November 14, 2025 • 30 mins
In this episode, we explore Michael Burry's massive short position on NVIDIA, questioning if it foreshadows a market correction or reflects AI's dependency on GPUs. We assess Burry's claims of inflated tech valuations against data showing ongoing reliance on older GPUs by Frontier AI labs.
We also touch on Microsoft's moves with OpenAI, the implications of the GPT-5.1 launch, and innovations from Google’s DeepMind.
------
🌌 LIMITLESS HQ: LISTEN & FOLLOW HERE ⬇️
https://limitless.bankless.com/
https://x.com/LimitlessFT
------
TIMESTAMPS
0:00 The Big AI Short
2:39 The Thesis
7:02 Hardware Demand
9:27 GPU Economics
11:42 The Crash and Burn
13:20 Energy Constraints
17:01 Microsoft Owns it All
18:14 ChatGPT 5.1
23:21 Google 3D Rendering
29:00 Final Thoughts
------
RESOURCES
Josh: https://x.com/JoshjKale
Ejaaz: https://x.com/cryptopunk7213
------
Not financial or tax advice. See our investment disclosures here:
https://www.bankless.com/disclosures
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Josh:
If you were alive in 2008, chances are you remember the market crashing, (00:00):
undefined
Josh: a catastrophic failure of the housing market that bled into the entire economy, (00:04):
undefined
Josh: and it wiped out a lot of entities. (00:08):
undefined
Josh: And if you'll remember, there was one person in particular who profited a tremendous (00:09):
undefined
Josh: amount off of this happening, and his name is Michael Burry. He made $100 million. (00:13):
undefined
Josh: Personally, collectively, he made a billion dollars off of shorting this market crash. (00:17):
undefined
Josh: He saw what was right early, and he doubled down on it, and he made a killing. (00:23):
undefined
Josh: In fact, so much so that it became a movie that you've probably seen named The Big Short. (00:26):
undefined
Josh: Now, Michael Berry recently has published a new position in recent SEC filings that we uncovered. (00:31):
undefined
Josh: And Ejaz has been all over this over the last couple of weeks, (00:36):
undefined
Josh: tracking the positions, trying to understand why he's making these things. (00:39):
undefined
Josh: And if this bet is correct in predicting the next big bubble to pop, (00:42):
undefined
Josh: which is seeming to be AI. (00:46):
undefined
Josh: So he predicted the housing bubble in 08. He made a billion dollars. (00:48):
undefined
Josh: He's predicting another bubble in 2025. five, is he going to make another billion dollars, EJS? (00:52):
undefined
Ejaaz: The short answer is, I don't think he will. I think he's going to get blown out. (00:58):
undefined
Ejaaz: So Michael Burry is back. He is short around $300 million worth of Nvidia shares. (01:03):
undefined
Ejaaz: So he's shorting, at this point, the richest, wealthiest company in the world. (01:09):
undefined
Ejaaz: And that is a big sign, basically saying that I think the air bubble is popping. (01:13):
undefined
Ejaaz: So the question is, is it? (01:17):
undefined
Ejaaz: Let's go through his thesis, Josh. I'm going to simplify his post here. (01:19):
undefined
Ejaaz: Which is basically GPUs run the AI world, right? (01:24):
undefined
Ejaaz: Trillions of dollars have been spent by Frontier AI labs collectively to train state-of-the-art AI. (01:28):
undefined
Ejaaz: But the thing with these GPUs, Josh, is that they have a lifespan, (01:34):
undefined
Ejaaz: right? They don't last forever, right? (01:39):
undefined
Ejaaz: And he estimates that the top Frontier labs. (01:41):
undefined
Ejaaz: Overestimating or artificially boosting the lifespan of the GPUs that they purchase. (01:46):
undefined
Ejaaz: And you might be like, well, that sounds boring. Why is that interesting? (01:50):
undefined
Ejaaz: Well, the thing is, if you have assets on your balance sheet, (01:53):
undefined
Ejaaz: that factors into your stock price. (01:57):
undefined
Ejaaz: So the point that he's making is all these big companies are artificially boosting (01:59):
undefined
Ejaaz: the lifespan of the GPUs that they've purchased so that it inflates their stock price. (02:03):
undefined
Ejaaz: And therefore, the real stock price is actually a lot lower. (02:08):
undefined
Ejaaz: In fact, he says at the bottom of this tweet, by 2028, Oracle will overstate (02:12):
undefined
Ejaaz: earnings by 26.9%, Meta by 20%, etc. (02:16):
undefined
Ejaaz: And so his big short, Josh, is on the behemoth that is supplying all these Frontier (02:20):
undefined
Ejaaz: AI labs, NVIDIA, saying that when the bubble pops, and he thinks the bubble's (02:25):
undefined
Ejaaz: going to pop now, he will make a heck of a lot of money on this trade. (02:29):
undefined
Josh: So is he saying the bubble's going to pop now, or is it going to pop in 2028? (02:34):
undefined
Josh: Because he's referencing those numbers in 2028. (02:37):
undefined
Josh: So I'm kind of curious how he's framing this is this like because i'm now i'm (02:39):
undefined
Josh: thinking about our positions um is this something that he predicts is going (02:44):
undefined
Josh: to happen in 2028 or is this more of a short timescale position if (02:47):
undefined
Ejaaz: We remember from his original short i think he held a position for about a year, (02:51):
undefined
Ejaaz: Or at least like the better part of a year. So I think he probably estimates (02:56):
undefined
Ejaaz: that it's going to happen sometime within that time frame. (03:00):
undefined
Ejaaz: But I have more evidence as to like why he might be thinking that the bubble is bursting. (03:03):
undefined
Ejaaz: Have you heard of what a neocloud is, Josh? (03:08):
undefined
Josh: The name is familiar, but I don't understand quite what that means. So please fill me in. (03:12):
undefined
Ejaaz: Okay, so think of like AWS or any of these other typical cloud providers. (03:17):
undefined
Ejaaz: But they're specifically focused on providing GPUs and data center pipelines for you. (03:21):
undefined
Ejaaz: So think of like an AWS just for AI compute. That makes sense? (03:27):
undefined
Ejaaz: And so what he did was he looked at the major NeoCloud providers and he looked (03:31):
undefined
Ejaaz: at their stock prices over the last week, Josh, and they were down. (03:38):
undefined
Ejaaz: So he upped his position and was like, okay, this is the bubble bursting. I'm gonna lay my claim. (03:43):
undefined
Ejaaz: There's an issue with this thesis, Josh. I think he's dead wrong. (03:49):
undefined
Ejaaz: And I'm about to walk you through four reasons why he's wrong. (03:53):
undefined
Josh: Okay. So then just to kind of like wrap my head around this, (03:57):
undefined
Josh: the NeoClouds are kind of infrastructure for people who don't want to build (03:59):
undefined
Josh: their own infrastructure. (04:02):
undefined
Josh: Kind of like you described AWS. If you want to get a data center, (04:03):
undefined
Josh: but you don't want to build a data center, you offload that responsibility to a NeoCloud. (04:06):
undefined
Josh: And he thinks that the cost that these companies are marking down these NeoCloud (04:10):
undefined
Josh: prices at, particularly the GPUs is much higher because the depreciation happens (04:14):
undefined
Josh: on a shorter timescale than the companies are writing off. (04:19):
undefined
Ejaaz: Correct. And NeoClouds is one part of the picture, but it's a great example (04:22):
undefined
Ejaaz: to lead with because so many frontier AI labs like Microsoft OpenAI actually (04:28):
undefined
Ejaaz: pay these NeoClouds billions of dollars. (04:33):
undefined
Ejaaz: Like Microsoft just signed a $19 billion contract with a NeoCloud provider called (04:36):
undefined
Ejaaz: Nebius, which kind of sent its stock price going up. (04:40):
undefined
Ejaaz: We talked about it on a previous episode. so like it's a (04:43):
undefined
Ejaaz: good example to kind of like lead with but the (04:46):
undefined
Ejaaz: take here is that these neoclouds are (04:49):
undefined
Ejaaz: basically as you said overestimating the life cycle of these gpus and therefore (04:52):
undefined
Ejaaz: they are wrong except that's not the case at all and i present to you my first (04:57):
undefined
Ejaaz: counter thesis which is he's wrong about the two to three year depreciation (05:02):
undefined
Ejaaz: cycle it's actually more like five to six in some cases even eight years. (05:06):
undefined
Ejaaz: These GPUs are used for more than just training AI models. (05:11):
undefined
Ejaaz: They are used for inference, and they're sometimes used just for general queries (05:15):
undefined
Ejaaz: and distribution for these AI models, right? (05:19):
undefined
Ejaaz: But the most staggering kind of example comes from the top doc, (05:22):
undefined
Ejaaz: the top neocloud called CoreWeave, who kind of shows the opposite to what Michael (05:26):
undefined
Ejaaz: Burry is estimating, which is, there are people booking up GPUs, Josh. (05:31):
undefined
Ejaaz: Two quarters in advance, six months in advance. (05:37):
undefined
Ejaaz: And these aren't the latest GPUs. These are GPUs that are like five years old, (05:40):
undefined
Ejaaz: three years old, in some cases, even longer. (05:46):
undefined
Ejaaz: So if you dig into the details of all these different NeoClouds, (05:49):
undefined
Ejaaz: you'll start realizing that it's the opposite of what Michael Burry is estimating. (05:53):
undefined
Ejaaz: One, these GPUs have a very long lifecycle. And two, they're being used for (05:57):
undefined
Ejaaz: way longer than people expect for really important things within AI. (06:01):
undefined
Ejaaz: They are oversubscribed, they are overutilized. (06:06):
undefined
Josh: This is interesting because you're getting like these really two counterpoints (06:09):
undefined
Josh: here where Michael Burry's subjective take where he's like, no, this is wrong. (06:12):
undefined
Josh: And then you have this objective number, which is CoreWeave. (06:16):
undefined
Josh: And they're saying, wait a second, we're actually selling out these old hard (06:18):
undefined
Josh: drives or these old GPUs. (06:21):
undefined
Josh: Two quarters in advance. So how do we kind of piece together who's right and (06:24):
undefined
Josh: who's wrong through this? (06:29):
undefined
Josh: Because it seems like the market demand, and intuitively it makes sense too, (06:30):
undefined
Josh: that there is no shortage of people who want to generate tokens. (06:33):
undefined
Josh: And even if you are paying a little extra in premium in terms of cost per kilowatt (06:36):
undefined
Josh: for those older GPUs, it's probably still worthwhile because the amount of money (06:40):
undefined
Josh: you could build on top of that is so large. (06:44):
undefined
Ejaaz: Forget about GPUs. Even the CPUs are being used for all the AI distribution (06:46):
undefined
Ejaaz: and coordination. I mean, look at this, AMD CPU tab has just gone up in this (06:50):
undefined
Ejaaz: week's projected earnings, right? (06:55):
undefined
Ejaaz: But to answer your question, it's like, okay, well, who's right here, right? (06:58):
undefined
Ejaaz: I want to kind of, before I address that point, I want to look at another actor (07:02):
undefined
Ejaaz: within this whole circular economy that we like to talk about, (07:07):
undefined
Ejaaz: right? Which is the Frontier AI Labs, right? (07:10):
undefined
Ejaaz: Google, who is an established company, they just had their first $100 billion (07:13):
undefined
Ejaaz: revenue quarter, are using eight-year-old GPUs to like fund their whole thing. (07:17):
undefined
Ejaaz: So again, another data point saying that I don't think these GPUs are old and or not useful. (07:22):
undefined
Ejaaz: And then if you look at NVIDIA themselves, they also have crazy amounts of demand. (07:28):
undefined
Ejaaz: They're oversubbed for years in advance for their Blackwell GPUs and all their (07:33):
undefined
Ejaaz: new GPUs going forwards. (07:37):
undefined
Ejaaz: So then the last actor that I would want to look at is where does all the demand come from? (07:39):
undefined
Ejaaz: Like who's buying these GPUs? Are they satiating demand? (07:45):
undefined
Ejaaz: And if you look at every other company like Google, who has a ton of end users, (07:48):
undefined
Ejaaz: OpenAI with chat GPT users, et cetera. (07:53):
undefined
Ejaaz: Every insider take points to there's not enough GPUs to supply all of this demand (07:55):
undefined
Ejaaz: that they're getting from the apps that they produce, from Sora, (08:01):
undefined
Ejaaz: from a bunch of these different things. (08:04):
undefined
Ejaaz: There's just too much demand. And in many cases, they don't have enough scale (08:05):
undefined
Ejaaz: to even kind of like meet this demand. (08:11):
undefined
Ejaaz: Jensen Huang had a conversation with TSMC, which is like their main (08:13):
undefined
Ejaaz: chip manufacturer asking them to increase the rates of production by 50%. (08:18):
undefined
Ejaaz: So the point I'm saying is Michael Burry, I think is dead wrong here. (08:22):
undefined
Ejaaz: He's definitely underestimated demand and he's underestimated the life cycle (08:26):
undefined
Ejaaz: of these GPUs. They are more valuable than he thought. (08:29):
undefined
Ejaaz: And the old GPUs are worth more than they were worth when they sold originally. (08:33):
undefined
Ejaaz: They're still selling within 5% of their contracts, which is just crazy. (08:38):
undefined
Josh: So maybe I have to ask a pretty dumb question or seemingly (08:41):
undefined
Josh: dumb at least and it's like why why is that so important to the bubble how companies (08:44):
undefined
Josh: factor off losses in gpus over time is this some real big existential threat (08:50):
undefined
Josh: to me it feels like a small part to a bigger picture and not quite as big as (08:55):
undefined
Josh: something that would totally blow out an entire market that's been built so far the (09:00):
undefined
Ejaaz: Biggest budget spend for any major ai company is on GPUs. (09:04):
undefined
Ejaaz: That's where all the hundreds of billions are going for from Microsoft. (09:11):
undefined
Ejaaz: That's where the majority of OpenAI's $1.4 trillion that they've committed over (09:14):
undefined
Ejaaz: the next five years is going to. (09:19):
undefined
Ejaaz: It's all to these compute providers. It's all to these neoclouds. (09:20):
undefined
Ejaaz: It's all to creating their own chips. It's all to Jensen Huang's pocket. (09:23):
undefined
Ejaaz: And so most of the CapEx bubble comes from GPUs. (09:28):
undefined
Ejaaz: So if the bubble is going to burst anywhere, it's going to be from over-leveraging (09:32):
undefined
Ejaaz: on GPUs. Do you want to know something crazy. (09:35):
undefined
Josh: Josh? What's that? (09:38):
undefined
Ejaaz: Even though hundreds of billions have been spent this year alone on GPUs, (09:38):
undefined
Ejaaz: the companies who have spent it haven't even made a dent on their balance sheet. (09:45):
undefined
Ejaaz: Remember, we're talking about Google here. We're talking about Microsoft here. (09:50):
undefined
Ejaaz: They make hundreds of billions of dollars in net profit per year. (09:53):
undefined
Ejaaz: They have plenty of flush cash and they're spending it on machinery that they think is going, (09:57):
undefined
Ejaaz: you and I maybe don't see on the enterprise side or on the end consumer side, (10:03):
undefined
Ejaaz: but they are obviously seeing it. (10:07):
undefined
Josh: Okay. That makes sense. So the depreciation really is a big factor because it's (10:10):
undefined
Josh: kind of artificially inflating numbers across the board. (10:13):
undefined
Josh: It's from the people who are borrowing it, the people who are issuing the GPUs. (10:16):
undefined
Josh: There's lots of high inflation that he's assuming is being baked into this. (10:19):
undefined
Josh: And eventually that inflation kind of fizzles its way out through either some (10:22):
undefined
Josh: really aggressive event or just over time degrading, which makes sense. (10:25):
undefined
Josh: Okay. I think i think i'm up to speed on this i think i get his case (10:30):
undefined
Josh: but what seems a little bit different this time is that (10:34):
undefined
Josh: the first short that was made was in 2008 it (10:36):
undefined
Josh: was against basically the entire stock market and particularly the housing market um (10:39):
undefined
Josh: that was not growing nearly as fast as the ai market is that's like shorting (10:43):
undefined
Josh: the internet and if you shorted the internet before 1999 or after like there's (10:47):
undefined
Josh: this very narrow window to be correct without getting blown out because what (10:53):
undefined
Josh: a lot of people don't realize is in order to short a company you have to borrow (10:56):
undefined
Josh: money against from somewhere else. (10:59):
undefined
Josh: And that borrowing comes with a premium. You have to pay a certain percentage (11:01):
undefined
Josh: interest in order to loan the money. (11:04):
undefined
Josh: If he's not right on this, or even if he's wrong by a couple of months, (11:06):
undefined
Josh: with the rate that the market is accelerating, it feels like it's a very difficult (11:10):
undefined
Josh: thing to get right because... (11:14):
undefined
Josh: You not only have to time it right, but you can't get blown out by the appreciation (11:16):
undefined
Josh: of things as fast as they're growing. (11:21):
undefined
Josh: Like I've never seen a hockey stick of an industry steeper than this one. (11:22):
undefined
Josh: So it seems like this is a really tough position to be bearish in or a tough (11:26):
undefined
Josh: time to bearish in at least. (11:30):
undefined
Ejaaz: Yeah, it's a it's incredibly risky trade to make. (11:32):
undefined
Ejaaz: And I mean, the cherry on top of all of this, Josh, is that he paid the price pretty dearly. (11:36):
undefined
Ejaaz: You sent a message to our group chat this morning and I had to read it like multiple times. (11:43):
undefined
Ejaaz: Michael Burry, this morning of our recording, announced that he is closing down (11:48):
undefined
Ejaaz: his fund, the very same fund which took out this $300 million short on NVIDIA. (11:54):
undefined
Ejaaz: He was down on his position pretty massively. He overstated his means. (11:59):
undefined
Ejaaz: And there's one line here, Josh, which kind of sums it all up. (12:04):
undefined
Ejaaz: He goes, my estimation of value in securities is not now and has not been for (12:07):
undefined
Ejaaz: some time in sync with the markets. (12:13):
undefined
Josh: Tough. Okay, so this connects the dots for me because I saw the beginning and I saw the end. (12:16):
undefined
Josh: I was like, okay, I know he goes out of business, but I do not know why or how (12:20):
undefined
Josh: he gets there. And I think this probably connects the story. (12:24):
undefined
Josh: Don't, again, to the point yesterday, don't bet against the optimist, (12:26):
undefined
Josh: dude. He had absolutely cooked. So where does that leave us then? (12:29):
undefined
Josh: So therefore, it is not a bubble and he was wrong or therefore it is a bubble (12:32):
undefined
Josh: and he was wrong with timing. (12:37):
undefined
Ejaaz: No, I like where you're sniffing josh because that something still seems a bit (12:38):
undefined
Ejaaz: off right it's like well okay if it's not gpus what could it potentially be also so. (12:44):
undefined
Josh: He he blew up his position but that doesn't ruin the hypothesis right like just (12:49):
undefined
Josh: the prices moved against him but it doesn't prove the thesis wrong is that right (12:55):
undefined
Ejaaz: It doesn't prove the thesis wrong on a long-term horizon because no one knows (13:00):
undefined
Ejaaz: whether any of these crazy spends are gonna be crazy in hindsight, (13:05):
undefined
Ejaaz: But for now, in the short term, all the fundamentals show that there is adequate (13:10):
undefined
Ejaaz: demand for the GPUs and there are end users that are willing to use these AI (13:14):
undefined
Ejaaz: products that require these GPUs. (13:18):
undefined
Ejaaz: It actually tells us the opposite story, which is there are not enough GPUs (13:20):
undefined
Ejaaz: in the world right now, old and new, that can satiate the demand for all the (13:24):
undefined
Ejaaz: AI products that are being served today, right now at this moment. (13:28):
undefined
Josh: Okay. Yeah. (13:31):
undefined
Ejaaz: But one thing that could signal a constraint, Josh, is energy. (13:33):
undefined
Ejaaz: The, put simply, the U.S. energy grid isn't up to par to supply energy to all (13:40):
undefined
Ejaaz: these GPUs so that they can do the job. (13:46):
undefined
Ejaaz: In fact, there's this clip that I have here from Satya Nadella of Microsoft (13:48):
undefined
Ejaaz: where he goes on to basically say he has hundreds of millions of dollars worth (13:52):
undefined
Ejaaz: of NVIDIA GPUs that are collecting dust in his fair weather data centers because (13:57):
undefined
Ejaaz: they don't have the energy to supply this. (14:02):
undefined
Josh: That seems uh troubling and i (14:04):
undefined
Josh: think this is to a point that is being made (14:07):
undefined
Josh: a lot where there are these gluts where (14:10):
undefined
Josh: they are and how they show themselves is going to be the thing for debate and (14:13):
undefined
Josh: this isn't the only hot clip we got this week there is this hysterical clip (14:18):
undefined
Josh: um of satya nadella ceo of microsoft kind of talking about his relationship (14:22):
undefined
Josh: with microsoft and i want to play this out before we give some commentary on (14:27):
undefined
Josh: it because i think of all the clips i watched this week. This was one of my favorite. (14:30):
undefined
Ejaaz: In our case, the good news here is OpenAI has a program which we have access to. (14:33):
undefined
Ejaaz: And so therefore, to think that Microsoft is not going to have something that's- What level of. (14:38):
undefined
Josh: Access do you have to that? (14:43):
undefined
Ejaaz: All of them. (14:44):
undefined
Josh: You just get the IP for all of that. So the only IP you don't have is a consumer (14:44):
undefined
Josh: hardware. That's it. Oh, wow. Okay. (14:48):
undefined
Ejaaz: Interesting. That's so good. That's crazy. (14:53):
undefined
Josh: And this comes off the back of Satya saying in a previous interview (14:55):
undefined
Josh: when asked about open ai he goes we are like above them we (14:58):
undefined
Josh: are beside them we are around them they have this full total (15:01):
undefined
Josh: controlling entity over open ai (15:04):
undefined
Josh: and this is it's like the delivery is (15:07):
undefined
Josh: hysterical the reaction is hysterical but it also it (15:10):
undefined
Josh: shows a lot of interesting dynamics between these large companies because during (15:13):
undefined
Josh: this interview satya also made mention of the fact that people who are building (15:18):
undefined
Josh: their moats around individual models are very fragile and very brittle in the (15:23):
undefined
Josh: sense that they are one company copying their model away from being worth significantly less. (15:27):
undefined
Josh: And we kind of saw that earlier this week with our Kimmy K2 episode. (15:34):
undefined
Josh: I highly recommend you watch it if you haven't, because (15:36):
undefined
Josh: It shows how fragile being a frontier model can be if that's your only business. (15:39):
undefined
Josh: And in the case of Satya and Microsoft, I mean, it seems like they have all (15:46):
undefined
Josh: the leverage in the world. (15:50):
undefined
Josh: If they want to fork the OpenAI code base right now and create their own chat (15:51):
undefined
Josh: GPT, they can do that. They own all of it. (15:54):
undefined
Ejaaz: Yeah, I've said this before, but I think the biggest winner of OpenAI is Satya (15:56):
undefined
Ejaaz: Nadella. He has struck the best deal. (16:02):
undefined
Ejaaz: OpenAI recently restructured their entire company. We actually did an episode on that. (16:06):
undefined
Ejaaz: Feel free to check it out. But one of the major takeaways from that was, (16:10):
undefined
Ejaaz: Satya owns 27% of OpenAI. And now he is able to engage in any model provider, not just OpenAI. (16:15):
undefined
Ejaaz: He doesn't have exclusive rights anymore. He can kind of like flirt with other (16:22):
undefined
Ejaaz: AI companies. And he did. (16:26):
undefined
Ejaaz: He engaged with like Amazon on a bunch of different things. (16:27):
undefined
Ejaaz: And he just owns all the IP until something like 2030 or 2032. (16:31):
undefined
Ejaaz: So he is one of the smartest chess players. (16:36):
undefined
Ejaaz: Honestly, unexpectedly for me, I didn't think I thought Microsoft was kind of (16:39):
undefined
Ejaaz: like boomer in this sets, but he's navigated it beautifully. (16:43):
undefined
Ejaaz: And one thing that I also need to give him kudos for is he has one of the biggest (16:47):
undefined
Ejaaz: moats when it comes to enterprise. (16:51):
undefined
Ejaaz: Like remember, Microsoft software is a really good consumer grade product, (16:53):
undefined
Ejaaz: but is mainly kind of making their money from all the enterprise stuff, (16:57):
undefined
Ejaaz: from Copala and stuff like that. (17:00):
undefined
Ejaaz: So beautifully executed from Satya here. (17:02):
undefined
Josh: I love this. Man, Satya, he's crushing it. (17:05):
undefined
Josh: And it's funny because you don't think of Microsoft as a serious player in the world of AI at all. (17:07):
undefined
Josh: I don't use any of their software. I don't use Copilot. In fact, (17:12):
undefined
Josh: he was asked about Copilot because Copilot was the single AI coding agent. (17:15):
undefined
Josh: It owned 100% of the market share. (17:19):
undefined
Josh: And now it's taken down to 25%. And he said, this is a great thing because now (17:21):
undefined
Josh: I still own 25% of a market that is now 10 times the size of what I previously (17:25):
undefined
Josh: owned 100% of. So for him, he's playing the positive sum game. (17:30):
undefined
Josh: He's very aware of the market dynamics at play. And for the cost of what was it, $10 billion? (17:33):
undefined
Josh: He got the largest shareholder position (17:38):
undefined
Josh: of open ai which was rumored to ipo at a trillion dollars (17:41):
undefined
Josh: so they not only see the financial upside but they get all the intellectual (17:45):
undefined
Josh: property most importantly the code base to either reverse engineer for their (17:49):
undefined
Josh: own wants or to just clone and use for whatever they want it for it they have (17:53):
undefined
Josh: it and it is a true chess move by satya bravo nicely done microsoft as (17:58):
undefined
Ejaaz: Of today as we speak today at open ai's 500 billion dollar valuation uh microsoft, (18:03):
undefined
Ejaaz: Microsoft stake is worth $136 billion. (18:10):
undefined
Josh: Not a bad deal. But that's not the only big news. We have exciting news on the ChatGPT front. (18:15):
undefined
Josh: This is what I'm most excited about because I use ChatGPT every day and this (18:22):
undefined
Josh: is hopefully going to change the way I use it. Maybe we'll see when we talk about it. (18:26):
undefined
Josh: The news this week is that ChatGPT and OpenAI, they launched GPT 5.1, (18:31):
undefined
Josh: which is a whole new upgrade to GPT-5. (18:36):
undefined
Josh: How different it is, we're not really sure. (18:39):
undefined
Josh: I know, EJ, as I kind of read through the highlights, it's warmer in terms of (18:42):
undefined
Josh: its like sentiment by default. So it'll be a little bit nicer, (18:46):
undefined
Josh: which is bizarre because I thought we were going, we were trying to go the opposite (18:50):
undefined
Josh: direction, like it's a little too nice. (18:54):
undefined
Josh: And they added some safeguards that we'll get into in a second. (18:56):
undefined
Josh: There's also much better instruction following, which is interesting because (18:58):
undefined
Josh: oftentimes when you instruct GPT-5 to do things, it doesn't always do them exactly as you want. (19:02):
undefined
Josh: So like if you say always respond in six words, it will oftentimes respond in (19:07):
undefined
Josh: more or less than six words. It just doesn't quite understand the instructions. (19:12):
undefined
Josh: And then one of the really fun things that I saw was they changed the different (19:15):
undefined
Josh: types of personalities. (19:20):
undefined
Josh: So you can choose between personalities on how you engage with the model. (19:21):
undefined
Josh: And the previous ones were default, friendly, and efficient. (19:25):
undefined
Josh: And now there are some newly added ones, which are professional, (19:29):
undefined
Josh: candid, quirky, nerdy, and cynical. (19:33):
undefined
Josh: So in the case you want to personalize ChatGPT to be more like those models, (19:36):
undefined
Josh: well, now you have the option to do so with GPT 5.1. (19:41):
undefined
Josh: So if I'm sick of it saying, you're so right, I agree, that's a great idea, (19:45):
undefined
Josh: that's a great question, you could just say, cut that out, and it'll actually (19:50):
undefined
Josh: listen for the first time, which is really exciting. (19:52):
undefined
Josh: EJS, did you find anything interesting from GPT 5.1 after going through everything? (19:55):
undefined
Ejaaz: Yeah, I spent the last like 12 hours playing around with it. (19:58):
undefined
Ejaaz: Can I just, I'm going to play Bad Cop for a bit, Josh. Cool. I don't care. (20:02):
undefined
Ejaaz: Like this is kind of like a nothing burger 0.1 update. (20:06):
undefined
Ejaaz: And it's honestly like I've come to expect more from open air. (20:11):
undefined
Ejaaz: So I'm kind of surprised that they've come through with this release. (20:14):
undefined
Ejaaz: On the point of different personalities, I think that's super helpful. (20:17):
undefined
Ejaaz: Because speaking for myself, I think that the current model is too agreeable. It's too sycophantic. (20:21):
undefined
Ejaaz: And kind of Sam heard this feedback and the original version of GPT-5 was less (20:28):
undefined
Ejaaz: sycophantic, but then a bunch of people were like, I want it to be more friendly. (20:34):
undefined
Ejaaz: I want it to be more appeasing. And so he's kind of like flittering back and forth. (20:37):
undefined
Ejaaz: This option gives a lot more flexibility for the end user, for me. (20:41):
undefined
Ejaaz: I'm like, I want someone that's maybe a bit more candid or a little more cynical (20:44):
undefined
Ejaaz: when I'm kind of doing my research and then maybe a little more friendly when (20:48):
undefined
Ejaaz: I talk about more personal stuff, right? (20:51):
undefined
Ejaaz: So the pliability is cool. (20:52):
undefined
Ejaaz: What I will say is like, you kind of were able to do (20:55):
undefined
Ejaaz: this before you could just go into settings user personalization (20:58):
undefined
Ejaaz: and type in a prompt this i guess makes it easier (21:01):
undefined
Ejaaz: because you just gotta click a button and maybe that helps a lot of people (21:04):
undefined
Ejaaz: um the other cool thing i guess is like uh when it comes to efficiency it uses (21:07):
undefined
Ejaaz: fewer tokens and gives you more output so you get a higher quality answer for (21:12):
undefined
Ejaaz: less energy for less compute which is probably going to drive down costs of (21:17):
undefined
Ejaaz: accessing this type of aid this type of Frontier AI, which is super cool. (21:20):
undefined
Ejaaz: But aside from that, it's kind of like, why did they do this? (21:25):
undefined
Ejaaz: One of the critiques I saw, Josh is like, (21:28):
undefined
Ejaaz: it seems kind of rushed. Like there's no benchmarks, no API release. (21:31):
undefined
Ejaaz: A bunch of the dev tools look a little sloppy. It's almost like they kind of rush this. (21:36):
undefined
Ejaaz: It's not, there's no formal kind of like Sam's doing a live stream around this. (21:40):
undefined
Ejaaz: It's nothing around like, you know, hey, here's how it compares against other (21:44):
undefined
Ejaaz: models. They just kind of like rush this out. And I'm kind of confused. (21:47):
undefined
Josh: Yeah. I suspect the public sentiment will start to reshape itself around how this goes. (21:51):
undefined
Josh: Like when, when Apple releases iOS 26, it (21:55):
undefined
Josh: was a huge release with liquid glass and all this cool new improvements and (21:58):
undefined
Josh: then 26.1 is an incremental improvement where it adds a lot (22:01):
undefined
Josh: of new features nothing really groundbreaking but the product gets light slightly (22:04):
undefined
Josh: better i think we can kind of view these mid-tier updates (22:07):
undefined
Josh: as that where they're just small incremental improvements like (22:10):
undefined
Josh: now it's a little more efficient like you said there's this (22:13):
undefined
Josh: new dynamic thinking time which is what they call like (22:15):
undefined
Josh: smart thinking and it allows it to think faster (22:18):
undefined
Josh: think more efficiently use less words that (22:21):
undefined
Josh: are fillers so it needs less tokens to operate like the chart that we're (22:25):
undefined
Josh: seeing and that just allows for one the (22:27):
undefined
Josh: model to think more because it uses less tokens to think (22:30):
undefined
Josh: and then two it just creates a little more efficiency unlock for open (22:33):
undefined
Josh: ai's gpus to open them up to do other things so i assume this update is probably (22:37):
undefined
Josh: for both people for the consumer and for open ai in terms of efficiency and (22:41):
undefined
Josh: slightly better product is it this amazing novel breakthrough it doesn't appear (22:45):
undefined
Josh: like it at all in fact it's just like a very marginal improvement but it's something it is like (22:49):
undefined
Josh: It is something, and I suspect we'll probably get more updates like this, (22:54):
undefined
Josh: where it's like, oh, well, it does this like one thing a little bit better. (22:57):
undefined
Josh: Do you notice it? Maybe like one in every 50 prompts, but it's not a big deal. (23:00):
undefined
Josh: And that's probably where we are for a little while until maybe Gemini 3. (23:04):
undefined
Josh: I don't know. We have, there are murmurs of Gemini 3 coming down the pipeline from Google. (23:07):
undefined
Ejaaz: I've been waiting for it for a month now. (23:11):
undefined
Josh: Yeah, that could hopefully blow things out of the water. But we're at this weird (23:14):
undefined
Josh: kind of stagnant period where we haven't seen those step function improvements (23:16):
undefined
Josh: in AI models in a little while now. (23:19):
undefined
Ejaaz: I mean, like the speculation that would be like we're kind of grinding to a (23:21):
undefined
Ejaaz: halt and that exponential up, that S curve where we see AI improving massively (23:25):
undefined
Ejaaz: and it's curing cancer may not be quite there yet. (23:30):
undefined
Ejaaz: But hey, listen, in the meantime, there are several parties that are super bullish (23:34):
undefined
Ejaaz: open AI. One of them is Masayoshi-san of SoftBank. (23:38):
undefined
Ejaaz: In this news, he sold his entire stake of NVIDIA worth $5.83 billion to buy OpenAI stock. (23:43):
undefined
Ejaaz: It's hilarious because all that money that is investing in OpenAI is only going (23:52):
undefined
Ejaaz: to get spent on NVIDIA GPUs. (23:55):
undefined
Ejaaz: Just hold NVIDIA. Like, why are you doing that? (23:59):
undefined
Josh: Masa's really, he's been making some questionable decisions lately. (24:02):
undefined
Josh: He sold NVIDIA too early and he missed the whole bubble he missed the whole (24:05):
undefined
Josh: like he's just been making some strange decisions here he (24:08):
undefined
Ejaaz: Wants to own 5% of NVIDIA do you know how much that would be worth today? I cannot believe, (24:12):
undefined
Ejaaz: So like, you know, someone like kind of like tweeted this. They were like, (24:19):
undefined
Ejaaz: you know, someone should look into what happened after he sold his entire NVIDIA stake in 2019. (24:21):
undefined
Ejaaz: Just for reference here, the GPT 3.5 that went viral across the internet released (24:26):
undefined
Ejaaz: two and a half years later. (24:32):
undefined
Ejaaz: So pretty, pretty insane thing there. And then in final news, (24:33):
undefined
Ejaaz: Josh, tell me about this. This looks super exciting. (24:38):
undefined
Josh: The Google DeepMind team is back with some pretty awesome news. (24:40):
undefined
Josh: We are very fond of the Google DeepMind team in particular here at Limitless. (24:44):
undefined
Josh: They are very good at building real world physics in a digital manifestation, (24:47):
undefined
Josh: which I think is a really important thing to be good at in a world where we're (24:52):
undefined
Josh: training robots to be good at physical stuff, but they have not quite had the (24:56):
undefined
Josh: training time to get good at physical stuff. (25:01):
undefined
Josh: And what I love is that we've seen a lot of these projects coming out of the (25:03):
undefined
Josh: DeepMind team. And today we got a (25:06):
undefined
Josh: An agent that is capable of navigating these 3D virtual worlds. (25:10):
undefined
Josh: So we saw previously there was a few models that can generate the worlds. (25:13):
undefined
Josh: Well, now SEMA 2, which implies there was a SEMA 1 that I was blissfully unaware (25:16):
undefined
Josh: of, allows these AI agents to actually navigate these virtual worlds. (25:19):
undefined
Josh: So now not only can Google generate the virtual games, they literally look like (25:24):
undefined
Josh: video games in this in this video, but they can also have the characters navigate (25:28):
undefined
Josh: these complicated experiences, understand how things work, kind of piece things together. (25:32):
undefined
Josh: It's amazing to see because it feels like as i'm watching (25:37):
undefined
Josh: this video it is how a toddler would play (25:40):
undefined
Josh: a video game it's kind of they're slowly moving it's figuring (25:43):
undefined
Josh: it out you could see kind of reasoning in real time and navigating (25:46):
undefined
Josh: these spaces and it's important to understand as you're watching this everything (25:49):
undefined
Josh: that is on this is generated by ai because previously (25:52):
undefined
Josh: these look like trailers to video games and those those video (25:56):
undefined
Josh: games would have required thousands of people who are game developers lots (25:59):
undefined
Josh: of time on the game engine and all of that is (26:03):
undefined
Josh: done now fully by ai so seeing this announcement (26:05):
undefined
Josh: was really exciting for me because i'm like one i just love video games and (26:08):
undefined
Josh: two like oh my god wait these ais not only are they able to generate the environment (26:11):
undefined
Josh: but now these ai models are able to actually engage with the environment and (26:15):
undefined
Josh: what downstream effects does that have on training humanoid robots and the like (26:19):
undefined
Josh: which we saw with the tesla episode last week yeah (26:23):
undefined
Ejaaz: I mean yeah to your point um we see a real life implementation of this with (26:26):
undefined
Ejaaz: tesla tesla Tesla has a world model which they use in their robots, (26:31):
undefined
Ejaaz: which is the Tesla cars and automated self-driving and the future optimist robots. (26:37):
undefined
Ejaaz: And what's cool about this is, Josh, one of the main things that makes your (26:42):
undefined
Ejaaz: AI models super cool or really smart or intelligent is data. (26:46):
undefined
Ejaaz: There's a scarcity of data. There's a scarcity of rich human data. (26:51):
undefined
Ejaaz: Like they don't necessarily see what we see through our eyes or hear what we hear. (26:56):
undefined
Ejaaz: These simulated environments basically synthetically create this data and enhance (26:59):
undefined
Ejaaz: any of these AI models, right? (27:04):
undefined
Ejaaz: And so Tesla kind of gets smarter by replaying a scenario where it almost crashed (27:06):
undefined
Ejaaz: over and over and over again in simulated environments until it gets it right. (27:11):
undefined
Ejaaz: And then that just gets transported to every single production ready Tesla that (27:15):
undefined
Ejaaz: anyone is driving so that if they found themselves in a similar scenario, they avoid it. (27:20):
undefined
Ejaaz: And that's just like one small use case for like what these robots will eventually (27:24):
undefined
Ejaaz: be benefiting from. So such a super cool update. (27:28):
undefined
Josh: Yeah. And if you scroll down actually to post EJs, you'll see a lot of similarities (27:32):
undefined
Josh: to what we spoke about with Tesla, where you could see the model kind of reasoning in plain English. (27:35):
undefined
Josh: You could see, oh, we equipped the pickaxe and copper mine. And then, (27:39):
undefined
Josh: okay, like I'll equip the pickaxe and it walks you through its chain of thought (27:42):
undefined
Josh: and then not only that but the post below it shows that it's self-improving (27:45):
undefined
Josh: so the more time it spends cycling through the more time it spends playing games (27:48):
undefined
Josh: the better it gets and these are early implementations of what what we kind (27:52):
undefined
Josh: of imagined ai would eventually be which is this self-recursive loop where it (27:56):
undefined
Josh: can improve without outside information and it's really exciting to see it happening (27:59):
undefined
Josh: in the virtual world, because I guess that's the only place it can make sense (28:04):
undefined
Josh: where it costs much less. (28:08):
undefined
Josh: You can't harm anybody if a robot goes rogue, but it can learn without the inputs (28:09):
undefined
Josh: of other people. So that is the exciting thing about this, I think. (28:13):
undefined
Josh: And I'm really excited for the DeepMind team, man. I hope they keep going. (28:17):
undefined
Ejaaz: Well, dude, like, do you remember our conversation with Logan Kilpatrick, (28:20):
undefined
Ejaaz: the head of Google's AI studio? (28:24):
undefined
Ejaaz: If you haven't seen that episode, definitely go and check it out. (28:26):
undefined
Ejaaz: One of his main takeaways or one of my main takeaways is all of Google's AI (28:29):
undefined
Ejaaz: byproducts feeds off of different models within the same suite. (28:34):
undefined
Ejaaz: So if their video model learns something cool, they can transpose that onto (28:38):
undefined
Ejaaz: their LLM, which is like in Word, so that it gets recursively smarter over time. (28:43):
undefined
Ejaaz: I just think it's like fascinating and world models kind of like manifest all (28:47):
undefined
Ejaaz: of these different things in one particular simulated environment. (28:52):
undefined
Ejaaz: So, so cool. Kind of hard to wrap my head around sometimes, but because it looks (28:56):
undefined
Ejaaz: like a video game to your point but but awesome. (29:00):
undefined
Josh: Yeah it's you have to break this gap between like your normal understanding (29:01):
undefined
Josh: when you see these videos and think like oh my god wait this is not normal this (29:06):
undefined
Josh: is actually all ai generated um but with that we have concluded all the fun (29:09):
undefined
Josh: new updates from this week it was a (29:12):
undefined
Josh: i'd say a medium week it wasn't anything crazy there was no crazy outliers but (29:14):
undefined
Josh: a solid week overall every week there's so much stuff going on and the progress (29:18):
undefined
Josh: i feel like we're almost getting immune to it how fast things are moving um (29:21):
undefined
Josh: but i hope everyone found this interesting if you did please don't remember (29:25):
undefined
Josh: to share with your friends, like, subscribe. (29:29):
undefined
Josh: Ejaz, I know you were manifesting some sort of send-off today. Do you want to share? (29:31):
undefined
Ejaaz: Yes. Similar to Michael Burry, you shouldn't be shorting the biggest podcast bubble in the world. (29:35):
undefined
Ejaaz: And Limitless is at the forefront over here. We're delivering you all the latest AI news. (29:41):
undefined
Ejaaz: And just like the trade that Burry made that he lost $300 million in, (29:46):
undefined
Ejaaz: you don't want to be shorting the best leading AI podcast in the world. (29:51):
undefined
Ejaaz: Give us a five-star rating. (29:55):
undefined
Ejaaz: Subscribe if you're not. What was the stat, Josh? 80% of listeners? (29:56):
undefined
Ejaaz: 83%? How many of them don't subscribe? Over 80%. (30:00):
undefined
Ejaaz: Over 80. That's nuts. It helps us out massively. (30:04):
undefined
Ejaaz: We are currently in the top 30 in the technology podcast across Spotify, (30:08):
undefined
Ejaaz: Apple, and wherever you listen. Help us get to the top 10. It would mean a lot. What do we say? (30:12):
undefined
Josh: Don't ever bet against the optimists. And by subscribing and by being a part (30:16):
undefined
Josh: of this, you are part of the optimists. (30:20):
undefined
Josh: And we are going to move forward and things are going to progress. (30:21):
undefined
Josh: And just don't bet against the optimists. So thank you for being optimistic (30:25):
undefined
Josh: with us and being on this journey with us and we'll be back next week with a (30:29):
undefined
Josh: whole new slew of episodes so thank you for watching and we'll see you then (30:33):
undefined
undefined
Josh: a catastrophic failure of the housing market that bled into the entire economy, (00:04):
undefined
Josh: and it wiped out a lot of entities. (00:08):
undefined
Josh: And if you'll remember, there was one person in particular who profited a tremendous (00:09):
undefined
Josh: amount off of this happening, and his name is Michael Burry. He made $100 million. (00:13):
undefined
Josh: Personally, collectively, he made a billion dollars off of shorting this market crash. (00:17):
undefined
Josh: He saw what was right early, and he doubled down on it, and he made a killing. (00:23):
undefined
Josh: In fact, so much so that it became a movie that you've probably seen named The Big Short. (00:26):
undefined
Josh: Now, Michael Berry recently has published a new position in recent SEC filings that we uncovered. (00:31):
undefined
Josh: And Ejaz has been all over this over the last couple of weeks, (00:36):
undefined
Josh: tracking the positions, trying to understand why he's making these things. (00:39):
undefined
Josh: And if this bet is correct in predicting the next big bubble to pop, (00:42):
undefined
Josh: which is seeming to be AI. (00:46):
undefined
Josh: So he predicted the housing bubble in 08. He made a billion dollars. (00:48):
undefined
Josh: He's predicting another bubble in 2025. five, is he going to make another billion dollars, EJS? (00:52):
undefined
Ejaaz: The short answer is, I don't think he will. I think he's going to get blown out. (00:58):
undefined
Ejaaz: So Michael Burry is back. He is short around $300 million worth of Nvidia shares. (01:03):
undefined
Ejaaz: So he's shorting, at this point, the richest, wealthiest company in the world. (01:09):
undefined
Ejaaz: And that is a big sign, basically saying that I think the air bubble is popping. (01:13):
undefined
Ejaaz: So the question is, is it? (01:17):
undefined
Ejaaz: Let's go through his thesis, Josh. I'm going to simplify his post here. (01:19):
undefined
Ejaaz: Which is basically GPUs run the AI world, right? (01:24):
undefined
Ejaaz: Trillions of dollars have been spent by Frontier AI labs collectively to train state-of-the-art AI. (01:28):
undefined
Ejaaz: But the thing with these GPUs, Josh, is that they have a lifespan, (01:34):
undefined
Ejaaz: right? They don't last forever, right? (01:39):
undefined
Ejaaz: And he estimates that the top Frontier labs. (01:41):
undefined
Ejaaz: Overestimating or artificially boosting the lifespan of the GPUs that they purchase. (01:46):
undefined
Ejaaz: And you might be like, well, that sounds boring. Why is that interesting? (01:50):
undefined
Ejaaz: Well, the thing is, if you have assets on your balance sheet, (01:53):
undefined
Ejaaz: that factors into your stock price. (01:57):
undefined
Ejaaz: So the point that he's making is all these big companies are artificially boosting (01:59):
undefined
Ejaaz: the lifespan of the GPUs that they've purchased so that it inflates their stock price. (02:03):
undefined
Ejaaz: And therefore, the real stock price is actually a lot lower. (02:08):
undefined
Ejaaz: In fact, he says at the bottom of this tweet, by 2028, Oracle will overstate (02:12):
undefined
Ejaaz: earnings by 26.9%, Meta by 20%, etc. (02:16):
undefined
Ejaaz: And so his big short, Josh, is on the behemoth that is supplying all these Frontier (02:20):
undefined
Ejaaz: AI labs, NVIDIA, saying that when the bubble pops, and he thinks the bubble's (02:25):
undefined
Ejaaz: going to pop now, he will make a heck of a lot of money on this trade. (02:29):
undefined
Josh: So is he saying the bubble's going to pop now, or is it going to pop in 2028? (02:34):
undefined
Josh: Because he's referencing those numbers in 2028. (02:37):
undefined
Josh: So I'm kind of curious how he's framing this is this like because i'm now i'm (02:39):
undefined
Josh: thinking about our positions um is this something that he predicts is going (02:44):
undefined
Josh: to happen in 2028 or is this more of a short timescale position if (02:47):
undefined
Ejaaz: We remember from his original short i think he held a position for about a year, (02:51):
undefined
Ejaaz: Or at least like the better part of a year. So I think he probably estimates (02:56):
undefined
Ejaaz: that it's going to happen sometime within that time frame. (03:00):
undefined
Ejaaz: But I have more evidence as to like why he might be thinking that the bubble is bursting. (03:03):
undefined
Ejaaz: Have you heard of what a neocloud is, Josh? (03:08):
undefined
Josh: The name is familiar, but I don't understand quite what that means. So please fill me in. (03:12):
undefined
Ejaaz: Okay, so think of like AWS or any of these other typical cloud providers. (03:17):
undefined
Ejaaz: But they're specifically focused on providing GPUs and data center pipelines for you. (03:21):
undefined
Ejaaz: So think of like an AWS just for AI compute. That makes sense? (03:27):
undefined
Ejaaz: And so what he did was he looked at the major NeoCloud providers and he looked (03:31):
undefined
Ejaaz: at their stock prices over the last week, Josh, and they were down. (03:38):
undefined
Ejaaz: So he upped his position and was like, okay, this is the bubble bursting. I'm gonna lay my claim. (03:43):
undefined
Ejaaz: There's an issue with this thesis, Josh. I think he's dead wrong. (03:49):
undefined
Ejaaz: And I'm about to walk you through four reasons why he's wrong. (03:53):
undefined
Josh: Okay. So then just to kind of like wrap my head around this, (03:57):
undefined
Josh: the NeoClouds are kind of infrastructure for people who don't want to build (03:59):
undefined
Josh: their own infrastructure. (04:02):
undefined
Josh: Kind of like you described AWS. If you want to get a data center, (04:03):
undefined
Josh: but you don't want to build a data center, you offload that responsibility to a NeoCloud. (04:06):
undefined
Josh: And he thinks that the cost that these companies are marking down these NeoCloud (04:10):
undefined
Josh: prices at, particularly the GPUs is much higher because the depreciation happens (04:14):
undefined
Josh: on a shorter timescale than the companies are writing off. (04:19):
undefined
Ejaaz: Correct. And NeoClouds is one part of the picture, but it's a great example (04:22):
undefined
Ejaaz: to lead with because so many frontier AI labs like Microsoft OpenAI actually (04:28):
undefined
Ejaaz: pay these NeoClouds billions of dollars. (04:33):
undefined
Ejaaz: Like Microsoft just signed a $19 billion contract with a NeoCloud provider called (04:36):
undefined
Ejaaz: Nebius, which kind of sent its stock price going up. (04:40):
undefined
Ejaaz: We talked about it on a previous episode. so like it's a (04:43):
undefined
Ejaaz: good example to kind of like lead with but the (04:46):
undefined
Ejaaz: take here is that these neoclouds are (04:49):
undefined
Ejaaz: basically as you said overestimating the life cycle of these gpus and therefore (04:52):
undefined
Ejaaz: they are wrong except that's not the case at all and i present to you my first (04:57):
undefined
Ejaaz: counter thesis which is he's wrong about the two to three year depreciation (05:02):
undefined
Ejaaz: cycle it's actually more like five to six in some cases even eight years. (05:06):
undefined
Ejaaz: These GPUs are used for more than just training AI models. (05:11):
undefined
Ejaaz: They are used for inference, and they're sometimes used just for general queries (05:15):
undefined
Ejaaz: and distribution for these AI models, right? (05:19):
undefined
Ejaaz: But the most staggering kind of example comes from the top doc, (05:22):
undefined
Ejaaz: the top neocloud called CoreWeave, who kind of shows the opposite to what Michael (05:26):
undefined
Ejaaz: Burry is estimating, which is, there are people booking up GPUs, Josh. (05:31):
undefined
Ejaaz: Two quarters in advance, six months in advance. (05:37):
undefined
Ejaaz: And these aren't the latest GPUs. These are GPUs that are like five years old, (05:40):
undefined
Ejaaz: three years old, in some cases, even longer. (05:46):
undefined
Ejaaz: So if you dig into the details of all these different NeoClouds, (05:49):
undefined
Ejaaz: you'll start realizing that it's the opposite of what Michael Burry is estimating. (05:53):
undefined
Ejaaz: One, these GPUs have a very long lifecycle. And two, they're being used for (05:57):
undefined
Ejaaz: way longer than people expect for really important things within AI. (06:01):
undefined
Ejaaz: They are oversubscribed, they are overutilized. (06:06):
undefined
Josh: This is interesting because you're getting like these really two counterpoints (06:09):
undefined
Josh: here where Michael Burry's subjective take where he's like, no, this is wrong. (06:12):
undefined
Josh: And then you have this objective number, which is CoreWeave. (06:16):
undefined
Josh: And they're saying, wait a second, we're actually selling out these old hard (06:18):
undefined
Josh: drives or these old GPUs. (06:21):
undefined
Josh: Two quarters in advance. So how do we kind of piece together who's right and (06:24):
undefined
Josh: who's wrong through this? (06:29):
undefined
Josh: Because it seems like the market demand, and intuitively it makes sense too, (06:30):
undefined
Josh: that there is no shortage of people who want to generate tokens. (06:33):
undefined
Josh: And even if you are paying a little extra in premium in terms of cost per kilowatt (06:36):
undefined
Josh: for those older GPUs, it's probably still worthwhile because the amount of money (06:40):
undefined
Josh: you could build on top of that is so large. (06:44):
undefined
Ejaaz: Forget about GPUs. Even the CPUs are being used for all the AI distribution (06:46):
undefined
Ejaaz: and coordination. I mean, look at this, AMD CPU tab has just gone up in this (06:50):
undefined
Ejaaz: week's projected earnings, right? (06:55):
undefined
Ejaaz: But to answer your question, it's like, okay, well, who's right here, right? (06:58):
undefined
Ejaaz: I want to kind of, before I address that point, I want to look at another actor (07:02):
undefined
Ejaaz: within this whole circular economy that we like to talk about, (07:07):
undefined
Ejaaz: right? Which is the Frontier AI Labs, right? (07:10):
undefined
Ejaaz: Google, who is an established company, they just had their first $100 billion (07:13):
undefined
Ejaaz: revenue quarter, are using eight-year-old GPUs to like fund their whole thing. (07:17):
undefined
Ejaaz: So again, another data point saying that I don't think these GPUs are old and or not useful. (07:22):
undefined
Ejaaz: And then if you look at NVIDIA themselves, they also have crazy amounts of demand. (07:28):
undefined
Ejaaz: They're oversubbed for years in advance for their Blackwell GPUs and all their (07:33):
undefined
Ejaaz: new GPUs going forwards. (07:37):
undefined
Ejaaz: So then the last actor that I would want to look at is where does all the demand come from? (07:39):
undefined
Ejaaz: Like who's buying these GPUs? Are they satiating demand? (07:45):
undefined
Ejaaz: And if you look at every other company like Google, who has a ton of end users, (07:48):
undefined
Ejaaz: OpenAI with chat GPT users, et cetera. (07:53):
undefined
Ejaaz: Every insider take points to there's not enough GPUs to supply all of this demand (07:55):
undefined
Ejaaz: that they're getting from the apps that they produce, from Sora, (08:01):
undefined
Ejaaz: from a bunch of these different things. (08:04):
undefined
Ejaaz: There's just too much demand. And in many cases, they don't have enough scale (08:05):
undefined
Ejaaz: to even kind of like meet this demand. (08:11):
undefined
Ejaaz: Jensen Huang had a conversation with TSMC, which is like their main (08:13):
undefined
Ejaaz: chip manufacturer asking them to increase the rates of production by 50%. (08:18):
undefined
Ejaaz: So the point I'm saying is Michael Burry, I think is dead wrong here. (08:22):
undefined
Ejaaz: He's definitely underestimated demand and he's underestimated the life cycle (08:26):
undefined
Ejaaz: of these GPUs. They are more valuable than he thought. (08:29):
undefined
Ejaaz: And the old GPUs are worth more than they were worth when they sold originally. (08:33):
undefined
Ejaaz: They're still selling within 5% of their contracts, which is just crazy. (08:38):
undefined
Josh: So maybe I have to ask a pretty dumb question or seemingly (08:41):
undefined
Josh: dumb at least and it's like why why is that so important to the bubble how companies (08:44):
undefined
Josh: factor off losses in gpus over time is this some real big existential threat (08:50):
undefined
Josh: to me it feels like a small part to a bigger picture and not quite as big as (08:55):
undefined
Josh: something that would totally blow out an entire market that's been built so far the (09:00):
undefined
Ejaaz: Biggest budget spend for any major ai company is on GPUs. (09:04):
undefined
Ejaaz: That's where all the hundreds of billions are going for from Microsoft. (09:11):
undefined
Ejaaz: That's where the majority of OpenAI's $1.4 trillion that they've committed over (09:14):
undefined
Ejaaz: the next five years is going to. (09:19):
undefined
Ejaaz: It's all to these compute providers. It's all to these neoclouds. (09:20):
undefined
Ejaaz: It's all to creating their own chips. It's all to Jensen Huang's pocket. (09:23):
undefined
Ejaaz: And so most of the CapEx bubble comes from GPUs. (09:28):
undefined
Ejaaz: So if the bubble is going to burst anywhere, it's going to be from over-leveraging (09:32):
undefined
Ejaaz: on GPUs. Do you want to know something crazy. (09:35):
undefined
Josh: Josh? What's that? (09:38):
undefined
Ejaaz: Even though hundreds of billions have been spent this year alone on GPUs, (09:38):
undefined
Ejaaz: the companies who have spent it haven't even made a dent on their balance sheet. (09:45):
undefined
Ejaaz: Remember, we're talking about Google here. We're talking about Microsoft here. (09:50):
undefined
Ejaaz: They make hundreds of billions of dollars in net profit per year. (09:53):
undefined
Ejaaz: They have plenty of flush cash and they're spending it on machinery that they think is going, (09:57):
undefined
Ejaaz: you and I maybe don't see on the enterprise side or on the end consumer side, (10:03):
undefined
Ejaaz: but they are obviously seeing it. (10:07):
undefined
Josh: Okay. That makes sense. So the depreciation really is a big factor because it's (10:10):
undefined
Josh: kind of artificially inflating numbers across the board. (10:13):
undefined
Josh: It's from the people who are borrowing it, the people who are issuing the GPUs. (10:16):
undefined
Josh: There's lots of high inflation that he's assuming is being baked into this. (10:19):
undefined
Josh: And eventually that inflation kind of fizzles its way out through either some (10:22):
undefined
Josh: really aggressive event or just over time degrading, which makes sense. (10:25):
undefined
Josh: Okay. I think i think i'm up to speed on this i think i get his case (10:30):
undefined
Josh: but what seems a little bit different this time is that (10:34):
undefined
Josh: the first short that was made was in 2008 it (10:36):
undefined
Josh: was against basically the entire stock market and particularly the housing market um (10:39):
undefined
Josh: that was not growing nearly as fast as the ai market is that's like shorting (10:43):
undefined
Josh: the internet and if you shorted the internet before 1999 or after like there's (10:47):
undefined
Josh: this very narrow window to be correct without getting blown out because what (10:53):
undefined
Josh: a lot of people don't realize is in order to short a company you have to borrow (10:56):
undefined
Josh: money against from somewhere else. (10:59):
undefined
Josh: And that borrowing comes with a premium. You have to pay a certain percentage (11:01):
undefined
Josh: interest in order to loan the money. (11:04):
undefined
Josh: If he's not right on this, or even if he's wrong by a couple of months, (11:06):
undefined
Josh: with the rate that the market is accelerating, it feels like it's a very difficult (11:10):
undefined
Josh: thing to get right because... (11:14):
undefined
Josh: You not only have to time it right, but you can't get blown out by the appreciation (11:16):
undefined
Josh: of things as fast as they're growing. (11:21):
undefined
Josh: Like I've never seen a hockey stick of an industry steeper than this one. (11:22):
undefined
Josh: So it seems like this is a really tough position to be bearish in or a tough (11:26):
undefined
Josh: time to bearish in at least. (11:30):
undefined
Ejaaz: Yeah, it's a it's incredibly risky trade to make. (11:32):
undefined
Ejaaz: And I mean, the cherry on top of all of this, Josh, is that he paid the price pretty dearly. (11:36):
undefined
Ejaaz: You sent a message to our group chat this morning and I had to read it like multiple times. (11:43):
undefined
Ejaaz: Michael Burry, this morning of our recording, announced that he is closing down (11:48):
undefined
Ejaaz: his fund, the very same fund which took out this $300 million short on NVIDIA. (11:54):
undefined
Ejaaz: He was down on his position pretty massively. He overstated his means. (11:59):
undefined
Ejaaz: And there's one line here, Josh, which kind of sums it all up. (12:04):
undefined
Ejaaz: He goes, my estimation of value in securities is not now and has not been for (12:07):
undefined
Ejaaz: some time in sync with the markets. (12:13):
undefined
Josh: Tough. Okay, so this connects the dots for me because I saw the beginning and I saw the end. (12:16):
undefined
Josh: I was like, okay, I know he goes out of business, but I do not know why or how (12:20):
undefined
Josh: he gets there. And I think this probably connects the story. (12:24):
undefined
Josh: Don't, again, to the point yesterday, don't bet against the optimist, (12:26):
undefined
Josh: dude. He had absolutely cooked. So where does that leave us then? (12:29):
undefined
Josh: So therefore, it is not a bubble and he was wrong or therefore it is a bubble (12:32):
undefined
Josh: and he was wrong with timing. (12:37):
undefined
Ejaaz: No, I like where you're sniffing josh because that something still seems a bit (12:38):
undefined
Ejaaz: off right it's like well okay if it's not gpus what could it potentially be also so. (12:44):
undefined
Josh: He he blew up his position but that doesn't ruin the hypothesis right like just (12:49):
undefined
Josh: the prices moved against him but it doesn't prove the thesis wrong is that right (12:55):
undefined
Ejaaz: It doesn't prove the thesis wrong on a long-term horizon because no one knows (13:00):
undefined
Ejaaz: whether any of these crazy spends are gonna be crazy in hindsight, (13:05):
undefined
Ejaaz: But for now, in the short term, all the fundamentals show that there is adequate (13:10):
undefined
Ejaaz: demand for the GPUs and there are end users that are willing to use these AI (13:14):
undefined
Ejaaz: products that require these GPUs. (13:18):
undefined
Ejaaz: It actually tells us the opposite story, which is there are not enough GPUs (13:20):
undefined
Ejaaz: in the world right now, old and new, that can satiate the demand for all the (13:24):
undefined
Ejaaz: AI products that are being served today, right now at this moment. (13:28):
undefined
Josh: Okay. Yeah. (13:31):
undefined
Ejaaz: But one thing that could signal a constraint, Josh, is energy. (13:33):
undefined
Ejaaz: The, put simply, the U.S. energy grid isn't up to par to supply energy to all (13:40):
undefined
Ejaaz: these GPUs so that they can do the job. (13:46):
undefined
Ejaaz: In fact, there's this clip that I have here from Satya Nadella of Microsoft (13:48):
undefined
Ejaaz: where he goes on to basically say he has hundreds of millions of dollars worth (13:52):
undefined
Ejaaz: of NVIDIA GPUs that are collecting dust in his fair weather data centers because (13:57):
undefined
Ejaaz: they don't have the energy to supply this. (14:02):
undefined
Josh: That seems uh troubling and i (14:04):
undefined
Josh: think this is to a point that is being made (14:07):
undefined
Josh: a lot where there are these gluts where (14:10):
undefined
Josh: they are and how they show themselves is going to be the thing for debate and (14:13):
undefined
Josh: this isn't the only hot clip we got this week there is this hysterical clip (14:18):
undefined
Josh: um of satya nadella ceo of microsoft kind of talking about his relationship (14:22):
undefined
Josh: with microsoft and i want to play this out before we give some commentary on (14:27):
undefined
Josh: it because i think of all the clips i watched this week. This was one of my favorite. (14:30):
undefined
Ejaaz: In our case, the good news here is OpenAI has a program which we have access to. (14:33):
undefined
Ejaaz: And so therefore, to think that Microsoft is not going to have something that's- What level of. (14:38):
undefined
Josh: Access do you have to that? (14:43):
undefined
Ejaaz: All of them. (14:44):
undefined
Josh: You just get the IP for all of that. So the only IP you don't have is a consumer (14:44):
undefined
Josh: hardware. That's it. Oh, wow. Okay. (14:48):
undefined
Ejaaz: Interesting. That's so good. That's crazy. (14:53):
undefined
Josh: And this comes off the back of Satya saying in a previous interview (14:55):
undefined
Josh: when asked about open ai he goes we are like above them we (14:58):
undefined
Josh: are beside them we are around them they have this full total (15:01):
undefined
Josh: controlling entity over open ai (15:04):
undefined
Josh: and this is it's like the delivery is (15:07):
undefined
Josh: hysterical the reaction is hysterical but it also it (15:10):
undefined
Josh: shows a lot of interesting dynamics between these large companies because during (15:13):
undefined
Josh: this interview satya also made mention of the fact that people who are building (15:18):
undefined
Josh: their moats around individual models are very fragile and very brittle in the (15:23):
undefined
Josh: sense that they are one company copying their model away from being worth significantly less. (15:27):
undefined
Josh: And we kind of saw that earlier this week with our Kimmy K2 episode. (15:34):
undefined
Josh: I highly recommend you watch it if you haven't, because (15:36):
undefined
Josh: It shows how fragile being a frontier model can be if that's your only business. (15:39):
undefined
Josh: And in the case of Satya and Microsoft, I mean, it seems like they have all (15:46):
undefined
Josh: the leverage in the world. (15:50):
undefined
Josh: If they want to fork the OpenAI code base right now and create their own chat (15:51):
undefined
Josh: GPT, they can do that. They own all of it. (15:54):
undefined
Ejaaz: Yeah, I've said this before, but I think the biggest winner of OpenAI is Satya (15:56):
undefined
Ejaaz: Nadella. He has struck the best deal. (16:02):
undefined
Ejaaz: OpenAI recently restructured their entire company. We actually did an episode on that. (16:06):
undefined
Ejaaz: Feel free to check it out. But one of the major takeaways from that was, (16:10):
undefined
Ejaaz: Satya owns 27% of OpenAI. And now he is able to engage in any model provider, not just OpenAI. (16:15):
undefined
Ejaaz: He doesn't have exclusive rights anymore. He can kind of like flirt with other (16:22):
undefined
Ejaaz: AI companies. And he did. (16:26):
undefined
Ejaaz: He engaged with like Amazon on a bunch of different things. (16:27):
undefined
Ejaaz: And he just owns all the IP until something like 2030 or 2032. (16:31):
undefined
Ejaaz: So he is one of the smartest chess players. (16:36):
undefined
Ejaaz: Honestly, unexpectedly for me, I didn't think I thought Microsoft was kind of (16:39):
undefined
Ejaaz: like boomer in this sets, but he's navigated it beautifully. (16:43):
undefined
Ejaaz: And one thing that I also need to give him kudos for is he has one of the biggest (16:47):
undefined
Ejaaz: moats when it comes to enterprise. (16:51):
undefined
Ejaaz: Like remember, Microsoft software is a really good consumer grade product, (16:53):
undefined
Ejaaz: but is mainly kind of making their money from all the enterprise stuff, (16:57):
undefined
Ejaaz: from Copala and stuff like that. (17:00):
undefined
Ejaaz: So beautifully executed from Satya here. (17:02):
undefined
Josh: I love this. Man, Satya, he's crushing it. (17:05):
undefined
Josh: And it's funny because you don't think of Microsoft as a serious player in the world of AI at all. (17:07):
undefined
Josh: I don't use any of their software. I don't use Copilot. In fact, (17:12):
undefined
Josh: he was asked about Copilot because Copilot was the single AI coding agent. (17:15):
undefined
Josh: It owned 100% of the market share. (17:19):
undefined
Josh: And now it's taken down to 25%. And he said, this is a great thing because now (17:21):
undefined
Josh: I still own 25% of a market that is now 10 times the size of what I previously (17:25):
undefined
Josh: owned 100% of. So for him, he's playing the positive sum game. (17:30):
undefined
Josh: He's very aware of the market dynamics at play. And for the cost of what was it, $10 billion? (17:33):
undefined
Josh: He got the largest shareholder position (17:38):
undefined
Josh: of open ai which was rumored to ipo at a trillion dollars (17:41):
undefined
Josh: so they not only see the financial upside but they get all the intellectual (17:45):
undefined
Josh: property most importantly the code base to either reverse engineer for their (17:49):
undefined
Josh: own wants or to just clone and use for whatever they want it for it they have (17:53):
undefined
Josh: it and it is a true chess move by satya bravo nicely done microsoft as (17:58):
undefined
Ejaaz: Of today as we speak today at open ai's 500 billion dollar valuation uh microsoft, (18:03):
undefined
Ejaaz: Microsoft stake is worth $136 billion. (18:10):
undefined
Josh: Not a bad deal. But that's not the only big news. We have exciting news on the ChatGPT front. (18:15):
undefined
Josh: This is what I'm most excited about because I use ChatGPT every day and this (18:22):
undefined
Josh: is hopefully going to change the way I use it. Maybe we'll see when we talk about it. (18:26):
undefined
Josh: The news this week is that ChatGPT and OpenAI, they launched GPT 5.1, (18:31):
undefined
Josh: which is a whole new upgrade to GPT-5. (18:36):
undefined
Josh: How different it is, we're not really sure. (18:39):
undefined
Josh: I know, EJ, as I kind of read through the highlights, it's warmer in terms of (18:42):
undefined
Josh: its like sentiment by default. So it'll be a little bit nicer, (18:46):
undefined
Josh: which is bizarre because I thought we were going, we were trying to go the opposite (18:50):
undefined
Josh: direction, like it's a little too nice. (18:54):
undefined
Josh: And they added some safeguards that we'll get into in a second. (18:56):
undefined
Josh: There's also much better instruction following, which is interesting because (18:58):
undefined
Josh: oftentimes when you instruct GPT-5 to do things, it doesn't always do them exactly as you want. (19:02):
undefined
Josh: So like if you say always respond in six words, it will oftentimes respond in (19:07):
undefined
Josh: more or less than six words. It just doesn't quite understand the instructions. (19:12):
undefined
Josh: And then one of the really fun things that I saw was they changed the different (19:15):
undefined
Josh: types of personalities. (19:20):
undefined
Josh: So you can choose between personalities on how you engage with the model. (19:21):
undefined
Josh: And the previous ones were default, friendly, and efficient. (19:25):
undefined
Josh: And now there are some newly added ones, which are professional, (19:29):
undefined
Josh: candid, quirky, nerdy, and cynical. (19:33):
undefined
Josh: So in the case you want to personalize ChatGPT to be more like those models, (19:36):
undefined
Josh: well, now you have the option to do so with GPT 5.1. (19:41):
undefined
Josh: So if I'm sick of it saying, you're so right, I agree, that's a great idea, (19:45):
undefined
Josh: that's a great question, you could just say, cut that out, and it'll actually (19:50):
undefined
Josh: listen for the first time, which is really exciting. (19:52):
undefined
Josh: EJS, did you find anything interesting from GPT 5.1 after going through everything? (19:55):
undefined
Ejaaz: Yeah, I spent the last like 12 hours playing around with it. (19:58):
undefined
Ejaaz: Can I just, I'm going to play Bad Cop for a bit, Josh. Cool. I don't care. (20:02):
undefined
Ejaaz: Like this is kind of like a nothing burger 0.1 update. (20:06):
undefined
Ejaaz: And it's honestly like I've come to expect more from open air. (20:11):
undefined
Ejaaz: So I'm kind of surprised that they've come through with this release. (20:14):
undefined
Ejaaz: On the point of different personalities, I think that's super helpful. (20:17):
undefined
Ejaaz: Because speaking for myself, I think that the current model is too agreeable. It's too sycophantic. (20:21):
undefined
Ejaaz: And kind of Sam heard this feedback and the original version of GPT-5 was less (20:28):
undefined
Ejaaz: sycophantic, but then a bunch of people were like, I want it to be more friendly. (20:34):
undefined
Ejaaz: I want it to be more appeasing. And so he's kind of like flittering back and forth. (20:37):
undefined
Ejaaz: This option gives a lot more flexibility for the end user, for me. (20:41):
undefined
Ejaaz: I'm like, I want someone that's maybe a bit more candid or a little more cynical (20:44):
undefined
Ejaaz: when I'm kind of doing my research and then maybe a little more friendly when (20:48):
undefined
Ejaaz: I talk about more personal stuff, right? (20:51):
undefined
Ejaaz: So the pliability is cool. (20:52):
undefined
Ejaaz: What I will say is like, you kind of were able to do (20:55):
undefined
Ejaaz: this before you could just go into settings user personalization (20:58):
undefined
Ejaaz: and type in a prompt this i guess makes it easier (21:01):
undefined
Ejaaz: because you just gotta click a button and maybe that helps a lot of people (21:04):
undefined
Ejaaz: um the other cool thing i guess is like uh when it comes to efficiency it uses (21:07):
undefined
Ejaaz: fewer tokens and gives you more output so you get a higher quality answer for (21:12):
undefined
Ejaaz: less energy for less compute which is probably going to drive down costs of (21:17):
undefined
Ejaaz: accessing this type of aid this type of Frontier AI, which is super cool. (21:20):
undefined
Ejaaz: But aside from that, it's kind of like, why did they do this? (21:25):
undefined
Ejaaz: One of the critiques I saw, Josh is like, (21:28):
undefined
Ejaaz: it seems kind of rushed. Like there's no benchmarks, no API release. (21:31):
undefined
Ejaaz: A bunch of the dev tools look a little sloppy. It's almost like they kind of rush this. (21:36):
undefined
Ejaaz: It's not, there's no formal kind of like Sam's doing a live stream around this. (21:40):
undefined
Ejaaz: It's nothing around like, you know, hey, here's how it compares against other (21:44):
undefined
Ejaaz: models. They just kind of like rush this out. And I'm kind of confused. (21:47):
undefined
Josh: Yeah. I suspect the public sentiment will start to reshape itself around how this goes. (21:51):
undefined
Josh: Like when, when Apple releases iOS 26, it (21:55):
undefined
Josh: was a huge release with liquid glass and all this cool new improvements and (21:58):
undefined
Josh: then 26.1 is an incremental improvement where it adds a lot (22:01):
undefined
Josh: of new features nothing really groundbreaking but the product gets light slightly (22:04):
undefined
Josh: better i think we can kind of view these mid-tier updates (22:07):
undefined
Josh: as that where they're just small incremental improvements like (22:10):
undefined
Josh: now it's a little more efficient like you said there's this (22:13):
undefined
Josh: new dynamic thinking time which is what they call like (22:15):
undefined
Josh: smart thinking and it allows it to think faster (22:18):
undefined
Josh: think more efficiently use less words that (22:21):
undefined
Josh: are fillers so it needs less tokens to operate like the chart that we're (22:25):
undefined
Josh: seeing and that just allows for one the (22:27):
undefined
Josh: model to think more because it uses less tokens to think (22:30):
undefined
Josh: and then two it just creates a little more efficiency unlock for open (22:33):
undefined
Josh: ai's gpus to open them up to do other things so i assume this update is probably (22:37):
undefined
Josh: for both people for the consumer and for open ai in terms of efficiency and (22:41):
undefined
Josh: slightly better product is it this amazing novel breakthrough it doesn't appear (22:45):
undefined
Josh: like it at all in fact it's just like a very marginal improvement but it's something it is like (22:49):
undefined
Josh: It is something, and I suspect we'll probably get more updates like this, (22:54):
undefined
Josh: where it's like, oh, well, it does this like one thing a little bit better. (22:57):
undefined
Josh: Do you notice it? Maybe like one in every 50 prompts, but it's not a big deal. (23:00):
undefined
Josh: And that's probably where we are for a little while until maybe Gemini 3. (23:04):
undefined
Josh: I don't know. We have, there are murmurs of Gemini 3 coming down the pipeline from Google. (23:07):
undefined
Ejaaz: I've been waiting for it for a month now. (23:11):
undefined
Josh: Yeah, that could hopefully blow things out of the water. But we're at this weird (23:14):
undefined
Josh: kind of stagnant period where we haven't seen those step function improvements (23:16):
undefined
Josh: in AI models in a little while now. (23:19):
undefined
Ejaaz: I mean, like the speculation that would be like we're kind of grinding to a (23:21):
undefined
Ejaaz: halt and that exponential up, that S curve where we see AI improving massively (23:25):
undefined
Ejaaz: and it's curing cancer may not be quite there yet. (23:30):
undefined
Ejaaz: But hey, listen, in the meantime, there are several parties that are super bullish (23:34):
undefined
Ejaaz: open AI. One of them is Masayoshi-san of SoftBank. (23:38):
undefined
Ejaaz: In this news, he sold his entire stake of NVIDIA worth $5.83 billion to buy OpenAI stock. (23:43):
undefined
Ejaaz: It's hilarious because all that money that is investing in OpenAI is only going (23:52):
undefined
Ejaaz: to get spent on NVIDIA GPUs. (23:55):
undefined
Ejaaz: Just hold NVIDIA. Like, why are you doing that? (23:59):
undefined
Josh: Masa's really, he's been making some questionable decisions lately. (24:02):
undefined
Josh: He sold NVIDIA too early and he missed the whole bubble he missed the whole (24:05):
undefined
Josh: like he's just been making some strange decisions here he (24:08):
undefined
Ejaaz: Wants to own 5% of NVIDIA do you know how much that would be worth today? I cannot believe, (24:12):
undefined
Ejaaz: So like, you know, someone like kind of like tweeted this. They were like, (24:19):
undefined
Ejaaz: you know, someone should look into what happened after he sold his entire NVIDIA stake in 2019. (24:21):
undefined
Ejaaz: Just for reference here, the GPT 3.5 that went viral across the internet released (24:26):
undefined
Ejaaz: two and a half years later. (24:32):
undefined
Ejaaz: So pretty, pretty insane thing there. And then in final news, (24:33):
undefined
Ejaaz: Josh, tell me about this. This looks super exciting. (24:38):
undefined
Josh: The Google DeepMind team is back with some pretty awesome news. (24:40):
undefined
Josh: We are very fond of the Google DeepMind team in particular here at Limitless. (24:44):
undefined
Josh: They are very good at building real world physics in a digital manifestation, (24:47):
undefined
Josh: which I think is a really important thing to be good at in a world where we're (24:52):
undefined
Josh: training robots to be good at physical stuff, but they have not quite had the (24:56):
undefined
Josh: training time to get good at physical stuff. (25:01):
undefined
Josh: And what I love is that we've seen a lot of these projects coming out of the (25:03):
undefined
Josh: DeepMind team. And today we got a (25:06):
undefined
Josh: An agent that is capable of navigating these 3D virtual worlds. (25:10):
undefined
Josh: So we saw previously there was a few models that can generate the worlds. (25:13):
undefined
Josh: Well, now SEMA 2, which implies there was a SEMA 1 that I was blissfully unaware (25:16):
undefined
Josh: of, allows these AI agents to actually navigate these virtual worlds. (25:19):
undefined
Josh: So now not only can Google generate the virtual games, they literally look like (25:24):
undefined
Josh: video games in this in this video, but they can also have the characters navigate (25:28):
undefined
Josh: these complicated experiences, understand how things work, kind of piece things together. (25:32):
undefined
Josh: It's amazing to see because it feels like as i'm watching (25:37):
undefined
Josh: this video it is how a toddler would play (25:40):
undefined
Josh: a video game it's kind of they're slowly moving it's figuring (25:43):
undefined
Josh: it out you could see kind of reasoning in real time and navigating (25:46):
undefined
Josh: these spaces and it's important to understand as you're watching this everything (25:49):
undefined
Josh: that is on this is generated by ai because previously (25:52):
undefined
Josh: these look like trailers to video games and those those video (25:56):
undefined
Josh: games would have required thousands of people who are game developers lots (25:59):
undefined
Josh: of time on the game engine and all of that is (26:03):
undefined
Josh: done now fully by ai so seeing this announcement (26:05):
undefined
Josh: was really exciting for me because i'm like one i just love video games and (26:08):
undefined
Josh: two like oh my god wait these ais not only are they able to generate the environment (26:11):
undefined
Josh: but now these ai models are able to actually engage with the environment and (26:15):
undefined
Josh: what downstream effects does that have on training humanoid robots and the like (26:19):
undefined
Josh: which we saw with the tesla episode last week yeah (26:23):
undefined
Ejaaz: I mean yeah to your point um we see a real life implementation of this with (26:26):
undefined
Ejaaz: tesla tesla Tesla has a world model which they use in their robots, (26:31):
undefined
Ejaaz: which is the Tesla cars and automated self-driving and the future optimist robots. (26:37):
undefined
Ejaaz: And what's cool about this is, Josh, one of the main things that makes your (26:42):
undefined
Ejaaz: AI models super cool or really smart or intelligent is data. (26:46):
undefined
Ejaaz: There's a scarcity of data. There's a scarcity of rich human data. (26:51):
undefined
Ejaaz: Like they don't necessarily see what we see through our eyes or hear what we hear. (26:56):
undefined
Ejaaz: These simulated environments basically synthetically create this data and enhance (26:59):
undefined
Ejaaz: any of these AI models, right? (27:04):
undefined
Ejaaz: And so Tesla kind of gets smarter by replaying a scenario where it almost crashed (27:06):
undefined
Ejaaz: over and over and over again in simulated environments until it gets it right. (27:11):
undefined
Ejaaz: And then that just gets transported to every single production ready Tesla that (27:15):
undefined
Ejaaz: anyone is driving so that if they found themselves in a similar scenario, they avoid it. (27:20):
undefined
Ejaaz: And that's just like one small use case for like what these robots will eventually (27:24):
undefined
Ejaaz: be benefiting from. So such a super cool update. (27:28):
undefined
Josh: Yeah. And if you scroll down actually to post EJs, you'll see a lot of similarities (27:32):
undefined
Josh: to what we spoke about with Tesla, where you could see the model kind of reasoning in plain English. (27:35):
undefined
Josh: You could see, oh, we equipped the pickaxe and copper mine. And then, (27:39):
undefined
Josh: okay, like I'll equip the pickaxe and it walks you through its chain of thought (27:42):
undefined
Josh: and then not only that but the post below it shows that it's self-improving (27:45):
undefined
Josh: so the more time it spends cycling through the more time it spends playing games (27:48):
undefined
Josh: the better it gets and these are early implementations of what what we kind (27:52):
undefined
Josh: of imagined ai would eventually be which is this self-recursive loop where it (27:56):
undefined
Josh: can improve without outside information and it's really exciting to see it happening (27:59):
undefined
Josh: in the virtual world, because I guess that's the only place it can make sense (28:04):
undefined
Josh: where it costs much less. (28:08):
undefined
Josh: You can't harm anybody if a robot goes rogue, but it can learn without the inputs (28:09):
undefined
Josh: of other people. So that is the exciting thing about this, I think. (28:13):
undefined
Josh: And I'm really excited for the DeepMind team, man. I hope they keep going. (28:17):
undefined
Ejaaz: Well, dude, like, do you remember our conversation with Logan Kilpatrick, (28:20):
undefined
Ejaaz: the head of Google's AI studio? (28:24):
undefined
Ejaaz: If you haven't seen that episode, definitely go and check it out. (28:26):
undefined
Ejaaz: One of his main takeaways or one of my main takeaways is all of Google's AI (28:29):
undefined
Ejaaz: byproducts feeds off of different models within the same suite. (28:34):
undefined
Ejaaz: So if their video model learns something cool, they can transpose that onto (28:38):
undefined
Ejaaz: their LLM, which is like in Word, so that it gets recursively smarter over time. (28:43):
undefined
Ejaaz: I just think it's like fascinating and world models kind of like manifest all (28:47):
undefined
Ejaaz: of these different things in one particular simulated environment. (28:52):
undefined
Ejaaz: So, so cool. Kind of hard to wrap my head around sometimes, but because it looks (28:56):
undefined
Ejaaz: like a video game to your point but but awesome. (29:00):
undefined
Josh: Yeah it's you have to break this gap between like your normal understanding (29:01):
undefined
Josh: when you see these videos and think like oh my god wait this is not normal this (29:06):
undefined
Josh: is actually all ai generated um but with that we have concluded all the fun (29:09):
undefined
Josh: new updates from this week it was a (29:12):
undefined
Josh: i'd say a medium week it wasn't anything crazy there was no crazy outliers but (29:14):
undefined
Josh: a solid week overall every week there's so much stuff going on and the progress (29:18):
undefined
Josh: i feel like we're almost getting immune to it how fast things are moving um (29:21):
undefined
Josh: but i hope everyone found this interesting if you did please don't remember (29:25):
undefined
Josh: to share with your friends, like, subscribe. (29:29):
undefined
Josh: Ejaz, I know you were manifesting some sort of send-off today. Do you want to share? (29:31):
undefined
Ejaaz: Yes. Similar to Michael Burry, you shouldn't be shorting the biggest podcast bubble in the world. (29:35):
undefined
Ejaaz: And Limitless is at the forefront over here. We're delivering you all the latest AI news. (29:41):
undefined
Ejaaz: And just like the trade that Burry made that he lost $300 million in, (29:46):
undefined
Ejaaz: you don't want to be shorting the best leading AI podcast in the world. (29:51):
undefined
Ejaaz: Give us a five-star rating. (29:55):
undefined
Ejaaz: Subscribe if you're not. What was the stat, Josh? 80% of listeners? (29:56):
undefined
Ejaaz: 83%? How many of them don't subscribe? Over 80%. (30:00):
undefined
Ejaaz: Over 80. That's nuts. It helps us out massively. (30:04):
undefined
Ejaaz: We are currently in the top 30 in the technology podcast across Spotify, (30:08):
undefined
Ejaaz: Apple, and wherever you listen. Help us get to the top 10. It would mean a lot. What do we say? (30:12):
undefined
Josh: Don't ever bet against the optimists. And by subscribing and by being a part (30:16):
undefined
Josh: of this, you are part of the optimists. (30:20):
undefined
Josh: And we are going to move forward and things are going to progress. (30:21):
undefined
Josh: And just don't bet against the optimists. So thank you for being optimistic (30:25):
undefined
Josh: with us and being on this journey with us and we'll be back next week with a (30:29):
undefined
Josh: whole new slew of episodes so thank you for watching and we'll see you then (30:33):
undefined
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com