AI Revolution LIVE: Analyzing Tech Stocks, Market Volatility, and Political Risk

Go.

Welcome to the AI revolution. I'm Cody Willard. We got Bryce Smith, works with me at the hedge fund. We got one of my dear old friends, John de Oliveira, who has been working in the AI industry since about two years after I met him, which I met him in 1994. So maybe since 96 about this. When I moved to Wall Street, I bought a one way ticket from Albuquerque to New York the day I finished my last class. I didn't stick around for graduation at college or anything. I got the hell out of there.

Cody, probably everybody doesn't know Cody was in my wedding. So that's a, you know, we're, we're.

Very close and dear friends. And John, what year did you move to Austin to start working for psych AI?

98.

So yeah, I had just. I'd been. I moved to New York in 96. John started working at Psych AI, which was the. One of the trailblazers and trying to monetize AI. I mean maybe just trying to make it make sense, make it work.

I'd say trying to monetize and not that being its strong point, but being like if you want to go like compared to folks who say that they're in it for research, they were really in it for research.

So what was your boss's name? Legendary. The legendary boss and founder of. Yeah. So anyway, we got a. A packed show today. There's. That sounded like I was reading a transcript from Fox Business Days. But no, we actually have a lot of cool stuff to go over there. Bryce and I have been working on some AI revolution apps. My daughter has gotten into app coding. My next door neighbor here at my office building is the NRCS National. Right. What's NRCS stand for?

And Nature and Resource Conservation Service. Natural. Natural Resource Conversation Service. It's a, It's a department of the usda.

You guys can ask AI what NRCS stands for and what department it's a part of as the bureaucracy or the federal government. Anyway, they do rant. They like, you know, it's like ranching and farming and you know, like support and they. He made an web. He used AI to make a web app last week and shared it with me. And it's incredible. And they're already using it at least locally here at the nrcs, so. And then there's a few. Oh, we got questions. Yeah, chime in with if you're watching on, you know, social media.

Hey, I have questions.

I know you. No.

It'S try. It's totally about what you're talking about. So like, I just want to know like, because what I get hung up on is, is like hosting how do you, where does this stuff kind of go and like run from?

How do you go from. I'm, I'm, I'm like previewing today's agenda for the show and you're like jump forward to like deep in the weeds, like trying to figure out the soil content. What kind of a question was that? To get started? We will address all that at the appropriate moment after we've shown you how amazing our apps are.

Good thing.

Your question is we paid up to google 350 bucks a month starting after I used $300 worth of crush credits and tokens and stuff in the AI studio and AI and Google cloud realm. Bryce, did I mess any of that answer up? We will come back and address this in more detail after we get it all introduced.

Johnny D. We've been, there's a few ways we've done it now, John. But Cody has spent most of his time in Google AI studio and that's where I was spending most my time too. And if you do it through there, it's all through Google's cloud platform.

So. And you know, they, they're, they gave us, there's like a, you know, promotion code thing. Not code, but you like, you sign. You know when you do it the first time. They're like, do it now. 350 bucks a month starting after you use 300 of credit. Which I thought, hey, I don't even know how much $300 of credit might be if. Or like if it's a day or a month or six months worth of token. Token using which Bryce also you got to make sure on the agenda today, if you'll just make a note in front of you and don't let, don't let us forget. Are you going to use pen and pad.

For note taking?

If it's clearly there's still like at this moment, it's just better to do it that way.

Yeah, I, I still like the paper note. I do the Gemini notes too for all of our meetings now where the Gemini takes notes and transcribes the call. But that's nice to have like so you have Gemini running the background taking detail notes. But if there's something like really high level that you're like, oh. Or you know, you have some kind of aha moment, like actually go ahead and write that down in your notes. It just does something for your brain, I think.

And all it even just helps you recall it later. I mean so, all right, anyway, let's, let's, let's jump in. I'm gonna, like, show John. Did I show you guys an app last week at all?

Yeah, I don't think you showed any apps. We showed some of my apps.

All right, so here's. Yeah, I'm going to start at what my daughter had me build the last two days. It's the most rudimentary, but also still amazing web app. We, we went to AI Studio. This is AI Studio. Oh, I have to share my screens first.

And, and initial context is that Cody nor his daughter are coders or software developers or have. I mean, Cody's built websites, I suppose, but other than that, basically no. No software experience.

Yeah, I've done some HTML coding, but very basic and. Oh, I've done some basic coding, actually. 10 coding.

Don't call HTML coding. It gets. It pisses off.

But can I call basic coding? Yes, because I literally type 10 print quote, C and then 20 print quote space O. And then you get down to the bottom, you're like 50, go to 10, and you could see Cody stream across the screen. This was when I was maybe 12 years old. It was not quite as deep as. I also took Fortran in college. The point of the story is that I have no idea how to code. And we were just like, we go. I signed into AI Studio up, see the URL there?

We can't see the URL.

Okay, AI Studio just again, Google it. It's. It's the future Googling things. You'll hear the term forever. All right, so anyway, you can like talk to it and then like, you know, you just. It's like chat about how to build AI Studio apps and stuff. But I've. Bryce showed me. Go down here to build and then, you know, you can type what you, you know, give it a prompt and it helps you get started. So my daughter is into crocheting and.

Why don't you go to the real URL we made?

Well, I will, but I just, I wanted to show people how this part sort of looks. So we just went in here and it's like make an. An. A web app that is a community for crochet people. And it started off with like two or three tabs, and then we just kept refining it and adding features. Eventually went to GoDaddy because Google doesn't sell URLs anymore. They sold that business to Squarespace. We went to Google and for five bucks a year, we bought CRO. And by the way, the AI made the logo, the name Everything we bought Crochet Dot Crochet Crush Club, and this is what it looks like when you show up. I log out so you can see what it looks.

We're not seeing that task code and.

The, the page sort of covered over.

All right, all right. Yeah, yeah, fine. That's fine.

Allow this.

Just bear with me. Remove. Let me present Stop Screen Share.

Like, I'm looking away, but I'm looking at my big screen.

That's one of the things you learn in. In law law school, John, is when you do your online arguments, you have to look at the camera and not at the screen.

Deep thoughts. All right, here we go. Here we go. Where's Crochet? Crochet Crush Club. Do you guys see this now?

Yes. Start crafting now.

All right, so it made the logo, it made the name. We bought the URL for five bucks a year. Bryce went in and pointed our web app to the URL after we spent some time on it. And so, like, you can see here, there's a tab across the top. There's eight different. This is the same as the tabs across the top. The. There are eight different things. Pages to use on here. Just get. Let's start. Let's get started. The AI Companion can select a home decor item. You're an intermediate user and you want cute Kawaii. As my daughter told me, it's like Hawaii, dad, but with a K. And it means cute in crochet. Apparently it does.

And if you leave one eye off of it, it means scary.

How do you know Hawaiian words now, John?

Not. It's Japanese. She's saying it's like Hawaii, but the word is Japanese.

Oh, that's my bad.

Yeah.

And you do know if you. A little bit of Japanese. This is not your first time pulling the Japanese card out on us.

Yes.

Yeah. So Cloud 9, these are what suggestions it has for you. And then you can click on one of them and it'll.

Wait. Did it just generate that? It made it up.

Yeah. AI generated this entire thing was AI.

Generated the pictures, the text, everything, right?

Yep.

And the instructions for how to.

Instructions for how to even the layout of all of this.

It's like it makes up a picture of, like, this thing that never existed. It's like, here's this crochet coaster, and then it makes up the actual real instructions that would actually result in that particular coaster existing in the world. Is that crazy?

That's that. Thank you. I, I, that's what I, I, that's how I feel about it. Like, I'M crazy. All right. And so then we were like, how about teaching people? I was like, lincoln, what else do you need? She's like, you know, like, what? Different. You need to know how to do all the different stitching. So we added a tab for this, and it came up with the name Stitchinary. Great name.

Oh, cool.

And then it tells you how to do it. Cool. Then you come over here. You can pick a color lab. And John, pick like a theme or a vibe.

It's a candy cane.

And.

Notice at the bottom left. My daughter wanted a mascot, which. It's basically the old paper clip for Microsoft Windows 98. But you can click on it and he'll.

I thought it looked like the old GoDaddy logo. Remember that?

Yeah, that too.

Yeah.

But anyway, look, it. You know, it tells you what you're doing on the page.

Oh, wow.

And now it gave you the candy cane things. Oh, yeah. By the way. So look, if you log in.

It has accounts. You can log into it.

Yes. I told it I wanted accounts and a social. You know, like a social community aspect to it. And then look, now and then I was like, hey, you should be able to save the palettes you create. I forgot. We could have. When we did this AI companion thing, we could have saved it also. Dream Machine. John, describe something you think someone should crochet.

A cube.

That is cube pillow. How about a Rubik's?

Oh, even better. I. But yeah.

All right, come on, come on.

I mean, come on.

Oops. Yeah, yeah, no, close. Right now we don't need to check our passwords. So look, you click on. You, like, create the pattern. That's adorable. For the rat. The adorbs, as my daughter would say. For the wreck.

Cody. Cody, Pull in on the Gen Z.

Oh, I've been trying so hard. It annoys my daughter so badly when I'm like, oh, that's us.

It's.

Anyway, so give it a second.

Say, do you know which Gemini models it's running specifically? Did you tell it?

I did not, Bryce.

I think it. I think it typically defaults to 2.0 flash.

I would think that's right. So, anyway, look, it's. Again, full instructions and the image and colors and everything.

When it. When you're making the app and you're saying, like, I want, you know, user accounts or something. Was it asking you like, oh, by the way, do you want user accounts? Or were you having to tell it everything up front?

No, I.

Not everything. Some of it. It came up with it on its own, but I Did specifically tell it, I want user accounts. And then I said, I want to be able to upload the log. I'd like to change your logo. You know, change the profile.

Yeah, thank you.

I want to be able to click through and see what my save palettes or patterns or shared projects are.

Oh, but also, you could be incremental. Like, you could tell it some things and it does something and you're like, oh, go back. I want this other thing too correct 100%.

And you can, you know, along the way, sometimes Bryce taught me that when you're talking to it, this over here at the code assistant, where you're typing in what you're asking it to do and everything, it can start to get confused if it has too many discussions going on and too much track record there. So you hit refresh, you sort of start over, and you're like, okay, thanks for doing that. Now let's. And then, you know, you don't think it. You just move on. And then you go to the next thing. But you clear the clearance pallet from.

Where you left off. You get to, you know, not starting over from scratch.

No, along the way over top right, you save it, copy it, download it, and here's how you deploy it to the cloud. We'll talk more about that in a second. But that's how we got it to work on the, you know, not just inside the AI studio, but actually on in the web thing. All right, so anyway, there, I just want to show you one last thing. At the end, I was like, hey, name one other feature that we could add to this web app that would blow people's minds. And it came up with the idea for a snap and stitch for your Take a photo, analyze the photo, and then if you want it realistically or kawaii, you choose. And it literally. Watch this. I just. I think it's just about the most amazing thing I've seen this week in the AI revolution. Come on, come on, come on, baby, now.

And while it's doing that, so the other thing you do is like, even just save picture of the thing you do create. You're like, this is one in the real that I made.

Right? So that's in the community page and you can upload your stitching and get an AI for it. So it literally walks you through each step to make it.

It's incredible.

It's incredible.

I give you. You get extra, extra points for integrating image capture in your website because even I hadn't thought of.

But again, before, that was AI's idea like I told I gave it the prompt. I need an amazing feature that using AI generation. I did. That was the prompt said I want an amazing feature for this web app using AI generation that uses AI generation that will blow people's minds. And it gave me the step and stitch idea. All right, next Bryce, if you'll pull up the Revolution AI website that you built while he's pulling that up last week.

It's that one.

Yeah, I shared it with you.

Oh that's right. That wasn't on, that was not on.

The call personal level. Bryce was so proud of his new web built. Again. I've built many websites over the years. Bryce has never built a website, a web app, anything. He's never taken a programming class, none of that.

So the, the idea, the reason we made this website is that Cody and I have now built a handful of these different AI applications with the AI Studio and we're like, well we need a place to aggregate all of the AI tools we've made and put them on the, you know, central place where people can find them on the web. So that's, that was the, the impetus for this. So we came up, you know, same thing, bought the Revolution Investing AI and I actually use a tool called Bolt New to create this website. And the reason it's, it's pretty cool is because it automatically integrates your, your back end and does a lot of other extra like functionality stuff that the Google AI Studio doesn't have yet. So you know I had a friend recommend it so I wanted to try it out but anyways we don't have to go through everything but you know it's got our kind of explains the philosophy of Revolution Investing. It made this based on a PowerPoint that we use you know, as part of our pitch. But then here's the kind of, the real me is these tools that, that Cody and I have built, we, this.

Is the web apps just AI generated web apps vibe coded through the AI Studio that we just showed you.

So this was part of a presentation we gave in Vail at an investment conference about Google. And so you know we had, we had already built this whole thing but I think we've, we've shown this on the call before the illusion call but anyways but now it's, you know we just embedded the iframe in this website so you can access the, the Google model here and it's got our other tools here. It's got a built in chat bot so you can ask questions about Revolution investing, you know, how Do I find revolutionary companies. And it will tell you. Based on Cody's, you know, framework for revolution investing. I integrated it with Cody's YouTube channel, where we are right now. And so it's got the. The real API to all of our.

Hi, John. There's John.

Yeah. And you know, put our bios on here. You know, did another API with our trading with Cody comic kind of theme.

That you did there, where you turned the picture into a comic. What is that?

Yes, I use. I use Chad GBT4O model to like make these thumbnails. Yeah.

What kind of thing is that? I mean, you got to tell it a certain sort of theme or.

Yeah, yeah. So, I mean, what. It. It just. It looks so much better to have a thumbnail. If you're a real YouTuber, you have to have a thumbnail, not just your stills of the video.

You know, maybe you'll do it to this revolution call sometime.

No, I did. That's how.

That's what these are done.

That's last week's Revolution. Yeah. So that's what. Like if you click. I don't. It'll be double playing here. But if you click this. I don't know.

Is it like that on YouTube now too? Is that the thumbnail?

Yeah, yeah, yeah, yeah. That was your idea a long time ago. Was going and change those. The pictures in the description. So that way, if people look at.

That AI generated face of Bryce being angry. Bryce doesn't actually have whatever that angsty anger, disgust. Yeah, that look right there. He. He. He doesn't actually have that in his emotional repertoire.

So let's. Let's.

So, all right, so Bryce, now let me show that real quick. This two. Two other apps I want to show real quick.

And then I had a question for Bryce. You said alt new. I didn't see it. I went there. It didn't. It wasn't a thing.

No. Bolt. B O L, T. Oh, bolt.

I know, Bolt. Okay. Okay, good.

All right. Share screen. One of the apps that we made this week. Or I. That I made actually all last week, I spent this whole time. And it's the car. It's the idea of showing the accelerating pace of innovation over time. And you can click on these things and again, I still need to improve this. That's what I need to go back. I haven't worked on this since last Wednesday, but this afternoon or tonight or tomorrow or something, I'll try to go in for an hour or two and vibe code with the AI Stuff like, hey, I want this when you Click on it to actually get completely big. And the the dates on this one actually go back to 3500 BC which is the advent of using the wheel writing and horse. The domestication of the horse. 3500 BC. So that's where I started my timeline. And some of them only go to one AD things like that. But if you'll notice what you can do here is you push play and it shows you that for hundreds of years you see the timeline going by at the bottom. Okay. Finally we get to this stuff. Okay. Oh we get. Wait we got more. And then bang in like the last 50 years. Specifically since the advent of. I'm gonna have to zoom in a little. Let me just rewind. Since electricity is really when it all started accelerating. But the applications of electricity have accelerated over time. And then if you click on any one of these icons it shows you you know who is the major key figure where it originated really got the harnessing of electricity done mostly in London I guess initially Faraday of course was in London. Societal adoption over time you can see it's basically 100% these days. The growth.000. And then we finally start in the 1800s seeing this accelerate. And we went from zero people having electricity to almost 7 billion people having electricity. You can click on this button to get a list of industries and names that are winners and losers and automatically it generates. Oh well since did it fast enough you can see winners people who sell electricity stuff and losers like people who sold ways of generating light or heat for electricity.

You just can't get a job as a lamplighter anymore.

You really can't. The well oiled industry has been in a secular decline for 170 years now.

Cody, what's the star? The kind of star diagram at the beginning. What do the different axes represent?

Okay so it's just the left. This one is transportation. And if you look it's horse drawn carriage, sailing, Roman roads. The magnetic compass. So it's like a sector or type of appearance medicine. Oh look it actually see this is the type of thing I need to go back and fix. This should say entertainment. This one should say medicine.

And then outer to inner is time.

Passage on the left and bottom are time is if you'll notice when I.

Go back they all build their way toward the middle.

Yeah, they all come towards the middle.

What's the right then if it doesn't also come in like that.

Say that again.

That's after like electricity has been invented. And then it goes from there right.

Into what Other the same ones, but continuing along. So there's really just.

Yeah, yeah, yeah, there's. It. Nothing changes through the middle. Except I just sort of wanted to organize the time horizon around the advent of electricity because I think it is the single most important application that's ever been invented. The. The biggest platform that's ever been invented in mankind's history. Other, like maybe the wheel is bigger than that. And I do think. Let me just finish, John. And then I do think AI revolution is like the next nexus of all applications and services.

So like top, like northwest to southeast on the thing is all the same one axis. No, northwest, other northwest. Yeah, up there. From there all the way straight down to Florida.

Yeah. So look. Yeah, it's board. This is. This should say entertainment, not medicine.

Okay.

Board games, theater. Okay. Theater came along at, you know, eight in the early hundreds and then. Oh, I mean, I went too fast. So I'll just go all the way. Look. So it goes there to musical notation, playing cards, photography. Electricity comes along right after electric, after photography and then you get to the phonograph, motion pictures, electric guitar. And if you'll notice, this is when the pace picks up. Right. There's so many more icons because the acceleration of the innovation in the entertainment industry. Video games, personal music players, compact discs, CGI and film, digital music players, MP3 players, online multiplayers, podcasting, video streaming, gaming, streaming virtual reality headsets, and all roads, by the way, lead to AI. AI, AI changes medicine, it changes transportation, it changes entertainment, changes medicine. The only one that maybe it doesn't really change is communication, except that maybe it just makes us communicate so much less because we're talking and engaging with AI instead of each other.

I don't know. It's writing our emails, you know.

Exactly. And answering other AI written emails and books even. So you get the idea. Yeah. And by the way, in communication there is the printing press that comes along here right after the postal service, sign language, optical semaphore, telegraph, on the way to fax machines, fiber optics, void.

Time wise, it's like a reverse, what do they call like a log scale.

So, yeah, inverse log scale.

Thousands of years at first in this, this amount of space here, and then you're showing only 100 years across this whole other space here. So it's gotten much. It's like zoomed in and going, here's all this stuff happening in a much quicker piece of time.

That's right. All right. And then finally here is the app that the guy next door. I'd be happy to give you guys, his name. I don't know if he wants me to. I should have asked first. But he built this again using. He saw Bryce. What did I tell you? He saw a replit. It's up here. Replit on the CEO of Replit. It was on Joe Rogan, apparently. And he heard it, and he was like, I want to check that out because I have all these ranchers that have to fill out all this paperwork and stuff. And he showed me. He took a picture of the paperwork, which is, you know, one, two pages, and the ranchers have to go in and fill all this out. And he put it. He took a picture and told the replic to build this. And now this is an example. And you can, like, add an assessment. You can look at your map view, zoom in and out. You can look at the soil data on that map view. You can draw your pasture on the map view. I'm trying to go back. He needs a button to go back. That would be good grazing plan livestock, add a new herd. It. It's just incredible. Is it not? Again, like, there's some sort of look you can. You use AI to recognize the. Oh, come on.

I've done that.

Dude, that's sick.

Right?

You guys with me? That's crazy.

Absolutely.

Like, it changes everything for the rancher. Instead of sitting there doing some archaic paperwork, he's literally, like, walking through his field using a smartphone or even smart glasses, and he's like, hey, take a picture. And then he can pull it up to the. Load it up to the A.I. oh, that's. Bryce named some random weed that grows in carazozo.

Broom snakeweed.

Broom snakeweed. John's not going to tell you what broom snakeweed is. He can.

I know, but I'm not going to tell you.

There you go. It's just not nice. Now the guy knows more about more things than everything you'll ever meet, but I don't think he knows any. He's terrible about mascots for professional sports teams. Any. I doubt he knows a random. What did you call it? Broom snakeweed.

Yeah.

So last month I was on a walk with my wife and. And we saw this magnolia tree and this other thing next to it, and she said, yeah, that's a magnolia. That's not. I said, yeah, it is. And it says, no, no, that one is. That's not right. So. So I took a leaf from each of them and I. By the way, I could be getting a story a little bit wrong about who said what. But anyway, put them down next to each Other took a picture of it and gave it to ChatGPT and it said they're not the same but this is like another. It's like a cousin of the magnolia or a different version of it. I said why are you saying that? This one's shinier but it still has the curviness to the outside of the leaf and it's like going through all this stuff and it's like ridiculous. Amazing.

Jensen Wong the other day was talking in an interview about he was refuting the idea that people using AI getting stupider and that's the case in point. It's like I've, I, We've put chat, GPT and Gemini on my 11 year old's phone. I am happy for her to use AI to learn and talk about anything she's using. What's the language Pro duo lingo with AI to learn Spanish. She uses AI some app to AI generated app with her sign language. It's incredible. Let's hit some questions real quick. Hi Cody, could you give me your views on Xerox? Xrx's Xerox is old. Is Xerox still publicly traded? Yeah, yeah.

So it's one of, it's one of icon's biggest holdings.

My wife was going through the Alphabet singing a song to my daughter Ramirez who doesn't walk or talk and just loves to hear us sing and stuff. And she was going through and doing A is for Apple, B is for and she went through and she knew a stock for every letter until she got to X and I was like Xerox or U S still. So anyway, rxrx Bryce, you know rxrx?

Yeah, it's recursion. Pharmaceuticals, biotech.

Yeah, I do. I have no opinion on it. I don't know. Rxrx, sorry Applied Bryce, what's the story with apply? What's a ply? Digital is the company they do what.

So they were bitcoin mining but now they are high performance computing and they.

Do like lending and finance. The answer, the short answer to your question is we're not fans of Applied. If you want to play bitcoin then buy bitcoin and you want to invest in gold, I say buy gold, not gold miners. There's too much execution risk, potential fraud. Who knows what their motivations are. I just. If you want to own, if you're. I don't ever invest in companies that mine commodities and bitcoin is a pure commodity in the sense that anyone can mine it. It's easier to mine Bitcoin. Than to mine gold.

What do you think about his comment that both have received Nvidia money?

Yeah, I don't know about the rxrx. That would actually sound sort of cool that Nvidia is like hey, you guys are at least promising enough that we should like help you use an AI applications to make your drug get to market or work out better, faster, whatever it might be. Yeah, applied. I don't know why Nvidia messed with it. I forgot about that actually. But yeah, one an AI drug company and the other provides AI infrastructure apparently. Man, I'm going with math at least on the second one I'll go with don't know RXRX at all. Sorry. And applied man. Just waiting to hear everyone's thoughts on Croc 4. Let's defer to our AI expert, John.

Yeah, so I mean it's, you know, it's passing all the tests. It's considered the best AI there is from every benchmark standpoint. Now when people get down and do their own kind of testing on it doesn't yet do quite as well on some of those. So I don't know whether it's been kind of working to the test and it just kind of. First of all, congratulations to the XAI team. Clearly, you know, a tremendous advance. But I mean you look at it from the AGI measurements, you know you get from what the highest score on an like I think on RKGI was like a 25 and now from Grok it's like a 50 given, you know, whether it's using reasoning, so forth. But it's clearly a big advance. Is it just another step in the leapfrogging? Like are we going to see the next.

Doesn't it feel like that, John? I mean that was well phrased.

I.

It feels like we're just getting, which I mean I don't mean to downplay these. And steps are advancing incredibly fast. Every couple three months there's this like leapfrogging, better LLM, better video, better model that someone has. You know usually there's about eight or nine different companies that are sort of leapfrogging each other. But it feels like that, right? It's just relentlessly leapfrogging.

Isn't there sort of an analogy to like semiconductor industry and the Moore's Law and you know, at one point in time intel was fighting with others and AMD and TSMC and trying to get the most advanced node and TSMC has sort of ran away with that now. But I mean it is like that Process like you, you can't just go from like 18 nanometer to 2 nanometer, right? You have to go through like, it's like this consistent pace. Every two years it, you know, gets cut in half or something. And so, you know, at the end of the day, if you make really good semiconductors and you're one of those companies driving, you know, more advanced process nodes, you're probably going to do all right as.

Except, you know, the it is, it does step through things in order to get somewhere, but it does it at a faster pace now than, than that example. So you know, one of the things, one of the kind of demos these different companies or whatever run these different projects run is to do some make a game happen, you know, describe in a game. And there's a real leap ahead with Grok4 because you know, you're seeing all these games and they kind of looked like what's the Mine. Minecraft. Minecraft. I forget the name. You know, you seem like things that look very blocky like Minecraft. And now Grok4 actually like went out to the web, found textures that are things that you kind of wrap around 3D objects to make them look more realistic and stuff, and made a game that just looks way better than you were seeing with these games. Like you look at it and people are starting to say give it one year and people will be making games that people would pay for just by describing them in with a prompt. Yeah.

Did you? I. This is part of probably what you're referencing. Elon commented on this guy who was like, hey look, I made this video game with just a few simple prompts using Grock4 and it really looks sort of like Doom. And there's seen a couple other ones that were just also equally impressive for just again, like imagine unleashing the random brilliant ideas for video games that random people in apartments and trailers around the world have in their head, but never had anywhere near the kind of access to talent or capital to build an actual video game and movie or whatever, you know, like high quality new eventually ar AR video game, not just a web video game. This stuff is going to happen at an increasingly faster pace. And by the way, that's why we're so bullish on inference for AI. The relentless demand for these incredible ideas to have the processing power behind making them work is just exponential beyond any of our capacity to comprehend.

Yeah, on. On that. One of the things I told Cody as we were building these websites this last week is Jensen and others have talked about how reasoning models are like resulted in this step change in token demand tokens. Or if you're not in the lingo, it's basically words, snippets of code. It's almost the equivalent of the like having a word count. And so, you know, if you're, if you're just one time. I did an experiment about six months ago. I wanted to see what the average token response was. Like, how many tokens did it take for just a typical chat with a chatgpt chatbot. And I did a bunch of tests across different models and the average came out to like, about, like, I think it was like, I, I came up with like 4,000 per day per user. You know, the average number of prompts you ask multiplied by the number of tokens per prompt that are needed. And building these websites with these reasoning models, I, I burned through 12 or no, like 20 million. I burned through 10 million tokens and I had it up again. So I went through like 20 million tokens building two websites. Right. And so clearly it is a order of two orders of magnitude larger, three orders of magnitude larger than my original. And but what I realized is that the reasoning models, it's not just that you're getting smarter responses, is that the reasoning models actually allow you to do useful work with them. Right. And so you, you need all those millions of tokens to build a website, not to get a really, really long, like deep research report. And I think most people on Wall street were just thinking, okay, now you can get like really smart answers. But it's actually all of that behind the scene token usage.

And this is where it's like the, the distribution of AI itself is exponential too. So whether it's Grok coming to Teslas now.

Yeah.

Or it's my daughter building a crochet website that has AI generated images and features in it. Like all of it is this like, compounding thing that you're using it to build the stuff. The stuff that you're building is using it. The stuff that is using it is being distributed ever more widely. And it's the democratization of the app revolution. And I just didn't realize just how powerful that really is. Because there's an app for everything, but not really. There wasn't an NRCS app for the random rural farmer in southeastern New Mexico until today, but now there is. Like, there wasn't an app that for teens who love to do crochet and want to be a part of a club, but today there is, and that uses AI generation. There wasn't distribute. There wasn't you know like the idea that now instantly we've talked about how Google has 2 billion Android devices out.

There and they have 3 billion Android devices. They have set seven different apps that have more than 2 million monthly. 2 billion users, 2 billion monthly actives.

And you again like instantly now Google Gemini is going to turn on in every one of those distribution points. And so the amount of capacity that it's going, the amount of infrastructure to do this stuff is obviously we're not even close. It's a hundred times more, A thousand times more inference capacity is needed over the next two to three, five years. Electricity is just to hit off that question. Electricity will have exponential growth in demand but it's like not as like hockey sticked as actual inference processing is going to be needed. And likewise you can mine electricity, it's a pure commodity. You can mine it with nuclear, you can mine it with solar, you can mine it with fossil fuels. And so I just set off the question. Stick with the actual AI infrastructure providers, not the electricity power generators, copper wires. Like the, the, the place where the differentiation happen is where it happens is where I want to be investing. John, you are going to say something always.

So what's the name of the XAI cluster in Memphis? Colossus. Yeah, I think Colossus. Anyway, a hundred thousand processors I think and he wants to hit a million, he wants to get to a million processors by the end of the year. That's 10 times in a little more than or in just a half a year. That's crazy.

Yeah, it really is. And then so much of the focus is on the training chips because that's the most cutting edge and it's exciting and that we need exponential growth and deployment of the ever, you know, whatever. Nvidia's best training, best, most advanced chip is we need that. And it's going to happen on basically a yearly basis if all goes according to plan here. But the fact is that all those chips that were used for training are now used for inference. And that is where the meat of like the expansion is going to happen.

Now if you test time it's the same, same thing. And inference really the part like while you're, while you're using it, what's, what's the deal kind of thing. So another is that hear a lot about lately is like we're gonna have our first like one person unicorn in the next whatever amount of time.

So the unicorn being a company worth a billion dollars.

Yeah, one person's gonna make a company worth a billion dollars with A single employee. Now on the way to that, you can already see these companies have what, like 50 people and they're. And they hit a billion. And they hit it in like 14.

Months or it's a good segue into the next question on the board. John Cognition this afternoon bought basically like the, the stranded assets and some of the ip, I guess, of Wind Surf because Wind Surf CEO and most of its main people were hired slash Aqua hired slash bought by Google after Open AI had tried to buy them. But Microsoft, Open AI's partner nixed the deal because they wouldn't share IP with Wind Surf. And it is the most dramatic turn. I mean, and again, it's like it happens on AI speed or something. Like this stuff happens in three weeks. This was like three weeks of drama, talking about billions of dollars. And how many people even work at this wind surf thing? 12, 20?

I don't know, Brian, I just looked it up. It says Google saying 273 employees.

All right, well, it's bigger than I thought. But still it's not one person. But John, it's 270 people with like a 8 billion dollar valuation. 3 billion.

They're largely in Austin because I heard that they have a chunk of them are people that worked at Trend Kite, where I worked, that are now. It's like Windsurf is like Trend Kite leftovers. And I use Windsurf on the job.

Okay, so John, what do you use it for? AI coding.

Yeah, it's.

It's an ide, right.

It's like cursor. You think of what you do, where you're getting the thing to write code. Well, it's like AI Studio, but where you're getting it to write code for you. But typically what it's good at is you've got a large code base. We've got, I don't know, half a million lines of code. And it's never going to fit all that in context, but it can search through it, grab the right pieces out, make it into the part that it's going to look at for the moment and then go and plan the next thing to go and do.

John, can I ask, because we've talked about GitLab some and it's a recurring question. We're analyzing and researching how GitLab and GitHub fit into this new stack of software development. As a developer, you're going to interface with a Cursor or a Windsurf or Replete primarily. And then as an organization, you're still going to commit your code to the code base, I guess and use a Git Manager, whether it's GitLab or Git.

Yeah, I use both and just, just like you described, I don't know what the financial gain is for GitHub. You know, the fact that we're doing that because it shows up as largely free. I guess they have their, you know, enterprise aspect to it, but I don't see how it, you know, really scales for its. But I think it does. I think because they have also they, they have Copilot which gets a lot of bad press. You know, Copilot is like Microsoft goes and, and is all over it and then they're like, yeah, people still don't love it. They going back to like the open AI or whatever. And then Microsoft has to wonder if they really can disconnect like they wanted to. That kind of thing.

Well, I think part of what GitLab is discussing is like a. Moving to like a more of a consumption based model versus a seat based model. So even if you aren't necessarily hiring more software developers, which I think in the long run we are going to see more software developers, not fewer. But even in the near term, if the amount of Usage going to GitLab and GitLab goes through the roof because of all this AI generated code that. Yeah. Like you were making, I think they'll have to change the model to, you know.

Yeah. So another thing about the GROK one that I, someone was describing, I don't know where I, where I heard it, but it's really interesting is how it proves this kind of thesis that maybe Musk has in general that he's done elsewhere and so forth. And that is that you're better off with lower complexity in kind of. And more and more kind of power applied to that. So it's like get a whole bunch of chips, keep your architecture simple and go make it happen. As compared to like, if you're like, oh, we do anthropic Claude and we kind of go in and paint in the, you know, we're very about, about how we, we tell it which pieces to do and we're very kind of artisan in how we piece together. You're screwed. Right. It's the, it's what? The way Tess is doing it, the way Xi Xai is doing it, it's like. Is the way you want to do it.

Cut out the parts and John, that's fascinating. I thought about it in that context.

Also. Go ahead. I was just going to note on the theme of acquisitions, slash funding, slash whatever mergers you have SpaceX doing around at 400 billion valuation and simultaneously investing $2 billion in XAI. And, and then now Tesla is potentially gonna, or they're gonna do according to Elon, we're a shareholder vote as Tesla shareholders to also invest in XAI. So there is a lot of money coming into XAI to buy even you know, to get to the 1 million.

That's just it. It's a buy bottomless pit of demand. Like money needed for like they can spend every dollar they can get their hands on buying Nvidia chips if they can get their hands on the chips at all. Which really the bottleneck right now is Taiwan Semiconductor needing to make 10 times more Nvidia's most advanced chips every year. But.

Talk about foundation models and you go like there's maybe it's rumor stuff but that anthropic is like block bleeding out the way that they're having to the amount that they're spending to keep the thing going as compared to what they collect with. You know, they loot, they lose money. Every time they do something with Claude code and get a new prescript subscription there you got like programmers who make it their life's work to squeeze every last, you know, kind of cycle out of the thing. It's like, oh, So I pay 300amonth. I'm gonna get you know, 3,000amonth worth out of it, you know, anyway, it's so it's, it's kind of tricky.

Quick run through with the more questions. We've got a bunch of questions today. Michael says what is cognition trade under? And Craig quickly responded correctly that cognition is privately held and Google obviously trades under the symbol G O O G or G O O G L depending on whether you want class A or class I see I've been holder forever.

What a guy on Twitter this morning was like Tesla is going to be the best pure play on a foundation model. And when they invest in xai. And I responded, I was like, well, what do you call Google? Like is Google. How is that not the best publicly traded pure play on a foundation model? Like we're Tesla bulls, but like you just ignoring Google, I guess he didn't respond ones.

No, he doesn't. He doesn't like you, Bryce. Pfizer announced a 43 dividend to the last. I don't know why. What I. All right, thank you. I'm not in Pfizer anymore. In case that's what that was in reference to. Isn't applied a Bitcoin miner. When I Don't know, dude. All the bitcoin miners are pretending they're high, high performance computing companies, AI infrastructure, like there, if anything bitcoin miners at this point, if they have locked in long term contracts for electricity, they might be an arbitrage on that electricity over the next three to five years or something like that. But I'm short the WGMI I own puts in the hedge fund were short. Not me personally, but in the hedge fund we're short the WGMI and some of the other bitcoin miners we own puts on some of these things we've had puts on applied, we might even have a few right now on it. So I'm not a big believer in this at all. I'm a skeptic that any of these publicly traded pure commodity miners, including bitcoin miners, are going to actually end up being well value.

You should mention the core weave acquisition of Core Scientific for the power and they, the Stock was down 25 or something the day that got announced because it wasn't as high as the market was anticipating.

Right. So that's just it. It was, and it was a pure arbitrage on the contracts for electricity that core we. Core Scientific has. Core Scientific does not have any customers or any real like AI revenue. Like it's, it's.

Anyway, I got a question.

Yes?

You're gonna love it or hate it. So what's the Levi's of crypto? Levi's was the, you know, kind of, it made its money and when it started, because of the, the original mining, the kind of the gold rush. So where they, they're like, all right, we don't, we're not mining, but we have to sell to the miners and they took off that way. So who's, who's going to make on selling to the people who are into crypto and bitcoin and so forth?

Cody, Robin Hood Coinbase.

Oh no, that's not bad.

Yeah.

Okay.

I mean Nvidia is what you would typically call like a picks and shovels, I guess.

Yeah, but that's. He was asking not about AI, but bitcoin and crypto and Nvidia's bitcoin crypto revenue used to matter on a percentage of revenue basis 10 years ago, but it's. I. What is it less than 2% of their revenue now? Like.

Right.

All the revenue is the advanced Nvidia chips, most of their revenue.

So like Robin Hood.

Yeah, Robin Hood Coinbase. Look, I'm a skeptic on the meme coins and all the infrastructure and ways to try to do that. I'm a skeptic on the treasury companies. I do have a really good friend who's actually trying to launch a Treasury company. It's exciting. I'll you know, learn. Love to watch him do it. He's a great person. I think it'd be if it's going to work he might pull it off but I don't think it lasts long. Basically a bitcoin treasury company and crypto treasury company right now what they do is they take investors money and buy the crypto with it. And the market for some reason for at least a limited amount of time here is paying two or some cases 50 times for the shares of value that the Bitcoin act or whatever the crypto assets on the balance sheet actually are. And I just. What about some kind of last that that arbitrage lasted at some point this stuff will probably over correct to the downside. And those companies, treasury crypto companies can very well trade it less than the value of the crypto assets they have. Because why would you want to have that buffer and potential for further fraud or theft or investigations or bad accidents or anything that can happen to publicly traded companies with if you actually want that the crypto you can just own the crypto.

Yeah, please be pure about it. What about anything to do with kind of a stable coin kind of thing?

Yeah, the best play on that right now and it's Circle publicly traded. It's just too expensive. Bryce and I wanted to buy it the date came public but it was up more than 100% after it came public and we didn't want to chase it and it hasn't looked back since. But I think we'll get a chance to buy it below 100 at some point in the near future, maybe before the end of the year.

They make Circle makes the usdc. It's the second largest US dollar based stable coin.

What do you think companies like SMR or even Enphase or which companies. Let's hit the first part in SMR lied to our faces. We do not like them. We are bitter about it frankly. We don't like liars. So no, I'm not a fan of smr. And like I said earlier, why mess with the energy providers? I, I mean nuclear is going to have good growth but I'd rather just buy the AI providers, not the underlying electricity providers which is as I noted on my web app earlier, 150-year-old industry. And it's a pure commodity. Enphase makes chips for like Digitizers for the solar industry. It's a good company. It's just like if you want to invest in solar, you should invest in the very best solar company. Bryce, what's the best solar company? John, not a trick question. What's the best solar company?

Tesla.

Well played sir, well played. Which companies to buy on dips? I've been in Nvidia, I kid you not with ever since 2016, 2017 with the split adjusted cost basis of $0.28 or something $0.70. It's insane so but we buy Nvidia on dips. We bought it at the the day the market bottomed in the Trump tariff tiff days in April. We bought more Nvidia for the hedge fund. I'm not buying anymore for my personal account because it's gotten to be my biggest position already and but yeah, we like Qualcomm, I like tech, I don't know Texas instrument. We bought at 130 again at like the bottom of the Trump tariff stuff. It's now 225. 230. No, I like give it some room. It's gotten expensive. It wasn't even cheap. 130. Qualcomm's probably the cheapest semiconductor stock on the planet. It could be a big beneficiary of the robotics revolution ahead. My boy SK said me or I don't know it could be a girl my or day whatever my person SK says I find copilot annoying. I've never heard anybody say anything else about copilot. Somehow Microsoft is at all time highs. Full disclosure, we're short. Microsoft does not talked about that already. Why don't I like Pfizer? I liked Pfizer. I wanted it to be the Taiwan semiconductor of the pharma of the AI pharmaceutical revolution. But it's not happening. Pfizer's fine, it's cheap, nice dividend. But I'm a revolution investor. The company's not all in on AI at this point or all in on robotics or all in on space. Something revolutionary. You know pharmacology is you got to go all in on the AI part of it if you're going to mess with it. SMR the liars. There you go. They haven't lied about making a profit but they, they told us they had contracts that apparently. Like what was the big contract with the University of Utah or something? Bryce?

Yeah. Well they were supposed to build a big nuclear plant for like a conglomerate of or consortium of utility providers in Utah and those guys canceled their contract because SMR couldn't deliver on what they said they could.

Which A. John, this is for you. SK would like to know, which AI do you think is best for coding? And then Bryce and I.

Could say either Claude or Gemini. I happen to use Claude more. And I think as an environment where you're setting up a project and you give it a lot of stuff to work with in the new term. Now, instead of talking about prompt engineering, you talk about context engineering. It's like, how much context and giving the right context, the right information can you give your AI so that it'll do the right stuff with it. So for me, Claude, and although when I'm using Windsurf, I'm using Gemini through Windsurf, for the record, by the way.

That'S very well phrased because Bryce and I have both been learning phraseology of AI code, requesting the prompts. Like, that stuff is a talent and an art form in its own right. I guess that's why it's called vibe coding, because you got to get in the vibe with the AI and how to talk to it and how to get results out of it. Bryce, your thoughts?

Yeah, totally. Like you. I think it's interesting. You have to learn how to express your intent clearly. Like, if your AI is not doing what you like, it's you. It's usually you not giving it clear intent and instructions. And so sometimes, like, if you just sit there and argue with. You're like, oh, you messed up. Fix it. And it's like, sometimes it can just fix it in one shot, but other times you need to, like, kind of massage it, walk it through how to go about fixing it.

Yeah, and the walk it through things. See this? Some people try to, you know, give it one sentence and have it go. And it's like, ah, I didn't do what I want. It's like, no, you walk it through you incrementally. You're like, let's go get at this part of it, and then that part, and you go get the thing through something I just did that was really cool. I was using Claude for this one. We've got a subsystem that does workflows, and it's probably about a million tokens of size. It's like 30,000 lines of code, probably just for this one subsystem. We had no documentation for users. And it read. I gave it the right pieces of code to look at, and it read through that, which is like, back end code of what goes to put together this workflow, what kinds of nodes, and this and that. Whatever it made front end readable for a regular person, a business person who's making workflows. Here's how you make it, what you can make with it. It made a 35 page manual and it took me a couple of hours back and forth with it to, to get what I wanted. But at the end of it I had, I had like a, a really well written manual on the product when it didn't exist.

Yeah, I, I, to answer the question, look, I mean I've, I've played with Chat GPT coding and I played with AI Studio from Gemini Google Coding. Even recently when I was playing, working on the crochet app for my daughter. I'll show it one more time because nobody, we've had a lot more people join since then. When I was, when I built, when I gave it the prompt, when I gave AI Studio the prompt to build this Crochet crush again, it came up with the name, it came up with the, the logo and it came up with three of the tabs up there. Gemini Chat GPT created something similar when I asked it to and it was okay, it wasn't quite as good but you know, it was a nuanced, maybe even personal preference kind of thing. But this is what the big difference for me was, was being able to come here and hit, deploy the app to the cloud and to hit that bang. And then when it did that, it gave me this URL which is just, you know, the web app URL that Google is Google Cloud is hosting this on. And we pointed it to the Crochet Crush Club URL. But being able to one stop shop from the prompting, the coding, the cloud service, the storage, the memory, the login, it's all my Google stuff. Like I got free Google YouTube TV when I signed up for the AI Gemini Ultra Pro. So Bryce is rightly, from the beginning, since he started playing with the various AI coding web app development tools, pointed out that only Google has the entire stack, including even the processors, the TPU processors that the Transformers that Google were the, of which they were the inventors.

Yeah, I'm not gonna pay 300amonth, so I've got to find a different solution for that. I guess the hosting side of things.

You know, I, I don't know how much of it is the hosting, John. I think part of it is, I don't even know. I just was like, sure, whatever, blah, blah, blah. I'm trying, I want to play with this stuff. It's research, but in the grand scheme of things, I think, you know, you could probably tailor it. You would know enough, you would be in the weeds enough. Be like, no, I don't need that kind of ram. I, I don't need that. I need more of this. And turn that down, turn that off. N N you I bet it wouldn't be 300 bucks a month for you.

Like N8N which is like the work, the, the AI autom automation stuff you drag out. It's like graphically you're dragging out different little nodes on the screen and connecting them and, and you're doing the same. You know, you're making it do stuff like with this, you know what it's like is like zapier or make or you know, the kind of stuff it lets you do. But with that it's again hosted, you pay for that or it's got open source version, you can run it on your machine. It's just hard. It's more complex to, to make it happen.

I bet the same is true for AI Studio. I'm sure it's a little bit more complex, but if you want to put in just a little bit extra time and you have the knowledge base like you do, I'm sure you would be like, oh yeah, it's 38 bucks a month, Cody, why didn't you do it that way? Bucks.

I do think that that advantage though for Google is flowing through to the bottom line the company. There was an article in the Information this morning about Google winning some major, you know, AI deals for the cloud from some AI startups and they actually stole some customers from aws.

Yeah.

And the Information interviewed those, some of those employees that involved with the deal and they just, I guess there was a lot of complaints about Google Cloud early on. It was too hard to integrate, it wasn't intuitive and I think they've totally flipped that dynamic and now Google Cloud is like the go to cloud for any startup these, these days. That, that was basically the gist of the article.

You put the new thing on top, the new stuff they're doing now and it's this extra layer and it makes it like who cares if Google Cloud original, you know, is whatever. Because that's not what we care about. So I was going to ask whether it was sort of stepping in the AWS territory.

It does. And you know, Bryce, even we both, I don't know, it's not on my screen right now and I don't want to go find it. But we, we interface, you know, we, we can go look at our cloud account at Google and it's it, it, I mean it's incredibly intuitive. It's fine. It's, you know, like it's modern and ready to go for anyone to look at.

The Google Cloud console has a Gemini thing built into it. So if you do run into an issue, you can just talk to Gemini right in the Google Cloud console and it'll like walk you through setting up a database or whatever, you know.

Yeah, yeah. Finally I would just remind everybody my last word and then you guys get your last words in. My last word is we are still in the dial up Internet phase of this AI revolution and that includes with this coding revolution and this web app development thing that Bryce and I have been playing with to learn about all of this and how much inference demand there is and how many different applications people can come up with and the how big this all can get. And it's so archaic still. Like when you're doing this, it reminds me of like using whatever that Apple Pro Studio video software stuff was from the 1990s and early 2000s. And like, like it's the interface itself that I was showing you of. All of that stuff is archaic and like not completely, but eventually we're going to get to the point where truly it even says there at the top of the AI Studio that multimodal is coming. You're going to get to the point where you're truly going to just be able to talk back and forth with your browser or app through whatever device form factor you're talking to it through and seeing the output on. But you're going to just Vibe code verbally and then when you finish you're literally going to be like, hey, go public, go find me a great URL and publish it to that URL. And all of the like still painful stupid slow steps that like Bryce had to learn how to do to get point this web app to that crush crochet crush club URL is painful. And the, this discussion about clouds and you know, like how much it costs and all of that. What am I using for what? I don't even know. It's so archaic. You just like we said last week, it will only get better from here. This is the worst it will ever be. Vibe coding is real and it's gonna like drug dealers, tire companies, restaurants, dispensaries, pharmacies, retail shops. Like the Vibe code and app for your store. Vibe code an app for your kid. Vibe code, an app for your interest. Like it's all, it's going to be insane. It's happening. So that's my Last words.

I'll kind of echo that. I was talking to a friend of mine who works at a big consulting firm, was asking him, you know, because I'm, you know, consultants have to be using AI right at this point. And he's like, well, I'm afraid of AI and I think I don't want to lose my job and I don't want to use all this AI stuff. And then my employer figure out they don't really need me anymore. And I was like, man, you know, I really think that's the wrong approach to take. I think the much better approach for your career would be to like be the absolute best AI person in your division of Deloitte and like be the guy that knows how to do all this stuff. And then I saw a video today of a marketing guy who made a, an advertisement on Gemini for Kalshi, the betting online betting website, and made a hell of a good advertisement. And like, it just, I think reinforce that point. Like if you're, if you're in marketing, you are clear and are not using AI, then you are going to be left behind and you should be worried about your job if you're not learning these new tools. But if you are learning these new tools, you're going to be way more productive, your company is going to be more profitable, your career path is going to be brighter than ever before. And so if you are one of those people in tech, in finance, healthcare, it doesn't matter. I mean the best thing I think you could do for your career right now is learn these new AI tools and start implementing them and how to.

Vibe it into it. Prompt. How to vibe prompt. If not vibe code, you got to vibe prompt even just when you're talking to the AI. Yeah, just real quick comment. Everyone was worried about small business because of Walmart and Amazon. This guy Craig says, I think we're at the ception of a small business boom. The democratization of app development, video game development, movie development, commercial development. Like any content you want to create is so much easier. I will add. Dang it. I just. Sorry John, but I got one last, last word and that is it's like, look, it sucks that blogs and written word website newspaper companies are getting screwed because the AI answering interface is revolutionizing how we interact with that content. And people are just not going to click through and read long form blog and articles like they used to. And like it sucks. But so did it suck for the newspapers when people quit reading print newspapers and it sucked for book companies when people quit reading Books and for a while it sucked when for musicians and record companies, when people quit buying CDs and records, but that you have to transcend it. And that's why, going back to that Revolution AI website that Bryce and I built to host all of our web apps, instead of trying to get traffic to come to my website where I write about Revolution investing, we are from now on going to take every article we write and spend some time with an AI studio or similar web app development platform and create a web app for that article, for that concept, for the what the companies in it, how they're interconnected, whatever it might be. And so you, you, you want to get click throughs where you're going to need some great web apps, you're going to need something that provides service to people that's more valuable than just reading a summary of your content on AI. And like, sorry, that's why you need to do what Bryce was just describing. Go immerse yourself in how to figure out how to make great content that is AI generated and immersed in AI so that it's an actual something people can interact with and learn from or use or have fun with or something. And not just whine about the fact that nobody comes to your blog anymore because Google or chat GPT summarized your content. Hey Gemini, write an 800 word article based on Cody's two last words rants.

I think they would probably be like 2, 000 words.

Those are why I wanted to cut it down. John, last word. And then we're shutting her down.

Okay, two things. Just real quick as response to the thing you were just saying is it reminds me of something in a previous company we were doing which was making, you know, interactive infographics kind of built into posts and articles so that we'd have like a data set that was being, you know, used for what we were writing about. And then that data set you could go and you know, drag whatever sliders and things and kind of see something happen to the data so that it reminds me of that from what you're saying. Yeah. And then another. But my last word is around the other original thing you were saying where it was like, you know, this is that we're in the dial up phase. And the thing I'm just thinking about with that is because I agree that's where we are, you know, is as it grows, what are the different kind of dimensions that it might differ along as we go along. For example, you take your grok for, you know, and somebody's off making some cool you know, game with it. And one thing is that game will get cooler and in the next year it'll be a VR game and also whatever. But do we get to where instead of a. This is just one of those dimensions is that, you know, the individual person versus the group of people. Do we get to where that thing being built is itself a kind of community activity and it's getting like improved by multiple people?

Of course. It's right there in front of our face. Right. Like for my daughter's crochet Crush club website, like literally open up the AI studio code, share it on a GitHub, GitLab, and let anyone who wants to improve it help us get it better. And then we, the community can improve the changes and upload it to the. Yeah, absolutely. That's the democratization doesn't have to just be an individual doing it. It could. The communities themselves can vibe together to get the right prompts moving.

And one step up on that is like, it's not like I take it offline and download it and do whatever, but it's like together we're like across with these people. And the thing that today would be one of those, you know, stream Steam live streams of playing the game will be the Steam live stream of building the game and people, you know, together in that respect. So you start to think about that and go like, what are these other. Not just it getting faster, not just it get doing something smarter, but what are the other dimensions you can apply onto it that make it all of a sudden like a different, you know, a whole different place? That's what I was thinking about.

That's awesome. Was a long show. Told you guys we had a full docket. Peace, love and happiness. Bye. Thanks a lot.

Thanks everyone.

Thanks, Donnie. Thanks, Bryce.