgehog unlike any other. He had bright sunflower petals instead of spines. Okay, it's bedtime for Larry. What does he do to go to sleep?

Larry curls up in his petal blanket, letting its softness embrace him. Luna sings a gentle lullaby, her glow dimming to mimic twilight. The scent of sunflower fills the air, and Larry drifts into peaceful dreams. Good night, Larry.

Gavin: Now, it's never going to be perfect. We're still learning about stuff here. But I can guarantee you as soon as this comes out, if you try it, show it off to grandma, show it off to anybody in your family.

Gavin: It is going to shock them and that is the next level like, right? There's only so many people that are going to spend time on chat. GPT and we know a lot of people have, right? It was one of the fastest apps ever to 100 million users. This is what's going to unlock the next like 500 million users. What people want is that sci fi moment of saying you're in a room and there's nothing around you and you [00:29:00] say like hey Open AI Can you tell me what this wrench I need to use for this machine is?

Gavin: And it tells you that is the future. That's the thing that I think is going to really be the unlock and open AI is going there in like two weeks, which is pretty crazy.

Kevin: They also demoed, , the multimodality image support, Gavin. So if you're having a conversation with chat GPT, you can send it an image and then make a query or a request based off of that. So where's this going to lead? This could lead to taking a photo of your fridge or your pantry and saying, 

Kevin: what can I make for dinner tonight? Hallucinate a recipe for me that's gonna use everything that's gonna expire on these shelves

Gavin: No kitten and cannons, no kittens and cannons, please. And the

Kevin: I hope you don't have that in your cabinet The demo that they used was someone who wanted to adjust their bicycle seat. Pretty simple task.

Kevin: They snapped a photo of the bicycle. ChatGPT would tell them, ah, you gotta do X, Y, or Z. They then took a closer photo, and they demoed the ability to draw on the image, which would help focus ChatGPT on an area. And in this case, they drew a little circle around a part on the [00:30:00] bike and said, is this the knob or the lever that you're talking about? And ChatGPT clarified. , keeping, again, the context of their conversation in mind, providing instructions for how to adjust this bike seat. And this could be a game changer for how we interact with AI moving forward. This is something, by the way, that Google's BARD already does.

Kevin: The problem is that BARD, holistically, Hasn't proven itself to be as capable or, , interesting, let's say, as ChatGPT. So the fact that they're rolling this out within their mobile app, so you're gonna run the same app, you're gonna have a conversation with it, you could even say, oh, hang on one second, I'll just snap you a photo of what I'm asking about.

Kevin: Boop, here's the landmark, or here's the area of town that I'm in, now guide me through it. And they did release a little bit of info. Regarding safety and security and some of the stuff that went into this. I think we're going to see chat GPT, at least with regards to image based queries, we're going to see it being a lot more.

Kevin: It's going to be more careful about [00:31:00] discussing sensitive things. Whether it's a question about the ethnicity or the gender of someone in an image,, identifying someone in a photo, who is this person where we don't know if that person gave consent or if the photo was taken in a public place.

Kevin: And. The other thing is medical diagnoses, which is something that you and I joked about on TikTok, but basically because the AI has a habit to sometimes hallucinate, to lie or make up something if it

Gavin: my god, that's pretty scary, right? You're

Gavin: like, You've got terminal cancer, you're going to die in six days. Wait, oh no, sorry, that was just a thing on the lens of the camera.

Kevin: WebMD used to be, I stub my toe on the coffee table. Oops! I give you about four days to live. And it's like, wait, what? No, it's, Oh, the area's red and sore. That's probably this rare disorder and you're gonna be done for. No, I just stubbed my toe!

Gavin: and that gets back to the trustworthiness, right? Like the one thing to keep in mind is that these are not fully trustworthy and these are always getting more and more of improved and I think open AI smartly is doing that We don't have to spend a ton of time in this, but , if you don't know this, Microsoft has invested a lot of money in open AI and [00:32:00] has partnered with open AI 

Gavin: microsoft themselves, is going to have Dolly 3 enabled in Bing chat, but also unveiled an entire AI suite of Windows 11 tools in co pilot, which is going to make interacting with your just. Basic computer, very AI centric and hopefully way easier. I can tell you as somebody who spent a ton of time on Mac.

Gavin: And got a gaming PC kind of for the show, windows is a disaster. I'm just going to say it right up here. I know this is probably going to like make a ton of people angry, but windows is the least intuitive interface I have ever used across the last 10 years of my life, and I've avoided it mostly.

Gavin: And I was a kid who grew up with windows, right? I grew up playing PC games, so I had windows my whole life. I don't know what they've done to it, but they've piled like multiple windows on top of other windows, 

Gavin: I would love an AI to be able to answer my mom's questions about how to get to the pictures from last year. Rather than me have to spend a half an hour on the phone with her stepping her through the thing.

Gavin: So windows if you could do that, you've solved a nightmare problem for me that I've had for [00:33:00] 50 years or whatever. 

Kevin: People talk a lot about the competitive advantage, the moat that Google has in the AI race, Gavin, because they've got so many accounts and so many users across multiple websites. Majority of the PC users out there are Windows users, and they own the lair on top of everything. And so when you can summon AI on demand... You might use the Windows AI to help you write a better something in the doc that you're in. Or search for those photos like you mentioned. Or, instead of giving the advice of, well, Google it if you don't know how to turn on dark mode for the system.

Kevin: Well now you just go to the AI and you'll click it and say, Can you help me find the setting for that? And the AI is just going to do that. So Microsoft has that layer. On top of everything. So no matter what service you're using, you're gonna have one click access to their AI.

Kevin: That could be very, very powerful. 

Gavin: So right now we're waiting on Google's next big release. Google Gemini is it's LLM that it's been working on for a while.

Gavin: The DeepMind founder has been involved in it. They've really put a lot on this, and honestly, we don't know a lot about it yet. We don't know. Is it like an image [00:34:00] generation software? Is it? we know it's going to be multimodal because, this Google CEO said it a thousand times, but we don't know what it's going to feel like.

Gavin: We don't know how it's going to be. We assume that some of the bar stuff is going to be there, but I am assuming this is there's going to be a massive leap, and we don't know what it is yet. 

Gavin: The other big news that happened this week is that Amazon. Invested 4 billion in anthropic. So Amazon, they are finally made their move and their move is the third kind of big player in the A. I. World after Google and Open A. I. Is anthropic. And we assume, according to announcements that that is going to be ingested into things like Alexa or all the things that are already there.

Gavin: And then finally... Facebook meta, , has been building its own things. And many people have like kind of championed Mark Zuckerberg for being open source.

Gavin: In the meantime, they are creating a really interesting, , side business of AI bots. , and they're going to do something kind of similar to what, , Snapchat did a while ago and they're going to release a few specific bots for users to play [00:35:00] around with, 

Gavin: so they've created an AI called Bob the robot They're supposedly loosely based on Bender from Futurama, but it's described as a sass master general and that is a sassy robot that taps into the type of Farcical humor that is resonated with young people.

Gavin: So let's just be clear

Kevin: This sounds like a character we rejected on, , last week's episode. And the week before that.

Gavin: Exactly. There's just so much stuff happening right now, so that hopefully catches a little bit up. I think we should ask Sheila what she thinks, because one of the things I think a lot about with what's going on is we are in a full blown race now, right?

Gavin: We're actually like six months out from that six month pause. , if you remember way back when, there was a, there was a memo that was sent around. It was like, hey, we should take a six month pause on AI. Well, we didn't pause, and here we are now. 

Kevin: We tried to hit pause, but we hit the fast forward on accident. And so now we're just moving at 2x,

Gavin: Exactly. So let's ask Sheila, an AI, what does she think about all these companies racing to release AI tools faster and faster? .

Kevin: It's a good question. Sheila, if you could put the monster, monster milk 

Gavin: Please put it [00:36:00] down. Put it down.

Kevin: when she sips now, it's frothing out of the sides of her lips. It's all over, it's really, it's unbecoming. I don't know if it's like... A Diet Coke and Mentos

Gavin: Oh no.

Kevin: it hits your mouth, apparently it just explodes.

Kevin: It's not good. 

SheilaPR: Oh, honey bunch. It's like watching squirrels on a sugar rush. All this hustle and bustle in the AI world. It's a spectacle, a real hoopla. Companies are popping out AI tools like popcorn, and it's a race to see who can butter the public up the best.

SheilaPR: Now, is it a good thing? Well, it's like biting into a mystery jelly bean. You might get a burst of delightful cherry or a punch of peculiar pickle. We're getting a smorgasbord of innovation, a buffet of brilliance. But let's not toss caution to the wind, like confetti, at a parade. We gotta keep our peepers peeled for any hiccups or hullabaloos.

SheilaPR: But you know what? No matter the tide, we'll ride the wave with a can of monster

Gavin: Oh no!

SheilaPR: to face the future with [00:37:00] fizz and flair. Sips, monster milk. Whoa! I feel like I'm surfing on a rainbow, riding the ripples of revolution. Whoop whoop! Let's keep the conversation rolling like a disco ball in the 

Kevin: sky.

Kevin: I don't, they, they spin, they don't really roll, but that's fine.

Gavin: They don't really in the sky either, but interesting,

Kevin: Oh, Gavin, I don't know about Sheila.

Gavin: Yeah,

Kevin: Maybe we gotta bring the Sassmaster General on from Meta.

Gavin: that's a good idea. Okay, let's move on to our next story, spotify just announced, , what we expected was going to happen maybe down the road from these larger companies, but it's already happening.

Gavin: They have announced a tool that's not out yet, but they have they announced it and they showed a couple examples of a tool that can translate it. Full length podcast into different languages. So you know, we've talked about a bunch of tools that could do this on the show. Hey, Jen has a specific tool in their labs right now, which will allow you to translate your voice and put it into a different language, but keep the same qualities of your voice.

Gavin: Kevin, I was thinking about this. Do you remember maybe it was a month ago or maybe five or six weeks ago, we had a cool [00:38:00] thing of the people did with AI. And it was like a, a one off guy had figured out how to do this himself. And he's like, I'm going to make a company out of this. And now here's Spotify 

Gavin: doing 

Kevin: was gonna say, Spotify announces that 15 startups just went out of business yesterday. Let's listen to a little bit of it so you can hear how it works.

Kevin: When you say the word toxicity, how are you defining that? You're saying like toxic chemicals,

Kevin: so here's Lex Friedman coming up now. How do you think English.

Gavin: crazy.

Kevin: The fact that as a content creator, you can create , whatever language you prefer, press a button and suddenly it's available to a worldwide audience.

Kevin: And it sounds like you. That's impressive.

Gavin: that's the thing that I think people maybe they might hear. Ah, that's cool. It can be different languages, what it's doing now is taking your sound your voice and turning it into that other language. And that's really, really remarkable. 

Gavin: The fact that Spotify is doing this and gonna make it [00:39:00] available, but it is like the 12th story on the list of things that are moving the needle this week in the AI world. But I do think for our audience, . It's something to be impressed by, because , 

Gavin: we make podcasts. And , if we could have our podcast heard by people in India or in China or in Brazil, all these places that speak different languages, like that is a very, very cool thing. .

Kevin: Yes. Yeah. And if we could also just get it heard by native English speakers in America, that would be impressive too, Gavin. I think it

Gavin: Yeah. Sheila, Sheila, come on. What the hell? That's what we wanted you on

Kevin: another

Gavin: Oh no, no,

Kevin: her hands are shaky right now,

Gavin: Let's ask, let's ask Sheila one question. I want to know what is inside monster milk. She's so interested in talking to us about monster milk. Can she give us a list of the ingredients of monster milk? 

Kevin: It's mostly lead paint. I guarantee it. Lead paint and taurine.

SheilaPR: Oh, sweet maracas and glittering galaxies. Inside this cosmic can of monster milk, you've got the ancient whispers of kombucha, a splash of starlight, No. a tickle of [00:40:00] tangerine.

SheilaPR: And a waltz of wild berries. Okay. It's a jubilee, a jamboree in a can. Okay. Whoop! Whoop! Let's dance through the dimensions with every sip. And let the flavors do the foxtrot on our taste buds. 

Kevin: Okay, that, I don't think that was an actual ingredient list, Gavin.

Gavin: Monster Milk is definitely not legal in the states. Let's be clear. It may be legal somewhere else, but you are definitely not going to be able to buy Monster Milk. 

Kevin: I got a feeling Sheila's making this in a basement somewhere. There's buckets,

Kevin: And she's just stirring up batch after batch. There's no label on the can! It's handwritten! You can see it!

Gavin: Let's get to our last news story, Kev. What happened here with Samsung and AI?

Kevin: Oh, finally, Samsung is going after my father, giving him a feature to replace his family, in all of his memories. Samsung has a Pixel 8 advertisement, , a real one, where they're showing AI faces technology. So if you have a photo of, in this case, your family on a carousel, but you couldn't get your dumb son to look [00:41:00] up and smile for two gosh darn seconds, Kevin, what's wrong with you?

Kevin: Enjoy this trip! Sorry. It's channeling my father. 

Gavin: This was like, I see your acting abilities. Very nice.

Kevin: Oh, it was, that was just, \, my recall ability. You mean, there was no acting there. I'll never forget that trip to Shasta Lake. 

Kevin: According to 9to5Google, there's no explanation of how the tech works, Gavin.

Kevin: We don't know if it's using AI to create a Laura, model of that person, and then generate different faces, or if it's pulling faces from other photos that you have, and then just massaging them into the photo. But you can see... Photos of a family on a carousel, and one kid's looking away and suddenly he's staring at the lens with a bright, beaming smile, which is exactly what you want in the YouTube thumbnail as you exploit your family for clicks.

Gavin: I will say this being the father of two teenage daughters. There is it's really hard to express how hard it is to take photos and I because every time you take a photo and I've given up in some ways like this is why there was a lot of photos of the [00:42:00] kids from like age 0 to like 8 or 9 and then there's way less because What they want to do, all kids now want to do is they want to have control over what the photos look like.

Gavin: So there's probably two parts to this. One is sure as a parent, this is great because it means that every photo you have, like you could have their faces, but no kid is going to allow this to happen either. So this is just another layer in the ability of photos not to be taken. Soon we're going to be a whole society without pictures, or it'll be just the most perfectly crafted, like Photoshopped AI images.

Gavin: And that's probably where this is going to, which

Gavin: is kind of 

Kevin: year, every year on your birthday, your face will be scanned. So that the family has an updated AI avatar, and then when you go to a location

Kevin: it's just gonna hallucinate the best version of that. There's gonna be nobody else in your Eiffel Tower photo. The Leaning Tower of Pisa? Oh, you could do this with it!

Kevin: You don't even have to do this! Just tell it that their kid was doing the thing where they're pretending to hold it up, and it will generate it.

Gavin: Honestly, this is a business we should start, by the way. You could easily, at every single tourist trap, create a thing where it [00:43:00] scans a face, and you just give them the perfect image, and you give them a You could even sell them like 20 of them in different places around the area. Kevin, what are we doing here?

Gavin: What are we doing? Yeah, exactly.

Kevin: be doing this. , listen, uh, this, this is a, it's a weird video, but it's actually an interesting piece of tech. 

Gavin: If you're not familiar at the end of each show or towards the end of each show, we do something unique and different with AI this week. I saw a tweet that I think a lot of people should understand what it means. And I tried to think like, how could I get other people to understand what this means?

Gavin: So the tweet is from Sam Altman, who's the CEO of course they are on the far edge of building AI tools 

Gavin: we don't know what goes on behind closed doors, but the tweet that he put out was this short timelines and slow takeoff will be a pretty good call, I think, but the way people define the start of the takeoff may make it seem otherwise. So. This seems like a simple tweet, but there's a lot of information that's baked into this to understand if you're not super familiar with the language and vocabulary around AI, short versus long [00:44:00] timelines, slow versus fast takeoff are both important things.

Gavin: So Kevin, I thought it would be fun for the demo today. is I thought we could create a third grade teacher who could bring this information to us in a very simple way and maybe we could ask her other questions as well. So shall we meet , , our special guest for the demo of the day?

Kevin: Oh, let's do it.

Mrs. Thompson: Now listen here, Kevin, Gavin, and all you listeners of AI for Humans. I am Mrs. Thompson, a seasoned educator with over four decades of experience molding young minds.

Mrs. Thompson: I stand before you, clad in a tailored dress with a ruler brooch affixed to my collar, and glasses perched upon my nose, ready to embark on a journey of learning with you all. In

Gavin: okay.

Mrs. Thompson: my classroom, there is no room for nonsense or tomfoolery. I will employ tales as old as time to unravel the complexities of A.

Mrs. Thompson: I. That's not a surprise. Ensuring that each and [00:45:00] every one of you grasps these concepts clear as day. But make no mistake, gentlemen. Gavin? I expect nothing short of your utmost attention and respect. That's right, Mrs Thompson. Let us proceed with the learning. And remember, I have my eye on you, Kevin and Gavin.

Mrs. Thompson: No monkey business on my watch.

Gavin: I wasn't expecting that but what did she say at the end?

Kevin: Shall we begin?

Gavin: That was kind of creepy, okay. All right. So let's start here. I this is a very I know we're you've got like 180 degrees from Monster milk woman Sheila to now mrs. Thompson, but mrs. Thompson is designed specifically This is what you could do with AI sometimes design characters to help you understand a topic.

Gavin: So let's feed

Kevin: But Gavin, hold on, I'm sorry, I don't know how to program. I'm not some big computer tech genius with the ones and zeros on my digi machine. How will you say create a character? I gotta be fuckin Walt Disney over here?[00:46:00] 

Gavin: No, you don't.

Gavin: It's so good. you 

Gavin: asked. Yes. That was incredible. You're almost as good as open. You're almost as good as chat GPT four. So here's how you do this. Let's be clear. Create a character in chat GPT for is as simple as asking it to help you create the character and what I often do when I'm doing a prompts, I'll go in and I'll say I would like to create a dialogue with a character like this.

Gavin: And then I literally say, here's the things I think it would be. And I say, ask me the questions of what else you need to know to fix this. So I did that a couple times with this character, Mrs. Thompson, and I asked it specifically for a teacher, somebody that was probably third, fourth grade, a younger elementary school teacher who could explain things super easily to us.

Gavin: But also I wanted to add a little bit of personality, so I made her stern and I gave her kind of like a background where Kevin and I were going to be, the troublemakers in the class, which is probably both what I know that I was, Kevin, I'm sure 

Kevin: Yeah, oh, 

Gavin: well. Yes, yes, but the two of us are both not well liked necessarily by every teacher we had.

Gavin: So now what we can do is we've taken and we've created this character. We can feed that [00:47:00] character who has all the knowledge of open AI fed through this lens. We can feed that character Sam Altman's tweet, and I'm curious to know how she will explain it to us as if we are her third grade students.

Kevin: So I've asked her to do just that. I've asked Ms. Thompson, Can you explain the following tweet? Again, short timelines and slow take off will be a pretty good call, I think, but the way people define the start of the take off may make it seem otherwise. Simple little tweet, but there's a lot to unpack there, and we're gonna hope that Mrs.

Kevin: Thompson does it, and maybe we'll shoot some spitballs at her while she's thinking! Spoo!

Gavin: Oh good. All right. It's 1940s,

Kevin: football through my

Gavin: Now it's probably doing something with a phone, but at the same point, you know,

Gavin: wonderful old days. Yes, 

Mrs. Thompson: Behave yourselves, gentlemen, and pay close attention. Imagine you're waiting for a giant beanstalk to grow, like in the tale of Jack and the Beanstalk.

Mrs. Thompson: Now, Mr. Altman is talking [00:48:00] about the time it takes for the beanstalk, or technology in this case, to grow tall and strong. He believes that the beanstalk will grow quickly, short timelines, but will start growing slowly, slow take off. 

Mrs. Thompson: However, people might be looking at the tiny sprout and thinking it's not growing at all, because they're not sure exactly when to start watching it, the way people define the start of the take off.

Mrs. Thompson: So, in simpler terms, the technology will grow quickly overall, but it will start off growing slowly, and people might miss the beginning of the growth, if they're not paying attention. Do we understand, or shall I fetch the dunce 

Kevin: cap? Wow. Wow, that's not fair, Miss Thompson.

Gavin: not that far.

Kevin: I was also playing games on my Nokia.

Kevin: I was playing Snake. I missed a lot of that. I'm

Kevin: sorry. 

Gavin: I will say we could [00:49:00] probably speed Mrs. Thompson's responses up a little bit. She might be on, something's going on with Mrs. Thompson. 

Kevin: you want me to throw her a monster milk?

Gavin: I don't think we could give her a monster milk, probably.

Kevin: whoop.

Gavin: Okay, let's ask one more question and then we'll end the show on throwing her a monster milk. Okay, what I want to ask her, so you've heard her talk about the idea of like fast takeoff, slow takeoff. So in the AI world, A slow takeoff is the idea that it will not immediately get to a crazy place really, really fast and that slow takeoff will allow us to have a better understanding how things work.

Kevin: We can adjust our daily lives, our economies, our artistic endeavors. We can adjust everything slowly as the technology takes off.

Gavin: So what I want to ask her further, because this is the thing I'm trying to understand even myself as an adult who follows this stuff. Is what it seems like Sam Altman is saying in this tweet is that the takeoff has happened already and most people just don't notice it. So I would ask her, does [00:50:00] she feel like this tweet is saying that we have already seen the takeoff begin? 

Mrs. Thompson: Behave yourselves, gentlemen, and focus on the matter at hand.

Gavin: we're here.

Mrs. Thompson: It does seem like Mr. Altman is hinting that the takeoff or the growth of this technological bean stalk has quietly begun. Unbeknownst to many people are perhaps gazing elsewhere, missing the initial slow sprouting of the stalk.

Mrs. Thompson: Now, if the takeoff has indeed started, it means the beans stalk is on its ascent. Even though it's growing slowly now, it will soon shoot up, reaching towards the clouds swiftly. This means we must prepare. Just as Jack had to face giants when he climbed the beanstalk, we too might encounter unforeseen challenges and opportunities as technology advances.

Mrs. Thompson: We gotta fight giants? It's imperative to keep a watchful eye on the growth. Adapt and learn as we [00:51:00] ascend the beanstalk of technology. Mrs. Thompson! Ensuring we are ready to face whatever giants lie in wait at the top. Mrs. Thompson, hold on! Are we clear? Or do I need to spell it out with alphabet blocks?

Mrs. Thompson: Mrs. 

Kevin: Thompson,

Gavin: Man, Mrs. Thompson, what is going 

Kevin: that a subtle hint? Do I need to spell it out with alphabet? Alphabet, parent company,

Gavin: Oh, damn.

Kevin: bard, and Gemini?

Gavin: What?

Kevin: is A G I? Ms. Thompson just confirmed it!

Gavin: Amazing. Okay. Well, you get a sense of what you can do with these characters when you create them on open AI, uh, which actually BT or really any of the LLMs, you have to tune them properly. You have to get a sense of it. And also like, just be aware. There's a whole vocabulary that if you're out there and you're kind of new to this stuff that you don't have to know right away, but things like slow takeoff and fast takeoff, learn about what they mean, , ultimately, I think what he's trying to say here, just my definition of this is that clearly that we've already started this takeoff in his mind, which is interesting because , I think a lot of people would be like, oh, [00:52:00] I don't believe it, but takeoff means that we're on the way to artificial general intelligence and that it might be slow, but it's already begun.

Gavin: And it probably began. And. Maybe years ago at this point, for all we know, so I think that's an important thing to know. Um, before we leave Mrs. Thompson, should we... 

Kevin: I have a gift for teacher, Gavin. I don't know if you 

Gavin: It's a bottle of monster milk.

Kevin: it's twelve cans. of monster milk 

Kevin: that 

Gavin: Oh no, oh no. How is she gonna, how are we gonna give her 12 cans? How is she gonna drink 12 cans?

Kevin: I'm saying it right now. I mean, I could... 

Kevin: So, Gavin, as a sweet gift for Miss Thompson, I went ahead and gave her twelve cans of Monster

Gavin: Oh, Miss Thompson.

Kevin: and told her it was a pleasure to have her here, we really appreciate it, and I wanted to give her the opportunity... With a fresh surge of monster milk in her veins to tell all of our delightful listeners and viewers To engage to like and subscribe and help us out because now miss Thompson is clearly part of the AI for humans

Gavin: Yeah, of [00:53:00] course. She's a lovely old lady. I'm really excited that she's here.

Kevin: that's right.

Kevin: Oh a whoop and a whoop to miss Thompson and to you Gavin Let's see what she has to say after 12 delicious cans Of Sheila's monster milk a little part of you feels bad when you do this to a character you just created, right?

Gavin: Absolutely, I'm terrified with what she's gonna say 

Kevin: This is like melting a G. I. Joe with a Bic lighter. This is not cool.

Mrs. Thompson: Oh, sweet pickles and jam. Whoop, whoop! The monster milk is bubbling in my veins like a cauldron of frothy frogs. Okay. Kevin Gavin, are we riding on a biscuit soaring through a jellybean jungle? Thank you for the time, or is it a rhinoceros? Whoop, whoop! Oh, dear audience, like and subscribe, or the marshmallow mice will nibble your night shoes.

Mrs. Thompson: Engage on the social platforms. Do the twist and shout. Make your fingers dance like spaghetti noodles in the wind. Yeah![00:54:00] 

Kevin: Whoop, whoop. 

Mrs. Thompson: Is it a plane? No, it's a parade of pancake penguins pirouetting on a peanut. Whoop, whoop. Let's all hop on the carousel of cosmic cucumbers and sail to the moon made of muffins.

Mrs. Thompson: Whoop, whoop. The world is a whirlwind of whizzing walnuts. 

Kevin: Whoop, whoop. Whoop, whoop indeed, Miss

Gavin: whoop. 

Kevin: you. 

Gavin: Now I understand that Sheila's whoop, whoop wasn't coming from some internal need to celebrate. It was actually the monster milk all along that causes the whoop. It's

Kevin: It'll whoop 

Gavin: symptoms. 

Kevin: Dear audience, I do want to reiterate, you gotta like or subscribe or else the marshmallow mice will nibble your night shoes.

Gavin: Yeah, everybody. Thank you so much for listening to another episode of AI for humans. And yes, as Ms. Thompson said. Monster milk or non, please subscribe. Please like us on all platforms. We are on YouTube We are on all podcast platforms.

Gavin: Leave us a five star view at Apple podcasts. [00:55:00] It really does make a difference for us Please spread the word tell people about it one of the interesting things we found about audio podcasts is you really have to go out Of your way to make sure people listen to them and we love the audio We, we are actually, I'm a heavy consumer podcast myself, and I want more people to know about the audio version of this show as well.

Gavin: So please go share it, tell people, find us on TikTok. Kevin made an amazing TikTok the other day about the Tesla's new autonomous robot. So there is stuff that you're not seeing on the show that's only on the socials. So make sure you go pick up there. And thanks again for coming here. We really appreciate having you and it getting to know everybody in our audience.

Kevin: , I do want to shout out Kuran's Music. It was one of the courageous listeners or viewers out there that left us a five star review on Apple Podcasts.

Kevin: Says, Good guys, Kevin and Gavin, have given me and the world a wonderful podcast. Been delaying my review. Interesting. He's been withholding. But the latest GoodGuyAI bit had me dying. Thanks for the wonderful infusion of humor and knowledge. I disagree with GoodGuyAI being a bit [00:56:00] worthy of coming off the sidelines for, but I appreciate it, Karan's Music.

Kevin: Thank you so much for that review.

Gavin: Thank you so much for tuning in. We really appreciate it. And yes, we will see you all next week. Thanks so much.

Kevin: Bye, everyone. 

Yeah. Yeah. Yeah.