Transcript
0:00 The question on everyone's minds, should you buy the SpaceX IPO? By the time you are hearing this, chances are SpaceX is about to go public or they have just gone public at what is [music] going to be the largest IPO in history and it's not even close. I think we have very strong takes and opinions on where SpaceX goes. I find the question that everyone is asking me is, should I buy is this overvalued? It's going to be trading, [music] I mean right now close to 2.1 trillion dollars beforehand.
0:25 That's an outrageous valuation and I feel very confident that we have some pretty good takes that to point in the direction of actually, yeah. Not only is it a good buy, in fact, it's undervalued and a lot of the news that we got this week from the S-1, from the new SpaceX releases backs that up. So, there's a lot of really strong prevailing winds pushing SpaceX forward and we are going to now address the question, should you buy SpaceX shares, should you get invest in the IPO and more importantly, why?
0:54 >> Yes, the answer is yes. Um, but before we get into [laughter] it, let's talk about the numbers of the IPO very quickly. So, they're raising 75 billion dollars, which is three times larger than the largest IPO that has happened so far. That was from Saudi Aramco back in 2019 and they raised just under 30 billion dollars. This is roughly around 3X that, but the demand for the 75 billion dollar raise is already 4X as you're seeing on the screen. There's over 250 billion dollars worth of demand with one institutional couple multiple institutions offering up 10 billion dollar chunks just to get involved here.
1:30 Now, if you're looking at like the price per share roughly, um, SpaceX's S-1 prospectus has put it at around 135 dollars a share. Now, what's interesting about this is typically with an IPO prospectus, you provide a range. You provide a range and you try and find where the match is. SpaceX just went right to the high end of its range and said, "It's 135 dollars per share." which values it at a at 1.75 trillion. But, recent reports from Financial Times and Bloomberg says that this number is actually now close to $162 per share, which gives them a crazy $2.1 billion valuation. Now, the demand for this thing, Josh, is absolutely insane.
2:10 They reserved I think it was around 30% of the share allocation that they're IPOing with for the retail, for the regular average person, and that was distributed to the likes of financial institutions and banks like Charles Schwab, as well as uh brokers such as Robinhood. And the demand initially required you to have like something crazy like half a million dollars in your bank account in some cases to get access to this thing, um but they reduced that kind of like limit, but it is the most in-demand thing. SpaceX is the most in-demand uh company in general, like whether it's just to buy the IPO or whether it's just to like work there. It is a a crazy scene that we haven't seen before.
2:45 Largest IPO raise ever. >> Yes, and it is also by far the most accessible IPO ever. We mentioned this in a previous episode, but 30% of the entire share float is being made available for the public, for retail, as a part of this IPO. This is three to five times more than the average, and that means that a lot of people are able to participate. So, there's a chance if you're watching this, you've actually been allocated IPO shares of SpaceX at the listing price. You don't have to pay that premium. We're going to see the market making happen today as you're watching this to see what is actually going to launch at, but I suspect a lot of people watching this episode are actually already shareholders and owners of SpaceX, and that's such a cool thing that everyone is able to participate.
3:20 Now, in terms of the valuations of SpaceX, why we believe this is undervalued, why we believe SpaceX is going to be such a remarkable company, there's a few pillars. The first is Starlink. It is internet in space from the sky. The second is manufacturing, their ability to manufacture Starship and get payload to orbit at a very low cost per kilogram. And the third, which is one that I really want to talk about right now, is AI data centers in space. We've talked about this a lot. What does this look like? How does it work? Well, earlier this week we finally got an answer, and it comes in the form of a satellite named AI 1.
3:55 AI 1 is the satellite that SpaceX is going to park on Starship, send into orbit, and actually put these things into space. And they're going to be training data centers, they're going to be beaming data back down to Earth. And now we finally have a rendering and we have some specs on what it's going to look like. In terms of power, there is a 150 kW peak with a 120 kW average on each one of these satellites. For reference, that's about 100 American homes worth of power per satellite. Um and then we're getting about 70 kW per ton. So, that's the efficiency part of the launch metrics where the more energy we can get per ton, the cheaper it costs. It's it's lowering those costs. We have compute slots that are interchangeable. And I think this is a really underrated part is the GPU modules that you're seeing on the screen on that animation we were just showing, they're actually hot-swappable. So, you could have Nvidia chips, you can have um Google chips, you could have Amazon chips there. What they're going to test this with initially is going to be Nvidia's Vera Rubin platform. That's going to be on version one of these things. And in terms of the radiator, because clearly no one believes that SpaceX is capable of figuring out thermodynamics in space, given the fact that they have how many thousands of satellites currently operating in there.
5:05 Well, they figured it out. They said they're going to have 110 square meter liquid radiator. And basically it's these gigantic sails that we're seeing on screen right now to dissipate the heat into deep space. Deep space is a huge place, even if you're in lower Earth orbits, geosynchronous orbit. And the wingspan of those will extend to about 70 m wide, which is wider than a Boeing 747. So, this is AI data centers in space. When we say that, this is what it looks like.
5:31 This is how it's going to function. It's a really big deal. I I want to shine a bit more light on this satellite, because previously, like Elon has just made bold claims saying that like, you know, the valuation that we're raising SpaceX at now at a $2.1 trillion valuation is primarily based on AI data centers in space. And then they kind of signed a few deals recently with Google and Anthropic saying, "Yeah, they're going to be buying these data centers in space." But we still didn't know how.
5:54 Now we have the answer and we have a bunch of the numbers and how it's built. Now, to be clear, what does the satellite look like? Picture a refrigerator that you have at home with 70 m wings. This thing is 95% so big. Basically, and that refrigerator is essentially a GPU server rack. It stores around 72 Nvidia GB 300 GPUs and the total compute power that it has is around 150 kW. Now, immediately after announcing the satellite, Elon also stated that he's filed to launch 1 million of these data center racks, these satellites, these AI 1 satellites into space. If he is able to pull that off, the total compute power that will be orbiting in space will be around 120 GW. That's around three to four X of the total data centers that currently exist on the entire surface of Earth right now. So that is a a hell of a lot of compute. But more so, currently there's around 14,000 satellites in total existence across every single other company in low orbit space right now. So Elon is like aiming for like multiples more of that for this particular satellite. Now, the wingspan is really important. You must be thinking like, you know, 70 m is absolutely ridiculous. Radiation, you need to disperse the heat. It's kind of like hard to do that in space. In fact, it's impossible to do that in space. So you need some kind of like really ingenious um cooling mechanisms. That's what that wingspan is for. Yes, it is large, but also you're in space. You have all of that space to be able to do that. And the way that these satellites are designed is different to the ground data centers in the sense that they're going to be beaming lasers between each other to form an interconnected grid that they can transfer data between each other. And that is the sole mechanism that is required to train and inference models, which is the entire point around SpaceX. Now, it's important to understand when you're looking at the valuation, when you're looking at the price, you might be thinking, "I don't know if this is overvalued. It kind of seems like overvalued, even though I like what Elon is doing." You have to understand that the whole premise is you're getting three companies for one share. It's SpaceX, yes, to be the kind of like highway to space, but then you have like the payloads, which is like all these GPUs that they're going to be training on the sun's energy, becoming a Kardashev type one civilization, which all sounds grand, but like have the numbers and the mechanism to be able to pull that off.
8:12 And then you have the model itself under xAI, and then you have the distribution through social media platform, through app stores, and all that kind of stuff. And it's not like Elon is depending primarily on his own companies to service all this demand for Grok entirely. He's signed major deals in the last, I think, 3 weeks with Anthropic, Google, and Cursor, which equates to around $56 billion per year. So, the previous critics were like, "I don't know if SpaceX is going to be profitable in general." Now we have the answer.
8:39 Like all this money that's being come from essentially Neocloud payments for Elon's existing Colossus 1, 2, and 3 data centers on Earth is paying for itself. It's paying for this entire expedition. >> Oh, it's an absolute monster of a company. And now I want to talk about a second pillar, because we've seen AI 1, what that looks like, what data centers in space look like. We also have visual graphics of the new Starlink satellites. And we've been talking about these new Starlink satellites for a long time now.
9:03 Version 2, and most recently version 3, which is going up on Starship. And the stats and analytics between these two, comparing them, is pretty remarkable. The the first generation of this, version 1.5 of Starlink, had 24 gigabits per second download, which pretty small. If we look at version 2 of the satellites, which are what's available today, that's what's going up in the Falcon 9 rockets, the smaller ones, those have about 96 gigabytes per second, and they're able to deploy about 27 of those per launch. So, that gives you about a total per launch of 2,600 gigabits per second. Now, when you look at version three satellites of Starlink, that's where the numbers get pretty silly. It's a 20 times multiple per launch on bandwidth relative to version two. This is enabled with Starship. We talk about Starship all the time.
9:51 They've been doing really successful test runs. They're going to continue that program. Once it starts becoming reusable, they're going to put these satellites on it, and it's going to deliver 61,000 gigabits per second to the internet. That is 60 satellites. That is a 20 x multiple. That is how you get to a true World Wide Web that exists in space. And when you think about what SpaceX is doing here, they're creating a high bandwidth clone of the internet that exists here, except in space. What happens with the internet here? We actually run wires underground through the oceans to connect the world. If we have satellites, you remove a lot of the threat vectors, you remove a lot of the censorability. You just You create this truly open and accessible network that's not only available to anyone with infrastructure, but available to anyone with access to the sky. And building that and making that available to everyone is such a remarkable business.
10:41 And we've seen that reflected in the Starlink subscriber numbers. We have a chart here that shows the Starlink growth, and it has been the most perfect exponential chart I think I've probably ever seen in my life. And it is the Starlink subscriber growth. As of When is this? May 23rd, it is now sitting at 12 million users. And you could see they've been adding an exponential increment. I mean, the latest 2 million users were added only in the last 81 days. They're averaging 25,000 and climbing. Elon and the Starlink team, they just recently revealed on an interview that they published a few days ago, the new version of these Starlink terminals, which are even smaller, even higher bandwidth, even more accessible in terms of price. And it's just this unbelievable service that it's going to be a no-brainer not to have. If you ever require internet anywhere that doesn't need fiber optics, you should be a Starlink subscriber. And that's what we're seeing in the growth. So, it's really it's an unbelievable company.
11:34 Starlink is one of those pillars that's just like it's it's crazy what they're doing. And no one else can come close to this. We we flew Virgin, which huge fan of. I really hope they figure it out, but their rocket just exploded on the launchpad last week, and their whole program set back, and they're the only ones that were close. Even if everything at Blue Origin worked as perfectly as they hoped for, they're still five years behind at minimum, probably 10. This is a full decade lead they have. So, there's essentially a monopoly on space that SpaceX owns. And as it becomes more in demand to send mass to orbit, there's only one company that can do it. And they can do it damn well.
12:06 >> So, this might be a crazy take, but Elon's uh superpower actually it's not a crazy take, this is obvious is scaling hardware from zero to one better than anyone else on the entire world. He's done it with electric cars, he's doing it with robots, and now he's doing it with space rockets. Now, there's another advantage that he has which translates directly from that, Starlink, right? So, you just mentioned that. Um you can repurpose Starlink's production lines to make this new AI one satellites. In fact, arguably, it's easier. See, when you're creating these Starlink satellites, they have this like phased array antenna, which basically is meant to like speak to the dishes that are on Earth to kind of like communicate and beam this internet back down to Earth.
12:48 That is a really tricky thing to produce, but the satellite itself is fairly easy. Now, with the AI one satellite, it doesn't need this phased array antenna. In fact, you just need the laser projections to be able to communicate and talk with each other. I mentioned it earlier, like transferring training your model, inferencing the model. That is arguably easier to do. And so, Elon already is not starting from scratch here. Normally he does zero to one, he's kind of starting from 0.5 to one to be able to pull this off. Now, in terms of timelines, um I think he's aiming for the end of 2027 to have a full gigawatt worth of these AI one satellites out in uh space. So, this is a very real thing. The timelines are, in my opinion, much shorter than we think, and he already has some of the world's leading AI labs committed, at least in theory, on paper, on in contractually, to pay and purchase access to these things. Like you mentioned earlier, these AI on satellites are GPU interchangeable. It's not just built for Nvidia, although that's the primary kind of case right now. You can fit TPUs, you can fit Amazon Tranium in the future.
13:47 He's meant to be the tall man, the transporter for all these GPUs to be put out there to kind of absorb all of that sun's energy and train whatever model. He's He's kind of like agnostic when it comes to that, and I think that is one of the biggest bull cases for SpaceX as I see it today. >> Yeah, think about how difficult it has been to build a data center. You need the What is it? Land, power, shell in order to get a data center sta- set up, and it is so remarkably difficult to get the approvals, to get the public sentiment to turn positive, to figure out where to source the energy from. It is this incredibly complicated process that's difficult and no one wants to deal with. And if you could just go over to Elon and say, "Yo, bro, I got um a couple gigawatts that I need in orbit.
14:24 Can you send them for me in the next few months?" That is such an easy outcome, and one that I think everyone will want to to use, because why wouldn't you? And I want to circle back to the IPO and why I said earlier in the show that a SpaceX should be the most valuable company in the world. We We barely even talked about xAI, and we barely even talked about the other entities that exist underneath it. But the idea is that as as AI continues to expand and continues to take up more of what we do on a day-to-day basis, it's very much accelerating in that world of bits that we talk about a lot. It's able to do ones and zeros very effectively and very efficiently. The difficult thing now, the bottleneck shifts to the physical world. of atoms where you have to actually build things. You have to create the physical manifestation of this AI to to allow it to be embodied or to give it enough energy and resources to train itself to become this AGI, the superintelligence. And when it comes to manufacturing, when it comes to real-world engineering, there is really no better team in the world than SpaceX in terms of hardware engineering. And when I look at other engineering teams who are really good at creating physical objects in the real world, manufacturing factories which are actually the product, and the factories produce these goods and services, but building the factory is the hard part, I also look at Tesla. And Tesla has done a remarkable job at building and manufacturing these objects at scale. And what is Tesla now working on? They cut down their Model S and their X line in exchange for Optimus, the humanoid robot. That humanoid robot is going to scale very quickly. That humanoid robot paired with Grok and XAI as the orchestrator is going to build this like very autonomous capability of actually physically manufacturing things in the real world. SpaceX and Tesla are probably going to merge together. In the case that those two companies come together, there is no instance in which they do not become the most valuable company in the world so long as they deliver on the promise. And I think the way you look at these companies, the way you value these companies is you kind of like look at the road map and you say, are the products and services that they offer in the future going to be significantly better than they are today? And if the answer is yes, then the idea is that the company should be significantly more valuable to match the added value that they're going to increase.
16:22 I very strongly feel that way. We have Starship. We have Optimus, the humanoid robots. If just those two things work and nothing else works, the amount of value that will be unlocked is unimaginable. It's really hard to fathom how large these companies can get. And we we have this perceived limit of what a market cap of a company can be. Nvidia is close to $6 trillion or something, but to imagine a company reaching 10, 20 trillion dollars, 30 trillion dollars doesn't seem unreasonable when you think about the scope and scale of a company like Tesla, of a company like SpaceX merged together under the rule of Elon Musk. It's just this unbelievable opportunity that finally the public has access to, and that's why I am particularly excited to invest in the SpaceX IPO. I I I think for the longest time uh the SpaceX IPO has been based on a lot of visionary ideas, um kind of belief and this undying faith in Elon, but I do want to touch more on like what the potential numbers might look like to make that multi-trillion dollar valuation that you're talking about, Josh, an actual reality. So, there there's two there's two kind of like perspectives I want to share here. So, like number one, Hamid's over here kind of like shared his like breakdown of like what each of these satellites basically equate to. So, if you launch one of these AI one satellites out into space right now today, it's the equivalent of basically half a rack of GPUs that you currently find in a data center here on Earth. So, it's around 120 to 150 kilowatts, it's around half a rack. Now, today's data centers, let's say it's worth like one gigawatt worth of compute, that's the equivalent of 4,000 racks at around 250 kilowatts. So, you would need to launch 8,000 of these AI one satellites into space to be able to equate to one data center one gigawatt worth of compute. And Elon is targeting that by the end of 2027, which is a lot sooner than most um people would have expected. Now, there's a lot of costs that go in this. The average launch cost actually, you know, if you're being kind of like medium to kind of like conservative is around 10 billion dollars just for like the launch cost and not the materials at all. Now, Elon's whole idea is that he's going to bring these launch costs down pretty dramatically over the next 3 years. And at 3 years in, he's saying that it'll be more cost effective than building a data center here on Earth. And so, if you do back in math with that in mind, right? 3 years from now or maybe less 3 years from now, a Vera Rubin rack right now, which is like Nvidia's latest GPUs that they're um rolling out, is worth around 8 million dollars. Now, if you couple that with the launch cost at around 200k launch cost per rocket, which is what Elon's aiming for, we are going to hit at around 10 million dollars to put a single Vera Rubin rack out in space. So, 12 million dollars on the ground, 10 million dollars in space, you can see a world where the cost economics end up becoming favorable, especially if you're not just using Vera Rubin racks. You might be using cheaper Google TPU rocks all Amazon training up. So, the idea is this is a big bet now. If you're purchasing this IPO, you aren't just in it for the near term, you're in it for this next three to honestly 10-year window where AI data centers in space become a much real thing, much more real thing.
19:27 >> Yeah, and this is a roundup. So, we're going to wrap this up in a second and get some more news. But, the final point that I want to make is you mentioned the cost of reusability for Starship. If you look at the Falcon 9 rocket, there's two stages and only one of them is actually reusable. So, there's still is a higher cost. With Starship, 100% of the rocket is reusable and the actual cost to launch it in the case that 100% is reusable is just the cost of fuel. And the cost of fuel on a Starship rocket is actually cheaper than the cost of fuel on something like a Boeing 747 because a 78% of it, I think, like almost 80% is just liquid oxygen. Liquid oxygen is like very abundant. We breathe it all the time. It is just a liquefied form of oxygen with the remaining 22 and 22% being just liquid methane. And methane is like it's the stuff you like cook on your grill with. So, this isn't very expensive. It's not crazy to think that they will need a lot of money to launch these things. And the idea is that that will bring the cost per kilogram down to sub $100. And that means if you are It does if you want to go to space, depending on how many kilograms you are, you just divide that by 100 and you're on your way. It's a cheap ticket. So, >> [laughter] >> So, yeah, go on a diet so you can get a cheaper ticket to space. But, that's I think that's kind of like the bull case why I'm excited. I have to ask you as we wrap up this section before we get into the next one, which is going to be Open AI building this like gigantic data center, is Ejas, are you going to be participating in the SpaceX IPO?
20:45 [snorts] Are you excited about it? Are you going to buy some? What What are your general thoughts? >> So, if Robinhood and my broker allows me to, which I've signaled interest, absolutely I will be as this goes live today. My thoughts in general, I wouldn't bet against Elon when it comes to hardware scaling. Like he's less proven on the software side of things. Like this is This is just me trying to be balanced here, but he has proven himself time and again on hardware scaling, and SpaceX is solving a hardware issue. One of the biggest hardware issues or challenges ever with AI data centers in space.
21:18 As everyone knows, I'm a big AI fan. So, if he's able to pull this off, more compute equals better models, better inference. He's He'll be able to serve agnostically any AI company. It's an obvious buy for me. >> Yeah, I think I'm I'm pretty aligned. Me, too. I have been So, I got lucky in the case that I had some private shares from about 5 years ago. So, I have been a SpaceX holder. I continue I plan to continue that going forward. I also requested some of the IPO. I'm also just planning to continue to buy. Like, dude, I can get a 10-plus trillion-dollar company for 2 trillion bucks. Sign me up. I think the people who are a little more time sensitive, who are looking to get a return sooner, that might be a bit iffy. I mean, $2.1 trillion is high.
21:53 There's still a lot of variables that need to be worked out. There's a chance that this thing goes public, shoots up, and then falls for a little while. I don't know. We're not traders, but I think as long-term investors, as as someone with a a low time preference, SpaceX at $2 trillion, two thumbs up for me. Let's keep the train going. Anyways, we probably should move on. This is around up. This is Friday. We got to fill the people in on what's going on. E just there's some news from OpenAI about a new gigawatt data center. We're back on Earth, by the way. We're We're talking about data centers on Earth.
22:22 What's the deal with this new OpenAI news? >> Yeah. So, um it's We're not We're out of the world of single-digit gigawatts now, Josh. We're We're We're in double digits. So, >> Run it up expeditiously. >> OpenAI has basically signed a major deal for a 10-gigawatt data center. For context here, the largest gigawatt data center, which is what Elon's building with Colossus uh 3, is aimed to be around 1.5 to 2 gigawatts worth of data centers. Um for a single site to own 10 gigawatts of compute, you need roughly around $500 billion. Now, obviously, OpenAI, they haven't IPO'd yet. Uh they might be pretty soon. Um they don't have that kind of money, so they're being backstopped by their best friend SoftBank and of course Nvidia. Jensen Huang is stepping up to backstop this entire data center campus in Ohio. It's going to be an absolute gargantuan beast of a site and it kind of follows in the uh trend of Sam Altman and OpenAI specifically, who have been incredibly aggressive about acquiring as much compute as they can to serve their models. Now, if you compare them to Anthropic, you might notice sometimes when you're using Claude models, the performance gets a little degraded or you might get limited or you might hit your limit really quickly. That's because Anthropic has, you know, fairly been conservative with computer and that's given them advantages in other areas. OpenAI has been the opposite and that's why it's always kind of like relatively free and flexible to use as much of your rate limit at OpenAI as you can, which has led to arguably a better developer experience, a better coding experience, a better LLM experience in general. And Sam is doubling down on that bet. He believes in the future, the demand for AI is going to outstrip the amount of available computer supply that he has currently on Earth. He is kind of not on good terms with Elon. You could say, you know, they recently just got out of a court case. Uh so, he's probably not going to be using data centers in space anytime soon. So, he's trying to secure as much ground space as he can on this new site. Now, what worries me a little quick uh a little about this is just the pure amount of money that you need. Now, obviously this is going to be done in phases. 10 gigawatts of computer is going to take absolute years to spin up, but it's an important bit of news that I think kind of lays into this next headline. Josh, I don't know if you saw this, but OpenAI has confidentially filed for an S-1.
24:33 They are going to be IPO'ing relatively soon. >> OpenAI is copying SpaceX with data centers. They're copying SpaceX with IPOs. Doesn't want to work with Elon. All right, I'm getting the trend. Yeah, but it seems like this is this is actually official now. Um, I hope they can build that data center. That would be amazing. But we do have a official filing of the confidential S-1. And there was also a leaked all-hands news report that I read that said um, Sam is kind of using this as an option. There is the option to go public soon, but depending on how their model progress goes over the next few months, they will then make that determination. His perspective is that the world will meaningfully change in the sense that they do get kind of like this runaway AGI self-fulfilling loop type thing, and he wants to be a private company in the case of that does happen sooner rather than later. This is, um, some pretty big grand visions that they're dealing with. So, we'll see, but now they do have the option. They have confidentially filed. Someone out there is reading through this and probably having a field day understanding the internals of how this company actually works. So, that's pretty exciting news.
25:32 >> Yeah, there was another bit of news that, uh, OpenAI reportedly is looking to slash the cost or prices to their subscriptions, which is kind of a a change in the directional trend that we've seen from Anthropic. Um, Google actually this week cut their prices as well. I think we're starting to see, uh, we're starting to reach a point where all these AI models, as good as they are, are starting to become more commoditized. They kind of all do the same things. The leap between the the prior version and the current version isn't as large as it used to be, like, say, a year back. And OpenAI is taking the first step to being competitive towards Anthropic and and Google. So, Google slashed the access to their, uh, AI studio to about $20 a month for their cheapest thing. I think OpenAI is planning to do the same. And I'm guessing like Anthropic and OpenAI are going to get into some kind of a bidding war. OpenAI kind of like taking its time to go public, as Sam suggested in that leaked memo, is very interesting to me because you see Anthropic and SpaceX kind of like racing to raise as much capital as they can. And if Sam wants to pay for a $500 billion data center, he kind of like, I'm confused where he's going to get the money from. Maybe he's just going to keep borrowing privately.
26:33 But, uh, it's just a unique case where OpenAI is is kind of breaking away from the pack, and I'm wondering if these financial decisions actually end up paying off. >> Yeah, we'll see. Um, we got to cover Fable next because it just released this week. You've been using it for a few days. I've been using it for a little while. It's it's a pretty amazing model. I've been using it for a lot of fun extracurricular stuff I found. Like I created the Minecraft clone to play around with it. I've been building kind of games and 3D spaces and I I found it to be really fun. It genuinely feels like this novel model architecture that I've been enjoying playing with. EJ, over the first few days that you've been using this, what are your initial thoughts, opinions, kind of like after the dust settles, how do you feel about the new model, Fable 5?
27:14 >> Um I think it is less sycophantic than Opus 4.8. I'm not going to lie. I When 4.8 4.8 came out, I used it primarily as my main model, but like it was way more agreeable than 4.7. We're now back with Fable 5 and that might sound like it's got nothing to do with intelligence. To me it is. It factors in the personality of the model and I'm able to like work with it longer over certain tasks. It just understands what I want to do. The second thing I noticed is I've been working on a couple of personal projects that involve a lot of code and research.
27:45 Um I can just one prompt uh Fable 5 and it is able to just work continuously for up to an hour. Now, I know software engineers listening to this are like, "Ah, I let my thing run for 5 to 6 hours or maybe even overnight." But for me, for the average person, it is becoming in- incrementally more valuable for a model that understands what you want, that doesn't come to you for every kind of like challenge it hits and just kind of like figures it out for himself. And I think this kind of trend of models doing longer horizon tasks way better than the average human is a subtle trend that is going to be increasingly valuable for corporate workers at enterprises as well as personal people that are like starting their own businesses or want to spin up multiple of these agencies instead of hiring an employee. I just think it's a wonderfully intelligent model and I'm excited to see further iterations of this mythos-grade um kind of type of model be released. But there was some pushback, right, Josh? Like there was some policies that came out that were basically like, "Hey, like you get access to this public version of Metis 5, but it's going to be heavily safeguarded." Now, one of the biggest criticisms that they faced was if you ask it questions like, "Hey, um I'm trying to train my agent to become like more efficient and use less tokens."
28:55 It'll take that as some kind of an inclination that you're trying to learn how Claude Fable 5 is built and uh Anthropic installs these anti-distillation maneuvers. Basically, redirecting you back to the old model, Opus 4.8. But, in this case, it would give you degraded quality intentionally because it thinks you're trying to like steal its blueprints. Anthropic has now walked that back because previously they weren't going to tell you when they were going to do it. And now they are telling you. They reversed that decision as of yesterday, basically saying, "Hey, we'll be upfront with you when we're going to intentionally degrade the quality of of your answer because we think you're trying to like copy our model weights."
29:30 Which is a very political maneuver and it's nice to see see that they backtracked on that. >> There you have it. So, there's the update on Fable and the final topic of the week. We filmed an entire episode about Fable and also this final topic, WWDC. So, if you have not watched either those, I highly recommend going to check them out. Pretty amazing episodes. We had a really strong week of fun topics to talk about this week. But, I do want to trace over WWDC particularly around Siri AI, which seems to actually be as good as they said it would be. And we have to cover this because that comes as a surprise. This is not the Apple of the past. This is the Apple that promises things and doesn't always deliver. So, the people have had to download the beta, actually fact-check, try it for themselves. And the initial reviews have been spectacular. It seems like it actually works better than what they demoed in real time on stage. It actually works faster. It understands context. It works very well on people's phones. And this is available through the developer program that Apple offers through their iOS developer membership.
30:23 It seems like we're actually going to have a pretty powerful Siri experience. We're going to have on-device AI that is actually contextually aware of everything that lives on your phone and off of it. And they actually addressed something that I found really interesting. Maybe you saw, maybe you didn't. It was about their um removal of Siri from Europe and from China in the sense that they will not budge on the privacy thing. They will offload some of the compute to cloud in the sense that or in the case that it can't run locally on your device. It's a little too challenging for that. And when they do, it's done so entirely privately. Europe doesn't want this to happen. So therefore Apple's like, "No, no, no. We're not giving you the back door. We don't even have the back door.
31:01 We are just not going to push it there." And I really admire the like standing strong on the stance of being private, of running everything on device. It seems like Apple has grand slam. And it comes maybe a couple years later than everyone would like, but this is something I am stoked to get. And I think the downstream implications of this are probably pretty big. There's a lot of casual users of something like ChatGPT who pay $20 a month to have an LLM that they basically use as a Google plus plus.
31:25 This now substitutes and subsidizes a lot of that all for free running local inference on device. So, very excited, very bullish about Apple in general. >> You know, it was funny. I was reading the release article for Siri AI. And typically at the footnote, you'll see, you know, "These regions are excluded." It will just be a minor bullet point in tiny, you know, writing. For Europe, they had an entire A4 page basically saying, "This is Europe's fault. Europe government is the reason that we can't release our product. And we are not budging because we just think frankly that this is ridiculous." They It was a larger piece than the China bit, which was just came in as like a a paragraph.
32:05 Obviously, China's like going to want access. >> everyone likes the Europe and China one. >> Exactly. But for Europe, they've been sort of anti-AI and anti-privacy for quite a while now. And it has resulted in slowed-down releases of the latest AI products and features in that region, which is just kind of like sad to see, especially like, you know, that's where I'm from. And it's annoying that like I can't talk to my friends about the latest features because they have to wait like months to a year to get access to the exact same thing. That being said, a lot of the haters for Apple, which included me, which was me a year ago, was like, "I don't think Apple's going to be able to catch up. We have been proven absolutely wrong and I want to make that very clear on camera right now. Apple has a distinct advantage which is they own the entire consumer hardware ecosystem, 3.5 billion devices, and now they have the hardware, the chips to run some of the best frontier models on your phone. They may not be as large as Anthropic's, they may not be as large as OpenAI's, but they're trained on your data and you can do something that none of these other AI labs can do which is ask Siri to basically find that text that your mom told you to do the thing and you forgot what it was or you know, bring back that memory of like when you were hanging out with your best friend doing blah blah blah. It has full context, grounded context on your entire thing and that is a feature we haven't seen and one that I'm personally super excited to use. So, that is all the news for today, folks. Josh, any any final thoughts?
33:22 >> You're caught up. No, that's everything. The vibes have been crazy high this week. The news has been insane. Lots of things that like I think we're probably personally excited about have been released. We got WWDC, we got a new frontier AI model, and we got the biggest IPO in history all in one week. It's crazy. It's like they're having a week. We also got the the Knicks dub yesterday. Please let's not let's not forget for the New Yorkers. EJ and I are both in New York, both big Knicks fans, huge. Hopefully the day after you're watching this, they win the championship. That'd be pretty amazing.
33:52 Given the vibes this week, I suspect that's that's likely. That feels probable just on a vibes-based opinion. Um but yeah, very excited about all the things that are happening. Very excited that you have made it here to the end of the episode and are still listening. So, thank you so much for all the support. As always, if you enjoy this, as always also, don't forget share it with a friend. Anyone who you think might be interested, anyone who asks you, should I buy the SpaceX IPO? Maybe send them this video and say, not only should you buy it, you should set I'm not sure I can't go that far.
34:18 >> Don't go >> [laughter] >> Make your own This is not financial advice. >> Actually, if you disagree with us, if you disagree with us, we actually want to hear from you. Like we have been combing through the numbers, we've been combing through the projections and we're like this kind of makes a lot of sense to us, but if you're adamant that we're wrong, let us know in the comments. We we're hearing. We're going to hear you. We're going to see you.
34:37 We're going to respond to you. So, we want to hear from you whether you hate our opinion or you love it. Like, tell us where we're wrong. Tell us where we're right. >> Yeah, and I'm also going to disagree with you. So, I'm ready to go to war in the comment section. Bring it on, baby. Tell me why SpaceX isn't the most valuable company in the world. Come on. But, yeah, that's it. Thank you guys always so much for watching. Have an amazing weekend. Enjoy the weekend.
34:55 Hopefully, by the time you're watching this SpaceX is is going up only. That'd be great. Um if not, I'm sure we'll hear about [laughter] it. >> Long term. The thesis is long term, guys. >> Yes, we're long term. This is This is a low time preference show, man. I'm here for a long time, not a good time. Or a good time and a long time, I guess. But, it doesn't matter the the timelines that happens on. But, yes. Thank you all so much for watching. Have an amazing weekend, and we'll see you next week for another set of episodes on Limitless.
35:18 See you. See you, guys. >> [music]