The AI Revolution Call

Hello and welcome to the AI revolution. I'm Cody. That's Bryce.

Hello, everybody.

No Johnny Boyd today. My friend, John d'. Oliveira. Bryce, I think you could consider him your friend too at this point.

Yeah, I think so. I consider John a friend.

Our friend, John D', Oliveira, who's been working in AI for 30 years and usually joins us for this call, is on an airplane headed back from Boston to where he lives. I won't disclose that. I don't think it's a secret, but I don't need to be saying it anyway. He's headed home. So Bryce and I have returned for a three week trip around the world. Literally, I added up the mileage. Share your screen, Bryce, or I'll share my screen. The chat GPT thing I sent or was that Gemini? I think I sent the total amount of distance we traveled.

It was Gemini.

Yeah, Here we go. You got it or am I got it.

Go for it.

I don't know how I. I just found it. Now I gotta find stream yard. Share my screen. This is what we did on our trip.

Can you make it larger?

I'm trying. I don't know. There we go. All right. How's that? Yeah, except now you can't see it all on one spot. All right, there we go. So traveling from Carrizozo to Albuquerque, Bryce was similar in his Las Cruces to Albuquerque jaunt. And then New York to jfk, New York. I mean Albuquerque to New York, New York to London, London to the Isle of Guernsey, back to London, London to Dubai, back and forth. Dubai and Abu Dhabi. I think they're actually we could add like another 150 miles because I think we did one more trip, Abu Dhabi to Dubai, which we should add on there because that would put us fully over 25, 000 miles. Yeah, there's a lot of traveling. We are back.

I emailed you, but it looks like that's greater than the. Slightly greater than the circumference of the earth at the.

Interestingly, I had looked that up also. After I saw that number, I was like, I wonder how that compares to if you went around the circumference of the earth and literally it's like 200 more miles than the entire circumference of the Earth that we traveled because of the going back and forth a few times and that we were doing like hops. Not just one big circle, but yeah, 25, 000 miles. Look at my eyes. I got bloodshot this afternoon. Jet lagged. Yeah, I'm hurting a little bit.

Today I, I slept a bunch yesterday this. Yesterday afternoon. So I feel all right, thankfully. But yeah, we're glad to be back.

Yeah, I'm glad to be back. Craig. You guys will be meeting Craig on these calls. I'm sure at some point he is joining our team as a new analyst and support Bryce's right hand. Bryce is my right hand. I'm. I don't know, I run out of analogy. I saw a guy today on Twitter who called himself. He said he's got the highest IQ in history.

Huh.

And then if someone took a screenshot of like he wrote a paper about himself as the highest IQ in history, and that's the reference by which he says he is the highest IQ in history. Really, I sort of like, yeah, 276 IQ, which I don't think that really exists because once you're like at 16165 on the IQ scale, you're like five standards of deviation away or something like 267 would be.

Well, yeah, yeah. I've never taken an IQ test. I have no idea. There's Doug.

Hello, Doug.

One of our other friends.

What's Doug's name on training with Cody?

Jamone.

Jamone. I love that.

Well, you're. You. You basically just dox. Jamone.

Yeah, sorry. Sorry, Dougie. Anyway, don't pick such a cool username if you don't want me to dox you.

So that was a no. I did read an article. This segues us into AI a little bit, but I shared with Cody. But it was written by a professor. He was talking about how AI agents have become useful. And you know, we wrote article. I wrote an article. We wrote an article, I don't know, six months ago, talking about the spectrum of intelligence, you know, going from primitive beings to say AGI or asi, which is artificial superintelligence. And you know, the current version of AI, even the chatbots are somewhere on that scale. Right. And test. Tesla fsd, I think is a good example. Like, is Tesla's full self driving perfect? No, but is it better than my one year old niece at driving a car? Absolutely. Is it better than a 12 year old driving a car? Absolutely. Is it better than the average 16 year old on a learner's permit? Probably is. Yeah. And so like, you know, as. As it moves up the scale, it's not a per.

Is it better than me? Yes, it is a better driver than Cody.

And so anyways, as you move up the scale of intelligence, there's value. Right. A 16 year old driver that can get you from point A to point B, most of the time has more valuable than an 8 year old who can't drive in terms of driving a car. So AI clearly has some value. Anyways, this professor was sort of on that paradigm and he was saying look, AI can't do my job. I don't know what his subject is, but he's a Ph.D. professor. But he's like a big part of our job which is very laborious and time consuming is reviewing research papers made by other professors and you have to replicate the research they've done and if it's replicable then you know it's, it's provable. And he says, you know, it takes hours, takes weeks to try to recreate those experiments and verify, you know, for peer reviewed research. And now the AI can do it 60, 70% of the time it's better than he would do. You know, it takes a few hours sometimes of back and forth with the AI. And I just thought, you know, that was a good use case where it's not, you know, a lot of AI companies like to say they have PhD level intelligence, but it certainly can't do everything a PhD can, but it can do some things a PhD can and it can do those things much, much faster than an average human physical PhD can do.

No, that's right Bryce. And it is a snapshot in time. I, I oft say we are in the dial up Internet phase of the AI revolution. And you know, whether you're, it's funny like right now when people are debating which model is smartest or most effective, most cost effective energy or metrics you're using, it really is like it's going to be silly to look back in three or four years, especially in 10 years and be like do you remember when they were like grading the LLMs on their intelligence and stuff? Like it is still so archaic that we're using it like we're using it, that agents are sort of like at this moment in time, the, the most obvious application that people are coming up with along with maybe coding and software and stuff. But you know, like the agents is sort of an obvious way of using AI. Sort of like email, text on an email was the most obvious use of the Internet back in the late 90s when dial up Internet was what we were using. And so at this snapshot in time, some of this stuff works well, some of it is getting really smart. It's certainly on the spectrum if, if you start at a dumb transistor and you work up to a Texas Instrument 1959 calculator, to a TRS 80 Radio Shack computer, to a Commodore 64, to a Mac, to a MacBook to an iPad, to this incredible smartphone we use today to Oculus and Vision Pro and like it it were, it's silly to look back and think of, hey, do you remember that Excite was slightly higher rated than AltaVista for a few months in 1998? And the other thing to remember, as bullish as we are on AI and the companies that we invest in the AI revolution there, most AI startups today won't make it 90%. I mean, that rule is pretty much how it's a standard in business. 90% of business startups don't make it. And the competition is going to be ferocious. And while the access to capital for someone who has a good resume that's related to AI or, you know, came from an AI LLM that someone's heard of, like you're. The access to capital is great, but so was it for pets.com and others. And so we don't mean to make, you know, be glib about how it's all going to work out because there's going to be a lot of pain, a lot of losses, a lot of ups and downs. And I guess we should hit on Bryce. It was interesting on the trip you had a few recurring themes that we ran into. I think one of them was is AI a bubble? And even the circular aspect of Nvidia invests in OpenAI and OpenAI takes the money and gives it to Oracle so that Oracle can buy Nvidia chips. And so then Nvidia has lots of money to invest in the next LLM that needs data center capacity. And so it becomes a circular thing. And there's been a. Whether I, I will say I think it's worse on Twitter than it is in the real world, people worrying about that aspect of it. But let's address it while we're at it. What are your thoughts on the Circle?

So I listened to Jensen Wong, our podcast after that announcement came out, and they asked him the question and we lost your video, Cody, if you did that intentionally. But anyways, and what, what I liked about it, he said, look, we're investing in these companies and it has nothing to do with the, you know, there's two sides of it. There's, there's the business side of it, the revenue that these companies are generating and the potential for profit. And the other side of it is in the investing side. And they can either raise investment from their profits or from outside capital. And he says, you know, as an investment, Nvidia thinks that OpenAI is a good investment. And he thinks it's going to be the next multi trillion dollar hyperscale cloud is the way he described it. And so he thinks OpenAI, sort of like Xai, is going to continue to vertically integrate and build its own cloud resources on top of using cloud services from Amazon and Core Weave and other, I mean, so from Microsoft and Core Weave and others. So for that reason, he wanted to invest in OpenAI. And then the second thing is they're moving to this model of leasing chips instead of selling chips, which is kind of the idea for this deal. And from the business end of it, Jensen seems to think that, you know, the Chat GBT and the services that open AI sells is going to be a fantastically successful business and they're comfortable, you know, helping their customer finance these chips, which are incredibly expensive, so that they can continue to grow and meet the demand that they're seeing. They just, they can't meet the demand. And so Nvidia wants to do what they can to help them scale even faster to meet demand.

Yeah, I mean, there's still so much more demand for compute than there is supply. And you've seen that evidenced if you've been using Gemini this week or chatgpt and the slowness in the responses and processing that it's doing when you ask questions. I think also it's worth noting that, you know, the vast majority of the spend in this AI revolution is coming from the most profitable, the most technologically savvy, the most advanced companies, the most, the biggest businesses that this planet has ever seen. And those guys, whether you're talking Google or Microsoft or Amazon, they're choosing to spend tens of billions or hundreds of billions of dollars building this AI infrastructure. And so as Doug asks, does that mean Coreweave is running at 100% 24. 7? I don't know that I would go that far because they leave a little bit of compute inside and not every moment of every day. Do you. Are you ever going to be at peak loads in that stuff?

But I would say that one.

Actually, you're right, Doug. Yes. There's so much more demand than there is supply that Core Weave is probably running as much capacity as they can at every moment right now.

I, I would say 100 of Core Weave's capacity is under contract. I don't know that they're. The servers are spinning 100 of the time.

That's right.

That's Right, it's all that capacity, compute capacity has been sold. Yeah.

Anyway, look, I mean this is, this is not a bunch of IPO money. It's not a bunch of these venture capital money. There is obviously some money that is circular here. I mean there is no denying that Nvidia is giving money to a company that is going to use that money to fund the purchase of Nvidia chips downstream. But Nvidia also is not just some random Nortel hiding it's less than arm lengths transactions with its biggest customers. ChatGPT is not directly a big customer at all really. So.

So this is slightly different topic but I've been playing with this new AI tool called Cluli. I don't know if you've heard of it Cody.

Yeah, let's see it.

It's wild I think. Let's see. How do I. Entire screen. This is a tool that listens to your meetings. I had it running in the background and what's weird about it is that unlike almost every other AI meeting tool, the people you're meeting with have no idea it's happening. Right. And so you could use it in an interview, you could use it in situations you're not supposed to use it. Right. Without other people's consent. But what's wild is it's, it's, it suggests questions, it suggests follow up things and then like you can fact check. So I just fact checked ourselves. Like is what percent of Corey's capacity is under contract? Nearly 100 is under contract. Fact check these things. Is Nvidia giving money to company that will use it to buy Nvidia chips? True, here's the source. I think it's crazy and then like if you want to like what should we talk about next? Right. You could say clearly gives instant back checks and context during meetings so you can catch errors. That's not a very good question actually because I just said that but I think it's, I think it's a wild, wild tool. And then, and then you can, you know it obviously it saves all this data and then you can reference the meetings later. It shows you what's coming up on your calendar and then it preps you for the meeting. I don't know how to get rid of this but like it's, you know, there's a meeting we have tomorrow, it says who we're meeting with has the backgrounds on the people that we're meeting with.

So I mean again it's such low hanging fruit. Like that is such an obvious use case for AI LLMs right now. And it alone is mind blowing. That's amazing. It is very cool. But you're right using nea.

Yeah, yeah. So it's just, it's sort of crazy to me that you can have this and then you can also hide it. So like if you wanted to like, see now it's. I made it invisible. So like if somebody was like, oh, let me see your screen or are you using Cluli, you can hide it.

I think they would have no clue that you're using Clulee.

Yeah.

Low hanging fruit there also. Bryce.

Oh yeah, yeah. Crazy though.

It is, it is another cool, another cool tool. And again, it's just so early. Like we'll see. Our minds will continue to be blown as video AI generation goes mainstream and as you know, robotics go mainstream. And Bryce, what was another recurring theme from the event? I mean the trip around the world.

Sure. One of the things we talked to lots of business people, you know, CEOs and other executives, founders, VC people and so they have insights into the enterprise software market. And so we talked a lot about that. And one of the recurring themes for me was that a lot of enterprises are using Microsoft Copilot or they've experimented with it because every enterprise on the planet uses Microsoft products. And so they've pushed Copilot out to everyone. But all, I think without exception they all sort of hate Copilot and would rather use just ChatGPT or Gemini or some Chatbot rather than the built in Microsoft version of the large language model. And so I think it just further demonstrated to me that Microsoft's kind of botching slash has botched this AI rollout thing. And I think there's probably some, a shoe to fall in the AI strategy of Microsoft at some point. As you know, I just. If Microsoft's not rolling out the best products, then they're going to lose in this race at some point.

That's right. And it is interesting, Bryce. It's clear that the, the, the positioning of Microsoft. You and I have talked in these calls a lot about how Google has seven applications that have more than 2 billion users monthly users on each of them. And including just Android devices as a perfect example, all of those things, each of those devices are potential distribution points for Google to put access to Gemini, Google's AI to in, in each of them instantly. And so it's instant distribution, instant usage for Google's gemini. Microsoft has billion plus PCs around the world and incredible, you know, 95 the higher or higher market share in enterprise. I, I'm on a Mac here. And you know, at one point Microsoft was 99 enterprise market share in PCs and they've lost share to Google. I mean well Google Chrome and Mac over Apple Mac over the years but it's still huge market share. And the biggest advantage they've got right now with their copilot is that it is distributed to anyone who's using Microsoft 360 or any of the other software as a service applications that you can use on your Windows machine. And, and you know that your enterprise data is already inside of that whatever Microsoft platform you're using there. And so I think that the single biggest reason without a doubt the recurring theme when we talk to enterprise users of AI, they were saying that they use Copilot because they know they can and they're not comfortable uploading sensitive customer data to Chat GPT directly or to Google Gemini or something. You know, and so that's that that is a in the dial up Internet phase moment, the snapshot in time that we are at in the AI revolution. That is a key advantage for Microsoft. It is a, a serious moat for now. But they will lose that moat quickly as all of the other LLMs, anthropic included Gemini, ChatGPT, everyone else sees that they need a solution for enterprise data to be secure and safe and not sent out and they'll, they're feverishly working on undercutting that moat, eliminating that moat. And so Microsoft needs to figure this out now or it's not the future is not good. Let's see what Craig says here. My experience I'm hearing co pilot is great to summarize. Yeah, I mean exactly. That's so pointless. Like yes, it's great at summarizing long email chains. I saw a meme the other day was like I honey come look. I had this, this LLM write this huge email complex email for me. It took my bullet points and turned into a big email. And then the other guy gets the email and he's like honey, come look. I had this LLM take this big long email and summarize it into some bullet points. That's right. The other applications are not that exciting and even summarizing long emails is hahaha. Because it's not exactly the most amazing thing. And really at this point, you know, you should be using AI like Bryce and I do every day where oh we want to research something. We're not just going to get the AI to do a deep dive research. We're going to build an AI application that we can interact with and use it to learn the app about the product we're using. So yeah, it's the upgraded paperclip guy. What was his name? Clippy.

Clippy.

Clippy.

That's a Mark Benioff quote. Yeah, that's Mark Benioff's favorite. Marvinoff loves to trash Microsoft. Marc Benioff is the CEO of Salesforce.

If I were a mean CEO boss that is, I, I would be like Bryce. I think your new nickname's Clippy. That would be really fun.

Yeah, that would not be a good nickname.

No, but it'd be insulting.

So the. I'm gonna ask you a follow up question from Cluley here.

Yeah.

How do you think Google's massive user base will impact the adoption of Gemini compared to Microsoft Copilot?

Well, we've already seen my Gemini while we were on our trip. Somewhere in the last three weeks it leapfrogged. It leapfrogged Chat GPT in the App Store. It is now the number one app in the App store for your smartphone, whether it's an Android device like Bryce has now or an Apple device like I have. And so I think that instant distribution is already impacted because people get used to using Gemini. And just like right now, at the beginning of this call, I wanted to reference something I had used Gemini for earlier. In this case, in that case I was talking about the total distance we traveled on our around the world trip. But I went back to Gemini to get that. And so there on a consumer level, not enterprise level, but on a consumer level, Gemini is going to be by far the most dominant AI out there. It's just, you know, Grok will be great for people on Twitter and some people will download the app and use it. ChatGPT has clearly got, you know, more than a billion users monthly right now anyway. And so it's not going to away or something but you know, Gemini will probably be the most popularly used AI interface for, you know, probably it'll have more like sort of search market share. 55, 60 of all AI interactions will be on Gemini five years from now.

I mean if you count, if you count Google's AI overviews which are powered by Gemini and AI mode, probably good argument that Google already has more users than OpenAI anyways if you just count all of the AI users of Google. Yeah, that's right.

All right, last questions, last comments. I got jet lag to deal with here. Price sweet.

This week you want to hit on Tesla? We got delivery numbers coming out. It's robo taxi AI pay package. I don't know, I don't think we've talked about any of that.

Yeah, look, Robo taxi clearly has continued to go to go. Well, it's very exciting that you know, there are cars driving around with just cameras, not lidar and they are successfully safely transporting human beings from point A to point B in Austin and soon in other places. Are they operating now in San Francisco too?

Yes. It's similar though. They have a driver.

Yeah.

You shouldn't say a driver. It's a safety monitor. Yeah, yeah.

Human.

Yeah, observer.

Anyway, yeah, so that's going great as I guess we should point it out because in one of our colleagues in a write up a week and a half ago was talking about how hard it will be for Tesla to scale from a few hundred Robo taxis presently to a million someday. And we should just remind people that look, the idea is that pretty much most Teslas on the road right now could be turned into robo taxis with maybe just some hardware, camera upgrades for the hardware, three versions that, you know, cars that are less than two years old really will probably just need a software download and to be unleashed and then you'll be able to add your car to a fleet of robo taxis going around and picking people up if you choose to. So I think that's all on, you know, as you and I wrote about when it seemed like no one else was taking it seriously more than I guess a year ago at this point when we wrote the ultimate Tesla book, it was we wanted to figure out what the numbers could look like and what kind of revenues and profits you could generate off of that. And it really is very exciting looking out. Not just this year, like this year and next year are fine, but really it's three to four years out if we've got millions of Robo taxis driving people at half the price. For Uber, for Ubers and Lyfts and Waymos, you know, Tesla's market share and ride sharing goes or taxiing goes to 80%. Cool.

You know a couple rapid fire questions here.

Sure.

Meta glasses Ray Bans. They just dropped not long after we left.

Everybody, I'd like to introduce Meta salesman. Meta Ray Ban Salesman leading the US and globally after last three weeks. Bryce Smith, Mr. Mike Metaglasses. Tell me you've got them there beside you, Bryce. Oh, that would have been anticlimactic if you did not have them there. How many pairs do you think you sold in London, Guernsey, Dubai, Abu Dhabi and Hong Kong over the last three weeks as you walked around showing everybody and raving about them to everyone.

Everybody loved them. I. But I sold 20 pairs. I know for a fact I've sold like four pairs. Seen the people that have bought them.

So yeah, yeah, you took some incredible footage. We got stuck in the riots in London for the. It was anti immigration slash antifa and riot cops lining the streets and smoke bombs going off. And Chance and Bryce and I were just trying to get back from a lunch to the, to the flat of the guy that we're partnered with on our fund over there and tried to get to the train station from there. And it was all very disconcerting and quite even stressful. But Bryce was, It's really cool. You know, he's taking pictures and, or recording his view of it all and it was interesting. We'll share some of that footage, I don't know on Twitter. Follow Bryce on Twitter. Throw your handle in there, Bryce, if you want.

I don't even know. I'll have to find it. But the new glasses are very exciting. I think they're going to be a big hit. They, you know, you can, you can see your text messages, you can see your map for navigation. You can see the AI responses. Instead of just hearing the AI responses, you can draft text messages. Like, I don't know. I think it's a lot of people like to, we're making fun of the launch event, but I think people are going to be surprised at how well they sell once they get out there into the hands of consumers and people start using them and posting about them and sharing videos they make with them. Oh, you can do one of the features they have, which these don't have, which I think is cool, is using WhatsApp. You can do a face, a voice video call with somebody. And so like, let's say like during that protest, I could have video called one of my family members or something back home or a news station or news station. Right. And they could see what I see and then I could see them in the, in the display.

Yeah, that's cool.

Yeah, that's gonna be it. I think that's gonna be Bryce.

Let's. The reason people were poo pooing the rollout was because like, like he couldn't connect with this tech guy. Like it was a disaster of a rollout and like, let me show you how cool this stuff is.

Yeah.

On the other hand, I, you know, like now we're, we're talking about that rollout and so it actually gets them a lot of press. And then, you know, they, I think everybody knows that like what they buy in in a few weeks and get delivered for Christmas is not gonna. Like Oculus isn't sitting still. Meta's improving it every day, including now. Here's Bryce.

Follow at.

@Search for Bryce Smith. I guess. I don't know. Nobody's gonna. Nobody can. Who types Smith underscore Bryce 12. Like, what is the worst username of all time?

That's why I can never remember my Twitter handle.

Yeah, exactly. You can't even remember your own Twitter handle.

Well, I wasn't early like a one of the first hundred thousand Twitter users like you to actually get my real name as my Twitter handle.

Yeah, yeah. And Cody Willard's not quite as popular his name as Bryce Smith around the US either. But that is still sort of inexcusable. It could have been like Bryce Smith funds or Bryce Smith Tech or something.

I was like in high school, I think, or college when I made that. 12 is the year I graduated high school. That's probably why I did that.

Anyway, I'm sorry I even put that up there. Everybody ignore that. You don't need to follow Bryce on Twitter. He'll post it to my. I'll have him post it to my Twitter handle.

So.

Which is at Cody Willard. Watch this, Bryce. Look, I don't have to look it up. I go, Cody Willard.

Bam. Done.

Shared. Look at that. I'm at Cody Willard.

Yeah, yeah, yeah. Only one of us here has been blocked by Elon, though. So a lot of. A lot of good your Twitter handle did for you.

Going below the belt, man. That's a low blow right there. All right. The other one I guess we can need to hit on quick is sats, baby. That is Echo Star. Let me just do the math for you guys here. We've been. We've been investors in SpaceX. This is about SpaceX, okay? We've been investors in SpaceX and my fund for six or seven years in the private markets long before it was cool to do that. We were the first Bryce do. Do the trump. Like, we're probably the best VCs. Nobody's as good a VCS as we are.

You got it?

That's terrible, man.

Terrible.

Anyway, Echoes. We. We've been forever trying to figure out how to buy SpaceX in the public markets without wildly overpaying it. And even if you want to buy SpaceX in the private markets right now, you're probably going to wildly overpay. There's going to be a commission of maybe 5% on it and then a carry fee of somewhere between 10 and 30%. And so however much it goes up, they take 10, 20 or 30% of whatever those profits are and it's a double layer SPV or a triple layer spv, which means you own a part of a special purpose vehicle that might own a part of a special purpose vehicle that might own a part of SpaceX equity. You got to hope it all works out. And not only that, you can go out and buy like there's a publicly traded closed end fund called dxyz and this is not a recommendation because, you know, it trades at 5x or 7x its actual book value for what it owns, SpaceX and open AI and stuff, but it trades at 5 or 6 or 7 times what the actual value of those things would be. And so you're literally paying it like SpaceX is trading at $212, not trading at that. But it's the most recent time when SpaceX bought its own shares back from employees and stuff. It a tender offer, it was at $212 a share. Buying DXYZ is like buying it at $1200 a share or more. And so we don't want to do that. So anyway, space, Bryce and I, about three or four weeks ago I turned to Bryce and said, you know what? Communicate with the wireless spectrum upon which these devices communicate and the Internet, satellite dishes talk, cell phone towers go back and forth. Spectrum is the water of the communications age. And Bryce, we should go do a bunch of work on spectrum. So what'd we do, Bryce? We built a Spectrum app. We built an AI app all about wireless spectrum. Who owns what we wanted to understand, most importantly space to space spectrum and space to earth spectrum. We wanted to remind ourselves of all terrestrial spectrum LTE 5G, who's going where microwaves open free plat megahertz ranges and how much is each range worth? What's called a pop. How much is each pop worth? So we built all this, did a bunch of work and EchoStar is the company that owned most of the spectrum that SpaceX needed. And I use it in past tense because guess what? The FCC basically forced Echo Star to sell that spectrum to SpaceX. It hasn't closed yet. It'll probably close in a year and a half. But everybody in the government and everybody in the private market now wants that deal to happen. It's probably really, really going to happen when and get cleared. And so SpaceX bought it for $20 billion, $2 billion of which is interest payments for Echo Star. $9 billion or 8 and a half billion dollars of which is cash. And most importantly to this story buried the lead. Here it comes. SpaceX gave them eight and a half billion dollars at $212 a share of SpaceX equity. And so you do the math. Right now the company just wiped out all of their debt because they also sold $24 billion of spectrum to wireless AT&TO AT&T wireless a week or two before this deal was announced. That wiped out most of their debt after taxes and everything. Sachs will have about a billion dollars plus in net cash on the balance sheet. They'll be generating about a billion dollars of EBITDA with their existing businesses and they'll own about $9 billion, $10 billion worth of SpaceX. The next round of SpaceX is probably going to be done at 250 a share. So 20% pop from the most recent round. So really it's like more than $10 billion of SpaceX equity. And the kicker, they still own. EchoStar still owns about five to $10 billion worth of other spectrum which they'll probably sell now. And so they're the net ass. You're you're paying maybe even a discount. You're certainly not paying a premium to buy SpaceX in a liquid publicly traded vehicle called EchoStar symbol SATS. I am not recommending. I'm simply telling you that we've been slowly but surely accumulating sets and I only told you that because Doug Mitchell asked how'd I do? Bryce? That's pretty good, right?

Good. You got it. I think that's good.

Smith_Price12@twitter.com.

You can also find me on Discord.

Cody Willard. Cool. Yeah, I bet I was Bryce. I bet I was one of the first couple hundred thousand people on Twitter access.

I'm one of the first hundred thousand people on Cluly.

I bet that's true.

I saw the CEO in an interview and he has five employees. It's pretty small.

Which let's my last word then will be there are going to be dozens if not scores of companies worth billions of dollars even in Europe because there's some brilliant entrepreneur, 18 year old kid in his parents flat. What do you call an apartment in French? Bryce and Perry. Anyway, that's building the next killer app. You know there's going to be some killer apps that are going to be coded with AI and don't have very many employees and going to be worth billions of doll. The democratization of app building is going to MIT a few billionaires, maybe a few hundred billionaires around the world in the next three to five years. Feel comfortable saying that? Maybe some kids in Africa, maybe some kids in South America, a few Europeans, some people in Asia, maybe even a Canadian. There might even be a Canadian that comes up with the killer app not named BlackBerry. Shopify. I guess Shopify. Shopify's worked out for the Canadians. Nortel, not so much. BlackBerry, not so much.

All right, I'm gonna give you the four bullet recap from Cluley.

Okay?

Okay. AI meeting tools like Cluli are providing real time fact checking, context, meeting prep, but raise privacy concerns. Number two, Tesla Full Self driving is considered safer than a 12 year old driver. It's being tested in cities like Austin, San Francisco. Corey's compute capacity is nearly 100 under contract due to high AI demand. And Microsoft Copilot is widely distributed in enterprises but often disliked compared to chat, GPT and there's more, but I need to upgrade to the pro plan to get the rest.

That made me snort. That was a great way to end the call. You guys are our normal regular viewers can click off now. But if you'd like to upgrade $99aminute version, you can stick around and Price and I will continue to talk.

Cool. All right.

All right, thanks everybody. Peace and love. Thanks, Bryce. We'll see you guys next week. We'll have John do back in the saddle and one of these days Bryce is going to build an AI app in real time for you guys. Stay tuned for that. We're scheduling it presently.

Cool. Thank you.

Thanks, Craig. Thanks Jamon and anybody else out there. Thanks.

Bye.