Limitless: An AI Podcast - The $META Investment Thesis: The Markets are Wrong

Episode Date: October 16, 2025

In this episode, we explor Meta's stock performance amid competition from Google and OpenAI. We discuss challenges with products like the Vibes app and evaluate the behemoth's long-term potential.------🌌 SORA CODES: DM US A SCREENSHOT ⬇️https://x.com/LimitlessFT------TIMESTAMPS0:00 Meta's AI Spending Frenzy1:01 The Bull and Bear Case for Meta Stock4:00 Hiring Insights and Market Reactions4:38 Evaluating Meta's Spending Strategy9:28 The Dilemma of Meta's Product Launches10:50 Unpacking the Ray-Ban Meta Displays13:19 Meta's Hardware Challenges18:05 Meta's Data Center Ambitions20:01 The State of Meta's AI Apps23:20 Meta Vibes: A Missed Opportunity25:16 The Future of Meta’s AI Models30:40 The Bull Case for Meta's Future35:41 Final Thoughts on Meta Stock38:36 Special Offer for Our Listeners------RESOURCESJosh: https://x.com/Josh_KaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠

Do you guys remember when Mark Zuckerberg spent $300 million to hire a single AI researcher? Well, this week he spent 10 and a half times that amount. $3.5 billion to hire one person from his competitor. At this point, I think it's evident that meta is relentless with the amount of money that they're spending to win the AI race. They've given away tens of billions of dollars in AI models for free to try and undermine their competitors, which failed, and they've already committed to spending $600 billion over the next five years here in the U.S.

to build our data centers to train frontier AI models. But one question remains unanswered, Josh. Is metastock a buy or is it a sell? And I ask that question because it's pretty clear that competitors like Google and Open AI are eating their lunch. They're beating them in apps. They're beating them in models. They're beating them in data centers. So in this episode, Josh and I are going to cover the bull case and the bear case to invest

in meta stock. Make sure you stay until the end because we have a special surprise for you from our friends at OpenAI. Josh, I want to show you a chart. I want to show you the meta stock chart. Over the last month, it is down 5.6%, which may not seem like much, but that is billions of dollars, whilst their competitors like Google, like Nvidia, have been up only. Do you have any thoughts about why this is? Because I'm personally confused, right? Zuck has spent, I think, $30 billion now to hire, like, a killer team to build out AI for the company. They have released a bunch of models and open source them, and they're investing in data centers. Why is this happening?

Yeah, I think the market's trying to understand the new meta, because this is a very different company than what it previously was. Previously, it was a social media company. They sold ads. They made a ton of money. They had billions of users. Now they're pivoting to, I mean, first it was a pivot to the Metaverse where they started building the virtual reality goggles. Now it's more of a pivot into AI and fully putting the balance sheet on investing in these large infrastructure projects, investing $3.5 billion on a single developer. Their CAPEX is going through the roof when we're not really getting the results to see it. So what I think we're seeing here in the MetaPrice and if you move to the six month chart is actually like fairly good, I believe.

It's up to. They've been doing pretty well. It's up well. People are trying to get a gauge for what they're doing right, what they're doing wrong, and how they could price this into the stuff. And I think this episode, we could kind of walk through the different elements of meta's business and kind of talk about what we like, what we don't like, because the market very much feels wrong when it comes to evaluating some things. But it is kind of right when it comes to evaluating others. So I think we're going to just kind of diagnose why the market is wrong for specific things that we have like a pretty unique interest in covering. Let's start with hiring, Josh, because like I, sorry, I can't get this headline out of my head. Let's give some context for the listeners here. this week, Zuck is rumored to have spent $3.5 billion to hire this guy called Andrew Tulloch from a company called Thinking Machine Labs, which coincidentally was founded by an OpenAI co-founder, which left Amiro Murati, who left Open AI and started at home thinking lab. What's so interesting about this is Zuck already made this guy an offer about two months ago for $1.5 billion, which he turned down.

But apparently everyone has their own price. Why this is shocking to me, Josh, is meta literally announced that they were not going to spend any more money hiring people for their AI team. And they've gone ahead and kind of broken that rule and spent the most amount that they ever had. But number two, and I have to be honest here, I'm going to eat some humble pie.

I remember on an episode about two months ago when Zuck went on this spending spree, I said, hey, we've got to give Zuck some time because it's going to take some time to build these AI models, to build out some apps. Now we're three months later, and he still hasn't produced anything to kind of show for it. So I'm kind of feeling like Zuck is misspending his money,

and he's not able to execute on building out the grandeur AI vision that he promised us. Where is the artificial general intelligence? Since they've started this huge hiring spree, the only thing we've really gotten is vibes, which was that, like, social media, TikTok, scrolling-based thing, Did it actually release you, Jazz? Is it even out? Because I haven't played with it if it is. No. Well, I mean, that's just something about both of us. When was the loss of you were on Facebook? I can't. I can't tell you. I'm not sure I even have Facebook still downloaded. It might be downloaded on my phone. I'm not logged in.

But I do want to talk about the spending because there is the bull case and the bear case both to be made on spending. What we're seeing Zuck do is he's really going all in. And that level of conviction matters because the company needs to also be all in. And they're spending a remarkable amount of money in doing so. The thing, though, is, is the thing, $3.5 billion is a drop in the bucket if they do this right. If they actually do create really high-quality AI experiences for their many billions of users, that generates a tremendous amount of revenue. And meta-stock is very, very clearly undervalue now if they're actually able to deliver on that. Because when you think about the costs associated with building of these giant data centers, they are tens to hundreds of billions of dollars that they're going to be

pouring into this. And if one single AI researcher can create one single algorithmic unlock or make a small incremental improvement that lowers the amount of compute required by even 10%. That covers his entire compensation and then some, with many multiples to spare. So the leverage of the decision making of these employees is astronomically high. And if you can get one person that can save you even one percent on $100 billion, that's a billion dollars. It's a lot of money. How much time are you going to give Zuck to do this, Josh? Because like, is it months?

Is it a year? Like, how long am I holding my metastock for? You have to give them time to release something. I know Vibs was something, but it was very much not. It was just kind of like a nothing burger. I'm not sure why they did that. It was kind of a waste. It left very bad, very bad, like, optics and a taste of my mouth because it's just like,

what are you doing with all this talent? You're making more AI, social media applications. But I want to give them an opportunity to actually release something substantial. Like, release a big model, release some sort of large product. I would say wait until version one of whatever they're working on to come out before evaluating is my take. My gut wants to basically say that I want to sell my meta stock. I hold matter. But my kind of brain is telling me, like it took years to build the first version of chat GPT.

And it's like no secret that Zucks joined this race kind of late. And he's trying to make up for that by spending the money to hire the talent. but the talent needs time to cook, as you say, right? So put quite clearly, I'm going to reserve my judgment until the next AI model that meta releases is out, whatever that is, if it's Lama 5 or Bahamas or whatever the hell they call it. I'm going to wait for that because the model, the chat GPT competitor, will basically determine whether they're good enough or not.

Yeah, that's interesting. Okay, so disclosures, EJS does have metastock. I have none, but I do, I guess I have a bulk case to make for them, which is that meta. Meta has a couple billion users, which is a ton. They also have a ton of data on those users in terms of preferences, what they like, what they don't like. They also have a algorithmic engine to serve those users exactly the type of content that they want, and that pairs very well with their advertising engine, which is world class. I mean, Instagram at least, because Instagram

uses the same engine, they're able to serve me ads that are better than basically any other ad that I get anywhere on the internet. So they have this really vertically integrated stack of users, technology to serve users and technology to serve their preferences. Now, if they can apply an AI layer on top of that stack where they can now coerce these, I don't even know how many billions of users they have, but many billions of users to further embed AI into their lives, I think that's a really unique advantage. Nobody really has the software stack that meta does and also the hardware capability. Now, meta is very bad. We'll talk about them. We'll have left a lot to be desired, but they're doing it. and they have a product and they're learning manufacturing and they are iterating quickly.

And if they can combine this really amazing software stack with billions of users and combine a compelling hardware product, which would hopefully be the glasses, and create this really nice ecosystem around it where you can augment your entire life with AI. That's a pretty unique position they're in because there's not many other companies that have that type of leverage. I mean, Open AI has the most users. They don't have the hardware. Google has a lot of hardware and kind of users, but their products suck.

Zuck actually has both. They just both need to be good and they both need to work together. So if they could do that, green charts ahead. Another thing that Zuck has repeatedly shown he's good at is taking all of this data and these users that you mentioned, Josh, and actually creating really fun like social features or products or network effects from that, right? He created Facebook, he acquired Instagram. you have the WhatsApp thing. Everything is like kind of cultivating this like ecosystem

of people just like communicating with each other as much as they can and there's probably no one better at it, right? From a hardware perspective, again, we'll talk about it in a second. I'm hoping that he can pull this off for AI. He's had a few shots on goal and in my opinion they've missed. He's had the meta chat bot AI assistant. And do you remember when news broke that all user,

prompts and conversations were actually made public. Really weird decision. He released the MetaVibs app, which is kind of like a TikTok competitor with all AI generated video. We're going to talk about that in a second. And that has kind of also flopped. So I commend him for trying. And he is like the master. He has so much history behind this. So I'm still bullish that he will be able to pull something off. And to be honest, and I know this will seem like a cop out, Josh. It is still too early. Like no one's created a sick AI social media layer. Some have taken some kind of bold attempts, but we haven't figured it out. We haven't had the final hardware form factor. We haven't had the final kind of like perspective on what this AI kind of like system will look like. I'll be just talking to chat

GPT for the rest of our lives or is there kind of something else? I think once that materializes, Zuck is going to come in like a cannonball and absolutely blow everyone apart. So I'm still reservedly bullish. I'm just short term. I'm kind of like, nah, whatever. Okay, well, you mentioned hardware doesn't have a form factor yet, the final form factor at least. We do have a form factor, and that is with these meta glasses. So can me please, I'm dying to talk about that. Okay. Please, let's talk about glasses.

I wonder why Josh is excited to talk about this. Buckle up. Right. So, some context. We released an episode about a week ago where meta announced this new AI glasses. They're called the Rayban Meta displays. And the reason why this kind of blew everyone's minds when they announced this news was, People kind of have never seen something like this before.

It's a pair of glasses that you kind of wear like a normal pair of glasses, but you see a screen projected on the right side of your glasses. So you can see things like apps that you see on your phone. If you're like walking somewhere, you can have a Google Maps that tells you where to go. You can answer text. You can take pictures as a camera embedded in the frame of the glasses itself. You've got speakers in the handles that like kind of like project audio into your ears like an airport. It was this new device that we haven't really seen since Google.

Glass, which kind of like flopped way back when. But now supposedly is the time. And this is the first bold step that Meta is taken. And we've got a tweet here. And for the listeners, I'm going to read it out to you from Shil Monot, which is, you know, he's been a meta bull for a while. And he goes, the Rayban meta displays are fulfilling the original promise of Google Glass. Society wasn't ready for the glass, to be honest. Now phones are out recording stuff all the time. Meta has done a better job of it. They look normal. $800, it's mine. I'm going to buy it. And then literally two weeks later, Josh, I have to share this tweet from the same guy, from Shil Monot. And it goes as follows. First impressions matter and meta really blew it with the

Rayband displays. I was ready to buy them on launch, but instead of letting me order them, they forced me to book an in-person demo at Best Buy. It was a complete disaster. And he goes on to explain his experience where the rep didn't know anything about the product. They couldn't get the glasses to work and he told him to go away and come back later where, you know, hopefully it works. So long story short, I think it was all talk and no walk. And I have to spend 20 seconds, Josh, before I pass the mic to you. I need to eat some glass and humble pie again because I was super bullish about this, dude. I was bullish because I'm tired of buying a new iPhone with like a better camera every year. I want something new. I want something that would blow my mind.

when the original iPhone came out. And so when META, when Zuck specifically took this step forward, I was applauding him because I was like, no one's been bold enough to do something like this, and I'm excited to see this worker, and it's, and it suck, it's meta. Like, they can't, like, release a shitty product, right? Apparently the product is kind of shitty,

and it's made me bearish about buying one, dude. I told you, like, you know, you asked me like, hey, have you bought these yet? I'm like, they don't have an appointment for me to walk it. I have to walk into a physical store to buy this thing instead of being able to order it online, which sucks to start off with. See, I'm confused, Josh.

Help me make sense of this. Well, I commend you for your optimism. And in a way, I envy it because I think when it came to evaluating these, the pragmatic part of my brain won. And I remember saying after watching the presentation, like, this is not a serious presentation. This is not a serious product. It's disappointing that that's the reality.

Meta really just has no business releasing these products to the public yet. And I respect the fact that they're doing it in public. They're shipping in public. they're testing in public. But these are very much developer versions that, I mean, if you could even call it that because they hardly work. I mean, we're seeing on screen now another disappointed review from Robert Scoble on X. He's been wanting them for a really long time.

I think everyone is. Everyone kind of has this aspirational idea of what the next type of hardware interface looks like and is very much glasses. But meta is just releasing just like horrendous products. One of the things that I think they made a really big mistake with is forcing people to go in store to purchase. but forcing them to go into a store that they don't control. If you remember the Vision Pro with Apple,

you didn't need to go in store, but you could go in store and get a demo, but you go into an Apple store, and it is a beautiful experience. They pulled out these brand new couches for it. They give you this fully walk-through guide. It showcases all of the amazing technology, and it actually worked.

To go to a Best Buy and have, like, the random guy from the neighborhood who hasn't been trained on this, who just read a little pamphlet from meta and said, okay, go show people how this works. and then to have to have that person demo a half-baked product with like pretty mediocre software and horrendous hardware, quite frankly. It just, it doesn't work. And that paired with the fact that there's supply constraints where even if you wanted to go

to most Best Buy's, they don't even have them in stock. The software doesn't work. The hardware doesn't work. It's a total disappointment. I think Zuck should do what they've been doing with the software and AI models and just don't ship mediocre stuff. Like with open source models, they were going to ship behemoth, which was their very large open source AI model.

And it never was good enough, so they never shipped it. And I think that's fine. I'd rather you just keep it to yourself. Don't ship garbage. What they're doing is they're shipping garbage. It's not a good product. It doesn't work well. And it's just, it's disappointing.

To kind of like push on the bear case, Josh, it seems kind of desperate to me. You know, they're shipping half-assed products which look like something an infant used back in the 1990s. I mean, like, look at this. Like, this is someone's perspective of what they can see using Google Maps on the glasses. And for those of you who are just listening to this, it looks like an old school projector that you used to have at high school back in the day where people, where the teachers would move slides.

I don't know whether you guys had this here in America, but growing up in England, this was like literally a lamp and a piece of glass. And you kind of like see what you see. This is what it looks like. It looks like something like some hacker made in the early 2000s. It doesn't seem awe-inspiring. So, number one, like, the tech just seems rushed and desperate. Another point is the point that you brought up on the episode of a week ago, Josh, which is,

I don't know if meta is playing in the right field. Like, do they have the hardware chops to even scale this hardware out? Like, who are their partners? Who are their suppliers? Like, the fact that you have to walk into a Best Buy a store that they don't even own is kind of tragic. And it doesn't speak very well of their brand. Now, when I look at the bull case of this, I'm impressed. that Mark is taking the shot.

No one else has taken the shot. No one else has tried to, no one else at the major level of like a Mac 7 company has produced a new device yet. And it's brave of Mark to be the first one to step out. And knowing the amount of cash reserves that Meta has,

they're going to invest in this. They're going to iterate really hard. They've got all these new AI researchers. They're going to work hard at, you know, making this better really quickly. And I wouldn't be surprised if version 2 or version 3 of this actually slacks.

Number two is they're making net new hardware as well. Don't forget the motion wristband, which they created, which is so, so cool. Like any kind of slight gesture of your fingertips and stuff will kind of like translate into the metaglosses that you're wearing. It helps you select an app or move to the next track when you're listening to music. So there is something here. I just think it's super early and it's super rushed. Again, this will be a matter of waiting, a couple more months or maybe six months to see

whether they can deliver on a second good version. Yeah, so here's probably where the market is wrong in the downside, where I'm looking at the stock right now, it's up 37.4% in the last six months. It's had an amazing run. The reality is that they've never shipped a great hardware product. They haven't really shipped any great software products that use AI at all. The meta Quest is just not that good.

These new meta ray bands are even worse. They have no track record. of shipping anything good. Everything they've shipped has been poor. There's no, in the case that they do ship good hardware, there's no ecosystem for it. So in the case that these meta-ray bands were fantastic, I still don't use any meta-applications. Like you said, I just, I don't use Facebook. I don't use WhatsApp. I don't use, I use Instagram sometimes, but I love my phone as my camera to interface with that because it's a far better camera than the meta-ray bands. So there's no world in which I would even want to invest in this ecosystem in terms of buying hardware products, even if they

were excellent. So meta has two really difficult problems to solve. One is building a useful software ecosystem that I actually want to participate in. And the other is building actually good hardware products. They don't have either. So the fact that meta is up 40% without releasing any good software, AI or not with their $10 trillion of employees or $10 billion of employees I just hired. And they shipped like a half-baked hardware product. It's like, why is this talk up 40%? I think they are betting on the fact that meta will be able to convert these users to cash at some point. But like, My God, they have a very long way to go and zero proven track record. The bullcase, like you mentioned, just briefly, is they have cool technology.

Like, the potential energy is there. They have unbelievable talent. They have this really cool technology in the wristband that can predict your neural signals and it can know what you want to tap before you tap it. They have the foundation and the roots to build something cool. They just haven't done it. So, hey, hopefully they do. And I think that 40% gain is probably indicative of people being kind of optimistic and hoping that they will.

Let's talk about Meta's bread and butter what they're actually good at. They're good at making social media apps, right? That's how they've made all their money. That's why they have all their billions and billions of users, right? Unfortunately, when it comes to AI meta apps, it's actually my most bearish argument for this entire episode for my Metro investment because I think two weeks ago,

they announced their first meta AI app. It's called Vibes, which is basically, kind of like a TikTok style app, except every single video that you scroll past is AI generated. And they had this kind of like cool promo video where they explained that it's super easy to make AI videos. You just type in a prompt. And if all of this is sounding familiar,

it's because it sounds like SORA, Open AI's app, which does pretty much the same thing, except Josh, there's one major difference between these two apps. One of them hit number one in the app store and is still number one in the app store. whilst the other has kind of like fallen down the app store rankings as well. Like I said earlier, like I don't remember the last time I opened the Facebook app. Why didn't they just create a new app?

And so the point I'm trying to make here is how has a company with, you know, trillions of dollars of value and billions of dollars to spend and some of the best social media app engineers ever to kind of coalesce in one single company flopped on their opening AI app whilst, sorry, pun intended. I guess. Open AI's app, SORA 2 is absolutely killing it. Like they've let a newbie come in and eat their lunch. See, this is the tough thing with the ecosystem is I'm not even sure how to access the Vib's app. Is it a standalone app? Is it built into my Facebook app? I don't know. I haven't used it, nor do I have any incentive to because, I mean, even though Sora came out later,

well, Sora, I use chat Cheptie. In fact, I love using chat Cheptie. And chat Chbett is very present on my timeline, too. There's a lot of people talking about it. There's a lot of people talking about it. There's a lot of people using it. I want to be engaged in that community. There's nobody on my timeline talking about vibes. There's nobody talking about how much they love their AI generated video on the vibes app. I haven't even seen any screenshots or videos of it, which is why I'm not even sure if it's live yet. There's just this huge disconnect between the velocity of, I guess, social acceptance of OpenAI's products versus METIS products.

And while they were both mostly the same, it would appear at least from the announcements, I very obviously just, prefer opening AI because that's where I am. They have my data. That's where my friends are. That's where I can scan in my face and make a cameo. And I think that's the most viral element of the SORA app versus Facebook's vibes, which just it's very sterile. It's very generic.

So I think like meta very clearly lost. This was the one on one they lost. Mostly because of the social virality brought to the application through SORA's cameo feature where you can actually insert yourself into the AI content. And it's, again, it's just a disappointment. You're stealing the announcement from Alexander Wang. They spent billions of dollars on him. He's in charge of this whole AI infrastructure plan.

And they're just releasing something that, again, it's kind of like half-baked. I'm not even sure how to use it. No one's talking about using it. There's nothing compelling about that product. So that to me is a swing and a total miss. I actually think it's the biggest swing. I think this is the bullet in the chamber, Josh, because of a few things. Okay, so number one, I'm going to pull.

up this tweet from Alexander Wang, who talks about this MetaVibs app, and he says, for this early version, we're partnering with Mid Journey and Black Forest Labs. Okay, what this instantly tells me is this isn't fully thought out. They've rushed this product once again just to get something out there, probably because they heard that open air was going to release their own social media app. Number two, this is kind of an emperor has no clothes situation, in my opinion, because what it tells me is you need a really good AI model at your core product offering in order to build any cool app on top of it. So either you use someone else's AI model or you build it yourself.

And Meta has tried building their own model in their Lama models. They open sourced it. And the fact is, does anyone talk about it anymore? I haven't seen it on the timeline. I haven't seen anyone of my friends, software engineers using it at all. So that kind of flopped. And everyone's kind of being quiet about it. And it's translating into their apps now. They don't have a good enough model to create good enough AI video to compete with the open AIs and the Googles of this world. They have V-O-3.

They're putting it in YouTube shorts. So META has the distribution. They have the users. But in my opinion, they don't have the foundational key component. And this in particular, the META vibes app, kind of reveal that the most to me. It's where I'm most bearish on it on META as an investor. at this moment. Now, I need to force myself as an investor to look at the bull case, right? Because either meta stock's going to zero because I don't care what anyone says,

AI is the future. If your company doesn't execute on it effectively, you are going to zero. So is meta going to zero or are they going to swing things around? One thing that I've seen suck perform really well at, Josh, is not only creating things from zero to one, but taking things from one to 100, right? We saw it when Snapchat came out, right? and Zuck was like, let's try and buy them. And they were like, no. And they were like, okay, cool, we're going to just do our own thing. We're going to create Snapchat competitor.

And here we go. And you had Instagram stories. He did the same thing with TikTok. They created Reels. And now all the famous SORA videos that are released in the last week are getting shared on IG Reels and they have millions and millions of views, right? So I'm still confident that Zuck will kind of like take this hit on the chest and be like, okay, cool.

well, whatever the good version of this ends up becoming, we're going to build a better wrap. And maybe, again, that is meta vibes 2.0. Yeah, it's getting harder to copy because like you mentioned, having that brain at the core of your infrastructure matters. And I think a big difference that we see between Open AI and meta is that open AI builds that brain first. And then they figure out a vessel to wrap it up with a product and ship that as a way

of distributing the brain. That's what we saw with Sora. They developed this amazing video generation software. They packed it up into a product. They shipped it. They created this amazing LLM. They packed it up into chat chepti, and they shipped it. So they developed the technology prior to developing the product.

What MET is doing is they're starting by developing the product, but they don't quite have the technology to place inside of the product, nor is the product particularly good. So Open AI, in this case, they beat META on product because they had the cameo feature, which was amazing for virality. meta didn't have anything uniquely compelling about it. And it's clear from this that they had to outsource the brain, the thinking, to Mid Journey in Black Forest Labs. So they only had to make the product. They failed at making the product. And they tried to make the brain, but failed and had to

outsource it. So from my perspective, they're getting just like crushed at all fronts. And I think it's increasingly difficult to see a company like Snapchat and crush it because you can't just copy a code base and run on a CPU, you need a gigantic GPU cluster in order to to copy these with the actual level of vertical integration that we see from a company like OpenAI. So in a world in which Zuck's competitive advantage was just crushing his competitors by copying them and making it slightly better, it's getting increasingly challenging to do. So that would be my concern as I'm just kind of observing the meta stock is like, hmm, okay, you haven't shipped a great product, you haven't shipped any good AI, you haven't shipped any good hardware. Like when is this going to happen?

pin. That's my big concern. It's like, when are you actually able to do this? Well, let's take a peek under the meta head, shall we? Let's actually look at the brain that they have right now, which is their AI model, Lama, right? The reason why we haven't heard from the Lama models for like the last, I think three or four months, which in the AI world is a heck of a long time is because they just suck, Josh. Like, I'm going to pull up this, this, this, this, um, bar which basically compares Maverick 4, which is their best thinking reasoning model. And it sucks compared to an old GPT4 model,

to Gemma, to Gemma, sorry, and through Claude 3.7 Sonnet. And if you're wondering, hey, Ejaz, why are you talking about old GPT models, Gemma models from Google and Claude models? It's because Lama hasn't released anything that is as good as the latest models from those same companies,

which tells me that Zuck and Meta are in a bit of a rut when it comes to forming that brain that you're talking about, Josh. Their AI model isn't good enough. And you mentioned that they were secretly working on a much, much bigger model, right, called Behemoth. And they didn't end up releasing it supposedly because it just wasn't good enough. And then, coincidentally, they went on this hiring spree spending $30 billion to hire, I don't know, whatever, 50 people or however many people they hired. and that was by already having an AI team, which has now gone quiet. Who is the guy that led that old AI team at Meta? Do you remember his name, Josh?

Oh, I am forgetting. I know what his face looks like. I'm forgetting his name, but all I remember is he would argue against all the top AI researchers at competitor companies who were saying, no, like, you have to be aggressive, we need to spend a lot of compute, we need to invest so much money in this. And he would just be like, nah, like there's a different architecture we can use, I haven't discovered it yet, but I'm almost confident that the way that you guys are doing it isn't the right way. That's Yan. Jan Lecoon. Yeah, Jan Lecoon, exactly. And fast forward six months,

and you've had people like Elon Musk, who was late to the race much later than, you know, meta-AI when they were kind of already focused and created this team with Yan Lecoon heading it, race ahead of them and beat them and spend more time on compute and more money on compute. So I guess the point that I'm making at is they've dropped the ball, creating the brain. Now, that's my bear case. The bull case is they've hired away all the smartest AI researchers

from the competitors that give them the best shot at building the best brain in the future. And my bull case tells me that they're not going to create it in a few months. They're probably going to create it in six months because remember, they're not trying to create Lama 5. They're trying to leapfrog

all the existing models that exist today. that's going to take a bit of time. That's going to take a ton of money. And he's spending the money, and hopefully he's spending that money on compute. So my guess is the next model that Lama, that meta releases is going to be killer. It's probably not going to be open source,

but it's going to be killer. Well, we'll know it's better if it's not open source, because open source, as we've seen from China, is very much a signal of you being at a disadvantage. So if meta comes out, they release a closed source model, it starts crushing benchmarks. That's good signal to look out for. But this, I mean, the episode,

This leaves me with a lot to be desired. I'm like, man, meta, there is a bulk case. And the bulk case is if meta can execute, they are better positioned than basically anyone else. They have the hardware manufacturing capability. They have the software capability. They have more users than anyone in the world. They, if they are able to hit a home run, it is going to be a grand slam. That feels like a low probability.

The reality is they haven't shipped any great software. They haven't shipped any great hardware. They have never created an original product that is compelling, at least in my personal opinion, and they have spent tons of billions of dollars on talent that seemingly isn't making any progress fast. Okay, so what you're looking at here is a graphic or a video showing how big the data center that meta is building currently will look like. It is the size of Manhattan, pretty much, or like the good stretch of the main island. So it's huge. And this is set to produce, I believe it's five gigawatts of compute, which is, don't get me wrong, a huge amount.

But if I were to compare it to Colossus 2, which is what Elon Musk's company XAI is building right now to power their next frontier model, I think that's going to be between 5 to 10 gigawatts. And, you know, Elon's already thinking about what the next data center is going to look like. And he's delivering on that really, really quickly, by the way. Like I think I read a statistic, Josh, where in the first five months of, building Colossus 2, he had already delivered 500 megawatts of power, which is just crazy and unforeseen, because usually it takes like a couple years to even set the foundations before you can start powering this up. And, you know, Jensen has said many times that Elon is the best at scaling

these things. Zuck, I haven't heard of Hyperion their data center since they announced it. Like, they had an extended timeline of going, and they gave a vague post about this, Josh, 2030 and beyond. That's when we're going to be like kind of like done with it. With the first two, two with the first gigawatt sometime in the late 2020s. And I'm like, dude, you need to ship this now. Yeah, I can say this with 100% levels of conviction that if meta has to compete directly with other companies with no edge and they have to compete on the front of building data infrastructures, they are going to get absolutely destroyed.

There's just no way. XAI is moving at an incredibly fast velocity. Open AI has partnered with the entire planet to help them build these data centers. Is there 25 gigawatts at this point? Yeah, 25 gigawatts. They're getting money from Saudi Arabia. They're getting money from Broadcom they're getting. They have custom chip fabs.

They're vertical integrating. They are going to absolutely slaughter meta on this front. So if meta is competing one-on-one on this and they're trying to win here, it's just not possible. They have to win elsewhere. They have to win on the software stack. They have to win on the hardware integration. Because this is cute.

But like, okay, go build it. Go build your 5 gigawatts. I love that you are aiming for that. Go build the damn thing. They're not making progress. So, yeah, this is pointing on the data center front, which is, which is, which is hard because it is so critical. And they have the wherewithal to build it.

They just clearly are not building it and just can't really compete at these levels. Can I give you my low IQ bold thesis, given that we've spoken about everything Meta AI right now, why I'm still foolish. The lower IQ, the better. So what are we got? Okay.

Okay, I have two simple arguments for why I think meta is a good buy right now. Number one, I've never seen Mark so aggressive as he is today. and he's putting his money where his mouth is literally. He is committed to spending hundreds of billions of dollars, probably hitting a trillion dollars over the next five years, to win at this game. So that gives him a high probability of winning.

It may not look like he's winning in the first couple of months, but who cares on the grand scale of things if this thing is going to be a life-world changing thing? Okay, number one is his aggressiveness. Makes me very bullish. Number two, meta's kind of too large, fail. And I might end up eating my words a year or two years from now, but they have so much money that is coming in from other sectors of their platform that isn't AI-related that they can

use to keep funding this thing until they hit it. So will they hit it the first couple of times? Maybe not. It might suck, but I bet you they're going to hit it on the second, third, or maybe even fourth time. Those are my two reasons why I'm bullish on Mesa. Okay, so we have a fun surprise to give to the listeners. But before we do, I just, I need to ask you, final question, which is, as a meta stock holder, are you planning to continue holding meta stock? Yes, I am. Okay. Nice.

I feel like that, that's interesting because to me, I concluding this episode, I'm bullish in the sense that I believe the number will go up. On a relative basis, compared to other opportunities, I am very bearish. To invest in meta at a $1.8 trillion valuation when a company like Tesla is at 1.3, It just feels like a misallocation in terms of potential returns, risk-adjusted returns, perhaps. But I do believe the case that meta will absolutely go up because AI is growing so fast. The GDP of the planet is growing so fast. And they will most certainly capture some percentage of that.

Just by being just by their nature, like you said, just being such a colossal company, they stand to benefit from the up wings of AI. So I guess at the end of the day, we're both probably bullish on meta. We both believe the company will go off. up, regardless of where their downfalls are, the stock price, if we're just talking on a basis of that, seems like it's going higher. But let's get to the little gift that we have, EJAS. You have sourced us some magical codes, right?

For the actual good product. Yes, we have. Okay, so if you remember, my most bearish take on this episode was like the meta vibes app sucks, but their competitor, Open AISORA, is absolutely awesome. And I'm loving it. And I'm still creating videos on their app, and it is like AI generated TikTok. But Josh, I'm sorry, it makes me laugh. And I love making cameos with my friends.

And I have seen so many SORA videos go viral on TikTok and Instagram reels that it's just too good to use. But there's an issue with this. Not everyone has access to it. They're keeping it super locked down and you need an invite code. So we went to our friends at Open AI and we said, hey, like, you know, we have a bunch of really awesome listeners who listen to this episode right till the end. like you guys are right now, and they deserve something special. And they were like, you know what?

We'll give you 500 SORA codes. And that's what we have. And if you guys want a SORA code, we just have a few very humble requests. If you enjoy our episodes, if you enjoy listening to the Limitless Show, please like, please subscribe to the videos and to the episodes that you listen to.

But most importantly, give us a five-star rating on whatever listing platform that you listen to this on, or if you're on YouTube, give us a thumbs up, subscribe to us. And if you DM Josh and I or the limitless account proof of this, you can reach us out on X or on our personal profiles or whatever that might be. Or if you comment and tell us that you have done this, we will send you a code. Josh, is that fair? I think that seems fair.

Yeah. So the intention is like, okay, we're starting. We have a fairly small show. We're trying to grow a small show. And what's really helpful in growing a small show is when people who enjoy it, they signal to the platforms that. amplify our shows that they enjoy it. So therefore, the platforms will want to amplify our stuff more. So when you go on a platform like Spotify and you rate it five stars, it really helps us

kind of elevate the rankings and get more distribution for more people to hear the show. So as a thank you for the service, and it doesn't even have to be five stars. If you think we're four stars, fine, I'll take it. But if you have a five star, it's really appreciated. But if you can take the time to do that, it is so very much appreciated. We, EJES was like very gracious in getting us these codes from the Open AI team. And it's like a fun little exchange that we could do. So we have plenty. We'd love to give them out.

We'd love to let you try them. As always, thank you for making it to the end. This was a little longer than usual. So thanks for sticking with us. And just being here for the journey. We are on a one-way trajectory up into the right. And it is thanks to all of you supporting us through the way. So tell us if you're going to buy your cell, please.

Yeah. And are you bullish-bearers on my stock? I'd love to know. Give me some price targets too. Are we crazy? Are we crazy? Maybe I'm just delusional about. in this bubble. But yeah, that's been the meta episode. Thank you all so much for watching,

for supporting for all the things. And yeah, we'll see you on Sora, if anything. And also the next episode. All right, peace, guys.