1 00:00:00,020 --> 00:00:03,760 Josh: We had a pretty big and exciting week in the world of AI and frontier technology. 2 00:00:03,920 --> 00:00:07,800 Josh: Lots of things happening. New AI hardware devices unveiled, rocket ships exploding. 3 00:00:07,960 --> 00:00:11,700 Josh: I want to start with the Microsoft event. Microsoft held their build event. 4 00:00:11,820 --> 00:00:17,480 Josh: Microsoft notoriously has been in the AI race only via companionship to OpenAI. 5 00:00:17,620 --> 00:00:19,500 Josh: They are the largest investor, largest owner of OpenAI. 6 00:00:19,740 --> 00:00:23,540 Josh: And that has been the connection. This week, they severed even more ties with 7 00:00:23,540 --> 00:00:26,900 Josh: OpenAI as they decided to build their own new hardware. 8 00:00:27,160 --> 00:00:31,400 Josh: Now, they released seven new AI models. they released two new AI devices and 9 00:00:31,400 --> 00:00:35,860 Josh: to hear that Microsoft is building an AI hardware device was a little shocking. 10 00:00:36,060 --> 00:00:40,080 Josh: So I went on this kind of like race to try to figure out more information to 11 00:00:40,080 --> 00:00:42,960 Josh: understand how it worked and what I found is these two devices that we're seeing 12 00:00:42,960 --> 00:00:48,460 Josh: on screen. We have a desktop device that looks kind of like an Amazon Echo Home 13 00:00:48,460 --> 00:00:50,160 Josh: Hub and apparently it's very smart. 14 00:00:50,280 --> 00:00:53,500 Josh: You can converse with it. It works with their new AI models and the second device 15 00:00:53,500 --> 00:00:56,520 Josh: is this small looking key card that's something you wear around your neck. 16 00:00:56,740 --> 00:01:00,220 Josh: It's very smart. It has a display built into it. And it's used for a lot of 17 00:01:00,220 --> 00:01:04,240 Josh: kind of practical in the job type use cases. 18 00:01:04,720 --> 00:01:09,080 Josh: What I found out that disappointed me is that all of this was a concept. 19 00:01:09,260 --> 00:01:13,100 Josh: They marketed this as this amazing product that everyone was going to be able 20 00:01:13,100 --> 00:01:15,000 Josh: to use showcasing all of the use cases. 21 00:01:15,120 --> 00:01:18,480 Josh: But the reality is, is that they're never actually going to ship these. 22 00:01:18,620 --> 00:01:23,000 Josh: In fact, this is just a concept for other companies to go off and get inspired 23 00:01:23,000 --> 00:01:25,300 Josh: to then build their own versions of. 24 00:01:25,420 --> 00:01:28,600 Josh: So here I was thinking, oh my god wait microsoft is making a comeback they have 25 00:01:28,600 --> 00:01:32,160 Josh: seven new models they're releasing new hardware and i think the reality is is 26 00:01:32,160 --> 00:01:35,740 Josh: that microsoft still kind of feels like they're in the same place they were 27 00:01:35,740 --> 00:01:37,120 Josh: prior to this build event 28 00:01:37,120 --> 00:01:39,960 Ejaaz: Yeah i've watched this video twice now and 29 00:01:39,960 --> 00:01:44,120 Ejaaz: i still don't know what these devices do um i feel like i've been transported 30 00:01:44,120 --> 00:01:47,760 Ejaaz: back a decade ago and amazon's just announced the echo again also why are we 31 00:01:47,760 --> 00:01:53,520 Ejaaz: wearing lanyards as like futuristic ai tech now i got hopeful when i saw this 32 00:01:53,520 --> 00:01:57,880 Ejaaz: announcement josh but i almost immediately killed that hope for one specific reason. 33 00:01:58,640 --> 00:02:05,100 Ejaaz: I don't know anyone that is conceivably using Microsoft's AI products to a larger extent, right? 34 00:02:05,280 --> 00:02:09,920 Ejaaz: So Copilot is like their flagship AI assistant, and they've admitted that their 35 00:02:09,920 --> 00:02:14,660 Ejaaz: Copilot product is way behind the likes of ChatGPT and Claude itself. 36 00:02:14,840 --> 00:02:19,580 Ejaaz: So if I already know that people aren't really using the Microsoft AI ecosystem 37 00:02:19,580 --> 00:02:23,020 Ejaaz: at the software level, why on earth would they use it at the hardware level? 38 00:02:23,100 --> 00:02:27,920 Ejaaz: Now, I dug in to a little bit about what this thing is supposedly meant to do. 39 00:02:27,920 --> 00:02:33,380 Ejaaz: And it's supposedly meant to be like an AI agent kind of like physical surface area. 40 00:02:33,500 --> 00:02:36,060 Ejaaz: So what I mean by that is it's meant to listen to the things that you do. 41 00:02:36,200 --> 00:02:39,960 Ejaaz: It's meant to see the things that you write or speak about and then it's meant 42 00:02:39,960 --> 00:02:43,060 Ejaaz: to kind of like ingest this data and improve your own personal agent, 43 00:02:43,160 --> 00:02:45,740 Ejaaz: your own personal model. There's one issue though. 44 00:02:46,220 --> 00:02:50,780 Ejaaz: Microsoft doesn't even have a functioning V1 of an AI agent that can do all of this for you. 45 00:02:50,920 --> 00:02:54,560 Ejaaz: They kind of brand it as Copilot but it's really ChatGPT and it's an older model 46 00:02:54,560 --> 00:02:59,080 Ejaaz: of ChatGPT and it's less effective than ChatGPT. So why wouldn't I just use ChatGPT? 47 00:02:59,180 --> 00:03:03,240 Ejaaz: Now, understandably, Microsoft recently has been trying to move away from their 48 00:03:03,240 --> 00:03:04,260 Ejaaz: partnership with OpenAI. 49 00:03:04,440 --> 00:03:08,040 Ejaaz: There was a bit of tension between Sam Altman and Satni and Odella. 50 00:03:08,220 --> 00:03:10,860 Ejaaz: And they're trying to kind of like sever that partnership and start becoming 51 00:03:10,860 --> 00:03:14,100 Ejaaz: a more multimodal operation. They've started using Claude for a lot of their 52 00:03:14,100 --> 00:03:15,820 Ejaaz: products and other types of models. 53 00:03:16,000 --> 00:03:18,400 Ejaaz: This is a step into the device side of things. 54 00:03:18,640 --> 00:03:22,640 Ejaaz: This wasn't exciting to me. And it maddens me so much that this is just a concept 55 00:03:22,640 --> 00:03:25,260 Ejaaz: that apparently they don't plan to build at all. 56 00:03:25,260 --> 00:03:30,980 Ejaaz: What did excite me was something that we covered earlier on in an earlier episode this week, 57 00:03:31,100 --> 00:03:36,480 Ejaaz: which was their new Microsoft Surface AI engineered laptop, which is meant to 58 00:03:36,480 --> 00:03:40,360 Ejaaz: run like one of Nvidia's latest chips, which allows you to kind of like run models locally. 59 00:03:40,500 --> 00:03:42,560 Ejaaz: But this doesn't really inspire me. 60 00:03:43,360 --> 00:03:47,200 Josh: Well, something that might inspire some people if the hardware isn't doing it 61 00:03:47,200 --> 00:03:50,380 Josh: is the software because Microsoft isn't just sitting there being dormant. 62 00:03:50,520 --> 00:03:53,780 Josh: They are working on their own AI models. They're working on large language models 63 00:03:53,780 --> 00:03:58,180 Josh: that I have to assume are somewhat derived from ChatGPT models because they 64 00:03:58,180 --> 00:04:01,820 Josh: have that proprietary license. They must be leaning on it to some extent. 65 00:04:01,960 --> 00:04:05,640 Josh: And what we have here is seven new models that they're working on in parallel. 66 00:04:05,860 --> 00:04:08,600 Josh: And initially I was a little frustrated by this too. I'm like, 67 00:04:08,720 --> 00:04:12,540 Josh: why the hell are they working on seven models at once? But in the context of 68 00:04:12,540 --> 00:04:14,680 Josh: this image that we're seeing on screen, it makes a lot of sense. 69 00:04:14,840 --> 00:04:19,960 Josh: The models are supposed to be working on specific modalities all together. 70 00:04:19,960 --> 00:04:23,900 Josh: So we have an image model. We have a large one and a flash one for smaller. 71 00:04:24,040 --> 00:04:27,300 Josh: We have a transcription model. We have a reasoning model. We have two voice 72 00:04:27,300 --> 00:04:30,460 Josh: models and a coding model. So they're building all of these in parallel. 73 00:04:30,940 --> 00:04:34,260 Josh: I have to imagine they're going to be building the orchestrator or at least 74 00:04:34,260 --> 00:04:38,700 Josh: using Copilot as the kind of orchestration harness engine for these models, 75 00:04:38,820 --> 00:04:39,660 Josh: knowing what to serve them. 76 00:04:39,860 --> 00:04:42,220 Josh: And I think it's at least a start. 77 00:04:42,520 --> 00:04:45,740 Josh: Microsoft has done pretty much nothing in the world of AI. 78 00:04:45,940 --> 00:04:49,620 Josh: I'm not sure anybody uses Copilot who isn't forced to buy their business. 79 00:04:49,780 --> 00:04:53,780 Josh: And I'm hopeful that these modalities, these models that they're working on 80 00:04:53,780 --> 00:04:56,320 Josh: will actually yield some sort of impressive results from Microsoft. 81 00:04:56,580 --> 00:05:02,240 Josh: Otherwise, I'm getting close to putting Microsoft in the Apple category of just like did not finish. 82 00:05:02,420 --> 00:05:05,620 Josh: Like they just like we're not able to actually execute able to deliver anything 83 00:05:05,620 --> 00:05:10,440 Josh: valuable in this ai model race maybe this changes things who knows but 84 00:05:11,160 --> 00:05:14,320 Josh: After watching this, I kind of didn't feel that way. 85 00:05:14,440 --> 00:05:17,280 Josh: It was like, okay, the hardware that you designed that was really cool looking, 86 00:05:17,420 --> 00:05:20,340 Josh: it's just a concept. The software, well, it's great on paper. 87 00:05:21,040 --> 00:05:23,580 Josh: These haven't been built. These haven't been deployed. It would have been awesome 88 00:05:23,580 --> 00:05:24,920 Josh: if we could actually play with them today. 89 00:05:25,140 --> 00:05:28,440 Josh: That is not the case. So there's still a lot of work that needs to happen for 90 00:05:28,440 --> 00:05:30,540 Josh: Microsoft, but that was the build event. 91 00:05:30,920 --> 00:05:34,960 Ejaaz: Yeah, I mean, the best thing that Microsoft has done in the entire AI race was 92 00:05:34,960 --> 00:05:39,020 Ejaaz: invest in their biggest competitor, OpenAI. When it comes to actually building 93 00:05:39,020 --> 00:05:42,420 Ejaaz: models, it's not exactly their specialty, but I applaud them for giving it a go. 94 00:05:42,920 --> 00:05:48,040 Ejaaz: Now, one might think that they're using OpenAI's IP. They have OpenAI's IP for 95 00:05:48,040 --> 00:05:51,300 Ejaaz: ChatGPT because they're one of the largest investors in that until I think it's 96 00:05:51,300 --> 00:05:56,100 Ejaaz: 2030 or 2032. So they can technically use this IP to create their own models. 97 00:05:56,300 --> 00:06:00,480 Ejaaz: But for one of these models specifically, which is the one I'm circling right 98 00:06:00,480 --> 00:06:05,620 Ejaaz: now, MAI Thinking 1, that is their main reasoning model. 99 00:06:05,740 --> 00:06:08,720 Ejaaz: It's their main orchestrator model that orchestrates all these other transcription, 100 00:06:09,040 --> 00:06:12,140 Ejaaz: image, video, and audio models. 101 00:06:12,520 --> 00:06:17,760 Ejaaz: And it hasn't been distilled or using OpenAI IP at all. 102 00:06:17,940 --> 00:06:22,020 Ejaaz: It's been built from the ground up, which is pretty ambitious for Microsoft. 103 00:06:22,160 --> 00:06:25,380 Ejaaz: Now, you might be wondering, okay, what are some of the advantages that it's unlocked? 104 00:06:25,520 --> 00:06:30,120 Ejaaz: Well, collectively, across all of these different models, it uses 60% fewer 105 00:06:30,120 --> 00:06:32,020 Ejaaz: tokens than your Frontier models. 106 00:06:32,020 --> 00:06:36,420 Ejaaz: But the models that it can kind of compare with and that it's on par with are 107 00:06:36,420 --> 00:06:41,200 Ejaaz: like three generations behind the frontier models that we're seeing from Anthropic and OpenAI right now. 108 00:06:41,320 --> 00:06:44,760 Ejaaz: So it's a good attempt, but it's got me thinking like, 109 00:06:44,920 --> 00:06:48,920 Ejaaz: I don't know why Microsoft is focused on this specifically when they should 110 00:06:48,920 --> 00:06:53,320 Ejaaz: be kind of like doubling down on like the AI agent harness and trying to build 111 00:06:53,320 --> 00:06:57,540 Ejaaz: a better AI experience across the humongous distribution that they have. 112 00:06:57,600 --> 00:07:01,560 Ejaaz: They have like hundreds of millions of co-pilot users across the world. 113 00:07:01,720 --> 00:07:03,960 Ejaaz: Like they should be able to kind of like monetize that somehow. 114 00:07:03,960 --> 00:07:06,300 Ejaaz: And I think they're kind of wasting that time here. 115 00:07:07,570 --> 00:07:10,850 Ejaaz: There was a bunch of Microsoft announcements, and one of them was announced 116 00:07:10,850 --> 00:07:15,670 Ejaaz: in partnership with NVIDIA, and it was with their Surface laptop. 117 00:07:16,250 --> 00:07:19,990 Ejaaz: And NVIDIA has been kind of like on a streak this week. So there's this conference 118 00:07:19,990 --> 00:07:24,090 Ejaaz: that's taking place in Taiwan right now called Computex. 119 00:07:24,250 --> 00:07:30,590 Ejaaz: And I'm kind of calling it NVIDIA GTC, which is their main conference 2.0. 120 00:07:30,710 --> 00:07:34,910 Ejaaz: It's basically Jensen's show, and he's taken the stage. Everyone's kind of like 121 00:07:34,910 --> 00:07:38,730 Ejaaz: crowded out the entire rooms and he's made a bunch of big announcements for NVIDIA. 122 00:07:38,810 --> 00:07:44,910 Ejaaz: One of the main ones is Vera Rubin NVL72, which is the kind of like it's the 123 00:07:44,910 --> 00:07:49,270 Ejaaz: next hottest generation that has actually been built for GPUs. 124 00:07:49,580 --> 00:07:53,220 Ejaaz: Is finally coming online. And there's a few partners that are bringing it online. 125 00:07:53,420 --> 00:07:56,240 Ejaaz: You've got Dell that we mentioned earlier this week, we put out an episode, 126 00:07:56,360 --> 00:07:59,460 Ejaaz: definitely go check that out, because Dell's actually a sleeper hit for AI investing. 127 00:08:00,300 --> 00:08:06,840 Ejaaz: CoreWeave and a number of other providers have helped stand up the very first Vero Rubin stack. 128 00:08:07,200 --> 00:08:12,140 Ejaaz: It looks incredibly complex and quite impressive from an engineering feat. 129 00:08:12,460 --> 00:08:15,320 Ejaaz: I don't know if you've seen a picture of this, Josh, but like there are pipes 130 00:08:15,320 --> 00:08:17,660 Ejaaz: and wires teeming out of this thing. 131 00:08:17,800 --> 00:08:20,920 Ejaaz: And it's actually bigger than me, if I sit on, I'm around like 6.3. 132 00:08:21,060 --> 00:08:25,680 Ejaaz: It's like a humongous type thing, incredibly impressive feat of networking architecture. 133 00:08:25,760 --> 00:08:29,220 Ejaaz: That is one of like the major announcements, like Jensen's kind of like leaning 134 00:08:29,220 --> 00:08:31,080 Ejaaz: into a lot of new hardware stacks. 135 00:08:31,240 --> 00:08:34,640 Ejaaz: Now, if you've watched an episode that we put out about a month ago, 136 00:08:34,820 --> 00:08:38,400 Ejaaz: we described why Jensen's kind of going all in on chips, if you didn't think 137 00:08:38,400 --> 00:08:41,120 Ejaaz: he already was doing this, by recreating 138 00:08:41,120 --> 00:08:45,680 Ejaaz: 15 of the core components that is required to build the best GPU. 139 00:08:46,060 --> 00:08:49,240 Ejaaz: Now, typically, GPU providers aren't recommended to do that because if something 140 00:08:49,240 --> 00:08:52,040 Ejaaz: goes wrong, you end up needing to delay the launch of your next model. 141 00:08:52,240 --> 00:08:58,680 Ejaaz: Not only did Jensen go for it, but he pulled it off. And so now we see these new GPUs come to life. 142 00:08:58,800 --> 00:09:02,880 Ejaaz: He's also entering a new market, which is the CPU market. 143 00:09:03,060 --> 00:09:08,000 Ejaaz: Typically, Jensen likes to play in his lane. He kind of like sits at his own level of the stack. 144 00:09:08,140 --> 00:09:11,060 Ejaaz: He owns the picks and shovels and he doesn't want to get involved in anything else. 145 00:09:11,180 --> 00:09:15,300 Ejaaz: But he's peeked over at Intel and AMD and he's like, damn, these guys are making 146 00:09:15,300 --> 00:09:16,900 Ejaaz: a bunch of money from CPUs. 147 00:09:17,060 --> 00:09:21,080 Ejaaz: And in a future world where AI agents are everywhere, we're going to need CPUs 148 00:09:21,080 --> 00:09:22,100 Ejaaz: to orchestrate these guys. 149 00:09:22,340 --> 00:09:26,940 Ejaaz: And so he announced the launch of NVIDIA CPUs, which currently is on track for 150 00:09:26,940 --> 00:09:30,640 Ejaaz: a $20 billion revenue run rate this year. It's pretty insane. 151 00:09:30,920 --> 00:09:33,380 Josh: This is the first time I've ever heard of Computex as an event. 152 00:09:33,480 --> 00:09:38,560 Josh: I didn't really even know that this existed, but it's basically the CES version in Taiwan. 153 00:09:38,780 --> 00:09:42,780 Josh: Like the consumer electronic show it happens every january in the united states 154 00:09:42,780 --> 00:09:45,740 Josh: it's where all of the cool new technology and hardware gets 155 00:09:46,430 --> 00:09:49,350 Josh: Announced with computex it felt like 156 00:09:49,350 --> 00:09:52,090 Josh: the gaming version almost of this and it's weird to 157 00:09:52,090 --> 00:09:55,190 Josh: say that because sure a lot of this is ai focused but a 158 00:09:55,190 --> 00:09:58,650 Josh: lot of the offerings a lot of the hardware was based on it's 159 00:09:58,650 --> 00:10:02,150 Josh: funny there's there's a lot of crossover it turns out between ai local 160 00:10:02,150 --> 00:10:04,970 Josh: inference running and gaming and that crossover is the 161 00:10:04,970 --> 00:10:07,870 Josh: gpu and famously nvidia stopped producing the 162 00:10:07,870 --> 00:10:10,810 Josh: higher end gpus that everyone was using in their computers to 163 00:10:10,810 --> 00:10:13,590 Josh: run local inference but also to run video games 164 00:10:13,590 --> 00:10:16,710 Josh: at very high fidelity what's happening here is through 165 00:10:16,710 --> 00:10:20,430 Josh: the release of the rtx spark which is pretty much the flagship announcement 166 00:10:20,430 --> 00:10:26,410 Josh: from this event that's the new super computer laptop macbook competitor it's 167 00:10:26,410 --> 00:10:30,570 Josh: now bringing the power back to the consumer not only for gaming but also for 168 00:10:30,570 --> 00:10:34,550 Josh: that edge ai compute and we have pricing for it which i thought was pretty interesting 169 00:10:34,550 --> 00:10:40,110 Josh: and competitive the n1 systems featuring this rtx spark they are going to cost $1,800. 170 00:10:40,450 --> 00:10:44,750 Josh: And that is for a very high end GPU that you can run locally on your machine 171 00:10:44,750 --> 00:10:49,150 Josh: to run edge inference if you want to try small models or to actually run very 172 00:10:49,150 --> 00:10:50,150 Josh: high fidelity video games. 173 00:10:50,270 --> 00:10:55,590 Josh: The higher end of the system starts at $2,900. That includes the N1X system. 174 00:10:55,750 --> 00:11:00,430 Josh: And that can run like pretty beefy language models locally on your machine. 175 00:11:00,570 --> 00:11:05,950 Josh: It has 128 gigabytes of memory. It has, I think, 6,144 CUDA cores, 176 00:11:06,070 --> 00:11:08,550 Josh: which for those who are unfamiliar with CUDA cores, it's a lot. 177 00:11:08,550 --> 00:11:10,070 Josh: You can just do a lot of AI stuff on them. 178 00:11:10,530 --> 00:11:13,270 Josh: And it's really impressive. The other announcement that I really loved is this 179 00:11:13,270 --> 00:11:15,330 Josh: thing called a NUC, an N-U-K. 180 00:11:15,530 --> 00:11:18,290 Josh: It's like they create these small computers. If you've ever ran an Ethereum 181 00:11:18,290 --> 00:11:22,050 Josh: validator or a crypto validator, odds are it was run locally on a little device 182 00:11:22,050 --> 00:11:23,750 Josh: called a NUC. They've built these... 183 00:11:24,430 --> 00:11:27,610 Josh: Single purpose machines now with these new graphical 184 00:11:27,610 --> 00:11:31,030 Josh: cards built into it so now you could have a desktop machine that isn't 185 00:11:31,030 --> 00:11:33,910 Josh: really a laptop per se but it's meant specifically for 186 00:11:33,910 --> 00:11:37,570 Josh: one purpose and that purpose is running local edge inference on your device 187 00:11:37,570 --> 00:11:41,470 Josh: running video games the highest fidelity these are really powerful machines 188 00:11:41,470 --> 00:11:45,650 Josh: and i have to imagine this is beginning to or i guess this is a continuation 189 00:11:45,650 --> 00:11:49,870 Josh: of the trend that we've been seeing of companies pushing these consumers There's 190 00:11:49,870 --> 00:11:53,230 Josh: two edge devices for this type of compute. It was very cool to see. 191 00:11:53,510 --> 00:11:58,670 Ejaaz: Yeah, and it wouldn't be a Jensen Huang conference if he didn't bless one of 192 00:11:58,670 --> 00:12:02,110 Ejaaz: the next companies that are going to be potentially a trillion dollar in valuation. 193 00:12:02,110 --> 00:12:04,110 Ejaaz: And he said it very explicitly on stage. 194 00:12:04,270 --> 00:12:07,310 Ejaaz: The next trillion dollar company is Marvel. 195 00:12:07,630 --> 00:12:09,990 Ejaaz: Now, I got to pull up the ticker. Yeah, yeah, please. 196 00:12:10,750 --> 00:12:15,490 Ejaaz: MRVL and please limit it to the last five days of price action because after 197 00:12:15,490 --> 00:12:19,890 Ejaaz: he said those words, the next day the stock boomed up. I think it was like 100 198 00:12:19,890 --> 00:12:22,630 Ejaaz: to 150 percent, something obscene like that. 199 00:12:22,770 --> 00:12:26,530 Ejaaz: And if you're wondering why Jensen brought the CEO and founder of Marvel, 200 00:12:26,590 --> 00:12:29,890 Ejaaz: if you're wondering what the hell is Marvel doing, I must say it has got nothing 201 00:12:29,890 --> 00:12:32,830 Ejaaz: to do with Iron Man or any of the cinematography around that. 202 00:12:33,050 --> 00:12:38,870 Ejaaz: It is a much more boring kind of company, which is focused on networking architecture 203 00:12:38,870 --> 00:12:41,330 Ejaaz: for all these GPUs and data centers. 204 00:12:41,470 --> 00:12:46,190 Ejaaz: So a problem that a lot of AI labs have is when they're scaling their data centers. 205 00:12:46,330 --> 00:12:50,750 Ejaaz: They get all these brand new expensive GPUs, but you need to be able to connect 206 00:12:50,750 --> 00:12:56,150 Ejaaz: these GPUs together and get them to communicate with them at lightning speed as quick as they can. 207 00:12:56,390 --> 00:13:00,590 Ejaaz: Sometimes these GPUs overheat, so you need to make sure you manage the electricity 208 00:13:00,590 --> 00:13:03,710 Ejaaz: and make sure that the data is transferred at the right amounts of times and 209 00:13:03,710 --> 00:13:06,450 Ejaaz: the right quantities overnight versus day. 210 00:13:06,510 --> 00:13:09,570 Ejaaz: There's a ton of complex networking architecture that is required. 211 00:13:09,730 --> 00:13:13,710 Ejaaz: There's wiring, pipelines, all this kind of stuff, Marvel as a company, 212 00:13:14,070 --> 00:13:16,610 Ejaaz: as one of the leaders that specialize in this. 213 00:13:16,770 --> 00:13:21,990 Ejaaz: And Jensen Huang placed a $2 billion check via NVIDIA into them. I think it was about, 214 00:13:22,500 --> 00:13:26,920 Ejaaz: in March, potentially. And then he made this announcement basically saying like, 215 00:13:27,020 --> 00:13:29,760 Ejaaz: you know, Marvel is the ones that are going to basically allow NVIDIA to kind 216 00:13:29,760 --> 00:13:34,400 Ejaaz: of like build out its moat as a data center kind of like expert. 217 00:13:35,040 --> 00:13:39,560 Ejaaz: And their stock went up a massive amount of time. And like he's pointing at 218 00:13:39,560 --> 00:13:41,480 Ejaaz: one very important thing, which I want to mention. 219 00:13:42,020 --> 00:13:45,940 Ejaaz: The next bottleneck constraint. So we've spoken about GPUs, compute. 220 00:13:46,140 --> 00:13:49,560 Ejaaz: We've spoken about memory and architecture for memory a lot here. 221 00:13:49,720 --> 00:13:53,660 Ejaaz: And those stocks have gone up. he's saying the next bottleneck is going to be 222 00:13:53,660 --> 00:13:57,600 Ejaaz: on power networking specifically and that's what Marvel specializes in they 223 00:13:57,600 --> 00:14:00,960 Ejaaz: specialize in like optical fiber networks and a ton of other stuff there so 224 00:14:00,960 --> 00:14:04,940 Ejaaz: that's going to be an interesting one to view but moving on, 225 00:14:05,880 --> 00:14:11,820 Ejaaz: there's been a massive explosion. I saw this all over my feed, Josh, over the weekend. 226 00:14:12,280 --> 00:14:17,380 Ejaaz: And it's actually pretty intimidating to an extent, but this is Blue Origin's 227 00:14:17,380 --> 00:14:24,560 Ejaaz: New Glenn rocket, which was being tested at Cape Canaveral and unfortunately exploded massively. 228 00:14:24,740 --> 00:14:27,300 Ejaaz: What is going on here? Why did this happen? 229 00:14:27,360 --> 00:14:30,220 Josh: Yeah, it's pretty intense. I actually had a post about this. 230 00:14:30,320 --> 00:14:32,440 Josh: This is one of my largest ever 5 million people saw. 231 00:14:32,600 --> 00:14:37,520 Josh: And it was because the comparison to this explosion is that it is roughly 20% 232 00:14:37,520 --> 00:14:43,040 Josh: of the energy of the Hiroshima atomic bomb that was released on the launch pad of Blue Origin. 233 00:14:43,280 --> 00:14:48,140 Josh: It is a horribly tragic happening for the space program, for the space program 234 00:14:48,140 --> 00:14:50,340 Josh: as a whole, but also Blue Origin as a company. 235 00:14:50,700 --> 00:14:55,280 Josh: The pad, unfortunately, is the only pad that is capable of launching the New 236 00:14:55,280 --> 00:14:57,360 Josh: Glenn rockets to Earth, and now it's gone. 237 00:14:57,540 --> 00:15:02,140 Josh: It cost them over a billion dollars to make, and SpaceX needed well over seven 238 00:15:02,140 --> 00:15:05,260 Josh: months to rebuild theirs after they had a similar issue. 239 00:15:05,420 --> 00:15:10,280 Josh: So this is a pretty large setback and it comes back against the problem that 240 00:15:10,280 --> 00:15:16,200 Josh: Amazon needs, I think, the number of 1,618 satellites up in orbit by July 30th 241 00:15:16,200 --> 00:15:18,280 Josh: to maintain its license with the FCC. 242 00:15:18,600 --> 00:15:23,360 Josh: It currently has about 300. So the rocket that was supposed to really help with 243 00:15:23,360 --> 00:15:29,560 Josh: this contract is now gone and the pad that was supposed to be launching it from is gone with it. 244 00:15:29,680 --> 00:15:33,340 Josh: Now, they did have a response that I saw that seemed somewhat reassuring. 245 00:15:33,520 --> 00:15:37,380 Josh: It was from someone within the company. I think he's the, oh, 246 00:15:37,440 --> 00:15:38,800 Josh: actually the CEO of Blue Origin. It's funny. 247 00:15:38,940 --> 00:15:40,160 Ejaaz: I think Jeff Bezos is the CEO. 248 00:15:40,560 --> 00:15:44,160 Josh: Turns out it's this other guy named Dave Limp. And Dave Limp said that now that 249 00:15:44,160 --> 00:15:47,560 Josh: we've had access to the pad and integration facility, we could share a good bit of news. 250 00:15:47,700 --> 00:15:52,020 Josh: The propellant farm, oxygen, liquid hydrogen, and LNG tanks, 251 00:15:52,160 --> 00:15:54,060 Josh: liquid natural gas, are all in good shape. 252 00:15:54,460 --> 00:15:58,180 Josh: This is good because those are very long lead times. So it seems like a lot 253 00:15:58,180 --> 00:16:03,100 Josh: of the core components that were going to be causing a longer delay 254 00:16:03,960 --> 00:16:06,660 Josh: aren't actually affected too badly there was a 255 00:16:06,660 --> 00:16:09,600 Josh: support tower that if you watch the video that thing got destroyed 256 00:16:09,600 --> 00:16:12,420 Josh: that's the strong arm that holds the rocket up totally got wiped 257 00:16:12,420 --> 00:16:16,260 Josh: out he said well good news we were planning to get rid of that anyway so they're 258 00:16:16,260 --> 00:16:19,000 Josh: very much making light of the situation they're trying to work their hardest 259 00:16:19,000 --> 00:16:21,980 Josh: i just wanted to share this because it was one unbelievable explosion great 260 00:16:21,980 --> 00:16:25,480 Josh: content to watch if you are seeing the video version of this episode but also 261 00:16:25,480 --> 00:16:30,000 Josh: that it may not be the worst setback in the world and in terms of getting these 262 00:16:30,000 --> 00:16:31,360 Josh: Blue Origin satellites into space. 263 00:16:31,740 --> 00:16:34,620 Josh: Thankfully, they still do have SpaceX with the Falcon 9, Falcon Heavy, 264 00:16:34,800 --> 00:16:38,800 Josh: and soon-to-be Starship. So a bit of a setback in the space race. 265 00:16:39,180 --> 00:16:42,380 Josh: It's unfortunate. SpaceX really could benefit from some competition. 266 00:16:42,380 --> 00:16:45,460 Josh: We don't want them going into their IPO as a complete and total monopoly. 267 00:16:45,680 --> 00:16:48,020 Josh: But here we are. They're the only ones that can figure it out. 268 00:16:48,160 --> 00:16:52,340 Josh: And it's funny, we took space for granted because SpaceX has just been landing 269 00:16:52,340 --> 00:16:56,500 Josh: boosters back on barges in the middle of the ocean for like five plus years now. 270 00:16:56,560 --> 00:17:01,420 Josh: And it turns Turns out space and rocket science really are like pretty difficult categories. 271 00:17:01,640 --> 00:17:03,580 Josh: So wishing the team all the best. A bummer, but. 272 00:17:04,270 --> 00:17:05,610 Josh: Alas, we will carry on. 273 00:17:06,550 --> 00:17:10,410 Ejaaz: It's probably like the best timing for Elon Musk. He probably saw that and he 274 00:17:10,410 --> 00:17:13,490 Ejaaz: was like, this is upsetting for the industry as a whole, but perfect for my 275 00:17:13,490 --> 00:17:15,610 Ejaaz: IPO that's happening in probably next week. 276 00:17:16,170 --> 00:17:21,650 Ejaaz: Moving on, there was this interesting stat that we pulled from Cloudflare, 277 00:17:21,790 --> 00:17:26,610 Ejaaz: which basically said that agentic traffic is growing so fast that it now has 278 00:17:26,610 --> 00:17:27,970 Ejaaz: surpassed human traffic. 279 00:17:28,190 --> 00:17:30,870 Ejaaz: It is over 50% of internet traffic online. 280 00:17:31,270 --> 00:17:35,030 Ejaaz: Now, I can't say that this surprises me, as much. 281 00:17:35,150 --> 00:17:39,950 Ejaaz: I think the adoption and growth of people using AI LLMs and setting up their 282 00:17:39,950 --> 00:17:43,270 Ejaaz: own agents, vibe coding, different kinds of things that kind of like run autonomously 283 00:17:43,270 --> 00:17:47,630 Ejaaz: 24-7 has increased exponentially at this point. 284 00:17:47,770 --> 00:17:50,030 Ejaaz: And I think this is only going to get more aggressive. 285 00:17:50,410 --> 00:17:53,990 Ejaaz: We were having a discussion prior to starting this recording, 286 00:17:53,990 --> 00:17:57,730 Ejaaz: and I was like pondering about a world where like, you know, 287 00:17:57,890 --> 00:18:00,730 Ejaaz: future kids or generations of kids are largely going to 288 00:18:00,730 --> 00:18:04,030 Ejaaz: interact with entities or beings online 289 00:18:04,030 --> 00:18:06,910 Ejaaz: that don't have a physical body or if they do it 290 00:18:06,910 --> 00:18:09,690 Ejaaz: might be robotic and we got into like a debate 291 00:18:09,690 --> 00:18:13,250 Ejaaz: around like what that might potentially mean like from an educational stance 292 00:18:13,250 --> 00:18:16,790 Ejaaz: right if you had a tutor that can kind of like personally tailor its lessons 293 00:18:16,790 --> 00:18:21,050 Ejaaz: to you for this particular child that might be end that might end up being useful 294 00:18:21,050 --> 00:18:26,170 Ejaaz: but for someone who is like engaging in entertainment or leisure with a type 295 00:18:26,170 --> 00:18:30,890 Ejaaz: of ai personality it gets a little blurry as to whether this is good or not. 296 00:18:31,010 --> 00:18:34,390 Ejaaz: And I guess some versions of this already do exist. 297 00:18:34,590 --> 00:18:38,390 Ejaaz: Like I watch a lot of YouTube, I kind of follow my favorite creators online. 298 00:18:38,570 --> 00:18:39,970 Ejaaz: I don't really know them. I 299 00:18:39,970 --> 00:18:42,430 Ejaaz: just know them through the screen and through the content that I consume. 300 00:18:42,650 --> 00:18:45,570 Ejaaz: So what if it means that it's AI generated completely? 301 00:18:45,970 --> 00:18:48,670 Ejaaz: Anyway, I think we're going to see a lot of this going forwards. 302 00:18:48,810 --> 00:18:51,570 Ejaaz: I think we're going to interact with a lot of businesses that are completely 303 00:18:51,570 --> 00:18:54,370 Ejaaz: autonomously run by agents, but we would never know because we interact with 304 00:18:54,370 --> 00:18:57,190 Ejaaz: them via text or they generate their own physical, sorry, 305 00:18:57,550 --> 00:19:03,270 Ejaaz: digital AI avatars, but just an interesting stat and one that has gone under the radar, I think. 306 00:19:04,080 --> 00:19:07,420 Josh: Yeah, well, there's been this concept of the dead internet theory for a very 307 00:19:07,420 --> 00:19:11,560 Josh: long time, which basically says that a lot of the internet has ceased to be 308 00:19:11,560 --> 00:19:16,060 Josh: human-centric and is now mostly composed of bot activity. And we've seen this before. 309 00:19:16,240 --> 00:19:19,160 Josh: We've seen this through mostly an election cycle. 310 00:19:19,260 --> 00:19:23,280 Josh: So you remember back in 2016, a lot of the propaganda that we were seeing on 311 00:19:23,280 --> 00:19:26,400 Josh: social media aggregator sites that were being published to the people, 312 00:19:26,580 --> 00:19:27,740 Josh: a lot of it was bot activity. 313 00:19:27,740 --> 00:19:32,260 Josh: So we've had at least the appearance of bot activity, particularly on curated 314 00:19:32,260 --> 00:19:34,520 Josh: algorithmic feeds for a very long time. 315 00:19:34,820 --> 00:19:39,420 Josh: Well, Cloudflare sharing is basically the idea that not only is it greater than 316 00:19:39,420 --> 00:19:44,360 Josh: 50% of the content that you see, but genuinely now 57% as of recording this 317 00:19:44,360 --> 00:19:47,900 Josh: based on their charts is bot activity, provably bot activity. 318 00:19:48,060 --> 00:19:52,420 Josh: So we've officially crossed the chasm. It's no longer a dead internet theory. 319 00:19:52,580 --> 00:19:55,780 Josh: It is a dead internet and we are living in a bot driven world. the 320 00:19:55,780 --> 00:19:58,820 Josh: downstream implications of this are going to be interesting 321 00:19:58,820 --> 00:20:01,540 Josh: to think about to discuss i'm sure this is a very 322 00:20:01,540 --> 00:20:04,540 Josh: long conversation we're going to be having for many many years 323 00:20:04,540 --> 00:20:07,620 Josh: going forward but it's official the flipping has happened as 324 00:20:07,620 --> 00:20:10,540 Josh: confirmed by cloudflare one of the largest internet providers i 325 00:20:10,540 --> 00:20:14,160 Josh: think the largest internet provider in the world so it's 326 00:20:14,160 --> 00:20:18,400 Josh: funny too they were guiding they say in the post for the end of 2027 and then 327 00:20:18,400 --> 00:20:22,260 Josh: they bumped their timeline up to early 2027 and now we're in mid-2026 and that 328 00:20:22,260 --> 00:20:25,660 Josh: flip has happened so things are clearly accelerating faster than i think anyone 329 00:20:25,660 --> 00:20:30,540 Josh: expected and what that means for the internet as a whole for the people the 330 00:20:30,540 --> 00:20:33,340 Josh: human users of it is to be determined 331 00:20:34,920 --> 00:20:40,040 Ejaaz: Now, one thing that we've spoken a lot about on this show is the idea of scaling laws. 332 00:20:40,240 --> 00:20:43,540 Ejaaz: So that follows the rule where if you want to create a better AI model, 333 00:20:43,660 --> 00:20:46,560 Ejaaz: if you want to train a better AI model, if you want to inference a better AI model, 334 00:20:46,740 --> 00:20:50,700 Ejaaz: you need a lot more compute that comes in the form of GPUs, it comes in the 335 00:20:50,700 --> 00:20:55,460 Ejaaz: form of energy or electricity to power these GPUs and to train these different models. 336 00:20:56,040 --> 00:21:02,560 Ejaaz: Sarah Fryer, the CFO of OpenAI, is one of the few people probably in the industry 337 00:21:02,560 --> 00:21:09,020 Ejaaz: that knows to the bones how true this statement is or how true this scaling law actually is upheld. 338 00:21:09,180 --> 00:21:12,600 Ejaaz: And she had a very clear statement on her appearance on the All In conference 339 00:21:12,600 --> 00:21:16,420 Ejaaz: where she said, compute through 2027 is completely sold out. 340 00:21:16,740 --> 00:21:21,500 Ejaaz: And that is the clearest statement ever that the scaling laws are still kind of holding. 341 00:21:21,600 --> 00:21:27,800 Ejaaz: We're still in the early parts of 2026. We've heard similar trends for AI memory 342 00:21:27,800 --> 00:21:30,780 Ejaaz: being sold out through the end of 2027, but not really on compute. 343 00:21:30,980 --> 00:21:35,220 Ejaaz: And we also know that compute data centers specifically is the current focus 344 00:21:35,220 --> 00:21:39,380 Ejaaz: of every major AI CEO in terms of scaling. They're trying to like... 345 00:21:39,900 --> 00:21:43,200 Ejaaz: Boots on the ground and build these things, hire all the technical employees 346 00:21:43,200 --> 00:21:44,820 Ejaaz: that are required to set these things up. 347 00:21:44,940 --> 00:21:47,920 Ejaaz: It is the most aggressive expenditure. Google just raised $85 billion. 348 00:21:48,400 --> 00:21:52,220 Ejaaz: All these IPOs that we're talking about, Anthropic, OpenAI, and SpaceX is all 349 00:21:52,220 --> 00:21:55,640 Ejaaz: to get more money to invest in AI CapEx. It's the number one focus. 350 00:21:55,800 --> 00:22:00,580 Ejaaz: It's Krishna Rao, CFO of Anthropic, number one focus. 351 00:22:00,780 --> 00:22:04,940 Ejaaz: So for Sarah Fry to also confirm this and say that she feels like, 352 00:22:04,940 --> 00:22:09,620 Ejaaz: compute is going to be short for a long period of time through 2030 and 2032, 353 00:22:09,900 --> 00:22:13,240 Ejaaz: is the clearest statement ever that OpenAI, Anthropic, 354 00:22:13,660 --> 00:22:19,100 Ejaaz: and the likes of other top major AI labs aren't going to go away when it comes to AI capex. 355 00:22:19,320 --> 00:22:23,860 Ejaaz: Now, one thing that this forces me to think is, or like challenge myself is, 356 00:22:23,980 --> 00:22:27,800 Ejaaz: it probably means then that we are not limited by capital. 357 00:22:27,920 --> 00:22:32,160 Ejaaz: We're not limited by kind of this AI bubble that is being inflated. 358 00:22:32,300 --> 00:22:35,220 Ejaaz: We're being limited by physical atoms itself. 359 00:22:35,420 --> 00:22:38,460 Ejaaz: Like we don't have enough, no matter how much money you have, 360 00:22:38,620 --> 00:22:41,900 Ejaaz: we don't have enough human capital to actually build these things. 361 00:22:41,980 --> 00:22:42,920 Ejaaz: And so we're constrained by that. 362 00:22:43,020 --> 00:22:46,860 Ejaaz: The regulations determining what land to build on it takes a lot of time. 363 00:22:47,000 --> 00:22:50,280 Ejaaz: And that backlog, in my opinion, is probably going to go on for a lot more. 364 00:22:50,460 --> 00:22:53,040 Ejaaz: So we're probably going to see a lot more investments of this trend. 365 00:22:53,660 --> 00:22:56,480 Josh: Sarah Fryer is incredible. I would encourage everyone to go and watch this. There's 366 00:22:56,480 --> 00:22:59,460 Josh: a strong case to me that she should be running the company because man she 367 00:22:59,460 --> 00:23:02,200 Josh: is incredibly well spoken and thoughtful and revealed a lot 368 00:23:02,200 --> 00:23:06,240 Josh: of things that weren't obviously apparent to me one of them talking about the 369 00:23:06,240 --> 00:23:09,720 Josh: time frames in which she deals on like i'm very much under the impression that 370 00:23:09,720 --> 00:23:13,640 Josh: these ai labs are thinking maybe 12 to 24 months out because further than that 371 00:23:13,640 --> 00:23:17,580 Josh: the supply chains are nearly impossible predict she was sharing that they think 372 00:23:17,580 --> 00:23:21,780 Josh: like up to six years out and she's placing guidance for 2032, 373 00:23:22,020 --> 00:23:27,040 Josh: which to me feels like this impossibly complicated task. So maybe she's crazy. 374 00:23:27,280 --> 00:23:31,300 Josh: Maybe there's some kind of merit to these ideas, but it was a really fascinating 375 00:23:31,300 --> 00:23:33,740 Josh: episode from the all-in guys just to understand 376 00:23:34,570 --> 00:23:38,290 Josh: Kind of where OPAIs see themselves and how they plan to navigate through this 377 00:23:38,290 --> 00:23:41,530 Josh: next five to six year period in the sense that, like you're saying, 378 00:23:41,650 --> 00:23:43,650 Josh: EJS, there's just not enough compute. 379 00:23:43,790 --> 00:23:47,290 Josh: So therefore, they're not worried about any slowing demand because the demand 380 00:23:47,290 --> 00:23:50,310 Josh: is going to continue to outpace that compute over the next six years. 381 00:23:50,410 --> 00:23:54,690 Josh: And so long as that demand outpaces the expense, these companies are going to become profitable. 382 00:23:54,850 --> 00:23:57,690 Josh: They're going to continue to go up. And I mean, obviously, she's guiding for 383 00:23:57,690 --> 00:24:00,070 Josh: this positive future, but it seems like they have the numbers to back it up. 384 00:24:00,250 --> 00:24:03,510 Josh: A second thing she mentioned that I found very, very exciting was she talked 385 00:24:03,510 --> 00:24:06,950 Josh: about my favorite, our favorite hardware device, the Johnny Ive design device. 386 00:24:07,150 --> 00:24:09,990 Josh: She actually referenced it by name. 387 00:24:10,510 --> 00:24:14,890 Josh: Johnny Ive was mentioned and I think it was earbud that was mentioned or earpiece. 388 00:24:15,050 --> 00:24:18,390 Josh: She said Johnny would kill her if she called it an earpiece, 389 00:24:18,690 --> 00:24:22,230 Josh: kind of acknowledging the fact that this is actually something that goes in your ears. 390 00:24:22,370 --> 00:24:25,270 Josh: The things that we have been seeing are somewhat real. 391 00:24:25,850 --> 00:24:29,270 Josh: Jason, one of the people who hosts the show, he referenced the puck, 392 00:24:29,390 --> 00:24:30,990 Josh: the earpiece. She did not 393 00:24:32,080 --> 00:24:34,940 Josh: Rebuke she kind of was just like yep it's it's happening 394 00:24:34,940 --> 00:24:37,920 Josh: it's amazing she's used it yes like 395 00:24:37,920 --> 00:24:40,740 Josh: people internally have been using it it seems like this incredible device 396 00:24:40,740 --> 00:24:43,780 Josh: and she confirmed that it will be announced before the 397 00:24:43,780 --> 00:24:47,200 Josh: end of this year so we have confirmation at least on timelines happening 398 00:24:47,200 --> 00:24:50,620 Josh: sometime i would assume around q4 where we'll get more information and 399 00:24:50,620 --> 00:24:54,740 Josh: that is the holy grail of hardware i cannot wait to get those to try it out 400 00:24:54,740 --> 00:24:59,320 Josh: just to see what it looks like so that was really exciting uh open ai also they 401 00:24:59,320 --> 00:25:04,040 Josh: launched sites which is a cool new feature back to like practical ai use cases 402 00:25:04,040 --> 00:25:07,420 Josh: that you can actually use um tell us about sites ejaz i know you've been looking 403 00:25:07,420 --> 00:25:09,240 Josh: through how these kind of work so 404 00:25:09,240 --> 00:25:14,200 Ejaaz: I'm a very visual learner and one thing that i interact a lot with when i'm 405 00:25:14,200 --> 00:25:17,800 Ejaaz: creating kind of like research and agendas with claude is like can you create 406 00:25:17,800 --> 00:25:22,320 Ejaaz: me like a visual artifact um this is basically open ai's attempt to kind of 407 00:25:22,320 --> 00:25:24,600 Ejaaz: like not only compete, but excel for that product. 408 00:25:24,780 --> 00:25:28,340 Ejaaz: So in their own words, with sites, Codex can turn your work, 409 00:25:28,540 --> 00:25:34,220 Ejaaz: ideas and plans into an interactive website or app your team can explore, use and share with a URL. 410 00:25:34,380 --> 00:25:38,720 Ejaaz: Now, why I enjoy this the most is typically when you create a visual artifact, 411 00:25:39,100 --> 00:25:43,140 Ejaaz: it's static, it's like an image, it kind of you can kind of tell that it's created 412 00:25:43,140 --> 00:25:46,440 Ejaaz: by an AI, because it uses the same font, it's structured in a very particular way. 413 00:25:46,620 --> 00:25:50,500 Ejaaz: And what I typically do with that is I take that I take it to code code and 414 00:25:50,500 --> 00:25:55,740 Ejaaz: I say, Hey, can you create an HTML version of this and enhance it in a B and C types of ways? 415 00:25:56,320 --> 00:26:02,400 Ejaaz: This is a one button, one plugin, one feature fix that OpenAI is doing via codex. 416 00:26:02,520 --> 00:26:05,200 Ejaaz: So it's, you know, they're creating this like super app and this is kind of 417 00:26:05,200 --> 00:26:06,500 Ejaaz: like a step towards that direction. 418 00:26:06,720 --> 00:26:09,600 Ejaaz: Well, when you ask for a visual artifact, it will now do this. 419 00:26:09,680 --> 00:26:11,940 Ejaaz: And what I like about this is it's easily shareable. 420 00:26:12,060 --> 00:26:15,200 Ejaaz: So like you can create your own webpage, you can share it with someone and it 421 00:26:15,200 --> 00:26:18,680 Ejaaz: introduces just a new medium and way to kind of read and learn about something. 422 00:26:18,800 --> 00:26:23,080 Ejaaz: Like I would love to scroll a custom website with graphics, infographics that 423 00:26:23,080 --> 00:26:25,500 Ejaaz: are kind of like responsive to the way that I move my cursor, 424 00:26:25,500 --> 00:26:29,420 Ejaaz: that kind of explains concepts very clearly to me, I would learn much faster 425 00:26:29,420 --> 00:26:33,260 Ejaaz: that way versus just reading a block of text or looking at a static image. 426 00:26:33,400 --> 00:26:36,740 Ejaaz: And I just think it's a really neat feature and I want to see more companies 427 00:26:36,740 --> 00:26:37,960 Ejaaz: rolling out features like this. 428 00:26:38,260 --> 00:26:41,660 Josh: Yeah, it seems like it's going out to business and enterprise price plans currently. 429 00:26:41,820 --> 00:26:43,960 Josh: So if you have one of those, congratulations, you can go and try it out. 430 00:26:44,080 --> 00:26:48,440 Josh: I assume all the other plans will be following shortly after. I'm excited for this. 431 00:26:48,740 --> 00:26:52,680 Josh: So many people want to have their own websites. They've dreamed of making websites. 432 00:26:52,860 --> 00:26:57,340 Josh: They've paid tens to hundreds of thousands of dollars to have other people build websites. 433 00:26:57,520 --> 00:27:00,240 Josh: And now it's a matter of clicking a button and getting something that looks 434 00:27:00,240 --> 00:27:03,560 Josh: really lovely and that is functional and that accomplishes all of your goals. 435 00:27:03,660 --> 00:27:06,920 Josh: So I think in terms of people, just the average person, this is such a powerful 436 00:27:06,920 --> 00:27:11,480 Josh: tool and use case to empower them to build these websites. So that's great. 437 00:27:11,720 --> 00:27:13,860 Josh: We have two more topics quickly that we could probably breeze through. 438 00:27:14,040 --> 00:27:17,780 Josh: The first is Grok launching with Composer 2.5. 439 00:27:18,060 --> 00:27:21,900 Josh: I'm just reading the headline here, making it competitive product to Cloud Code. 440 00:27:23,000 --> 00:27:26,520 Josh: When I think of Cloud Code, I think of, oh my gosh, this model is unbelievable. 441 00:27:26,740 --> 00:27:27,920 Josh: People are actually doing real work. 442 00:27:28,620 --> 00:27:32,200 Josh: Awesome. When I think of Grok, I think of, oh yeah, that's right. 443 00:27:32,300 --> 00:27:34,680 Josh: That's the thing I used like six months ago and haven't touched since. 444 00:27:34,820 --> 00:27:36,360 Josh: So what's changed over those six months? 445 00:27:36,760 --> 00:27:42,260 Ejaaz: Yeah. So it's this one layer that sits on top of the model called the agent harness. 446 00:27:42,540 --> 00:27:47,860 Ejaaz: Now, a few weeks ago, we covered XAI or SpaceX AI forming a large partnership 447 00:27:47,860 --> 00:27:49,300 Ejaaz: with this company called Cursor. 448 00:27:49,480 --> 00:27:52,960 Ejaaz: Cursor doesn't actually create their own AI model, but they have specialized 449 00:27:52,960 --> 00:27:56,740 Ejaaz: remote known as the agent harness, which basically sits on top of any frontier 450 00:27:56,740 --> 00:28:01,100 Ejaaz: AI model and tells it, you know, kind of how to work, where to direct prompts. 451 00:28:01,280 --> 00:28:05,540 Ejaaz: It kind of like fine tunes it in a very specific way that actually includes 452 00:28:05,540 --> 00:28:11,020 Ejaaz: a lot of information that is very valuable if you're able to own it as a company. 453 00:28:11,140 --> 00:28:12,100 Ejaaz: Cursor has been able to do that. 454 00:28:12,260 --> 00:28:16,960 Ejaaz: Now, in this partnership, Elon has basically been able to acquire the right 455 00:28:16,960 --> 00:28:20,600 Ejaaz: to buy Cursor within 30 days of their IPO. 456 00:28:20,860 --> 00:28:24,780 Ejaaz: And as part of this agreement, that means he can take Cursor's very valuable 457 00:28:24,780 --> 00:28:29,560 Ejaaz: moat, their harness, slap it on Grok 4.3, which is the latest model that he's 458 00:28:29,560 --> 00:28:33,060 Ejaaz: training right now that is, I believe, publicly available now as of this week. 459 00:28:33,400 --> 00:28:36,520 Ejaaz: And hey, presto, you have a frontier coding model. 460 00:28:36,680 --> 00:28:41,160 Ejaaz: That's how powerful the harness layer is over the model. It's not just how good 461 00:28:41,160 --> 00:28:43,740 Ejaaz: your model is, it's how good your harness is as well. 462 00:28:43,940 --> 00:28:47,100 Ejaaz: Like Dario and Sam have both confirmed this. 463 00:28:47,280 --> 00:28:51,340 Ejaaz: And actually, if you slap Curse's harness right now on top of Claude Opus 4.8 464 00:28:51,340 --> 00:28:56,380 Ejaaz: or GPT 5.5 high, you end up having a better coding model than those models stand on their own. 465 00:28:56,580 --> 00:29:00,740 Ejaaz: So Elon has been able to buy his way to the frontier top. 466 00:29:00,920 --> 00:29:06,220 Ejaaz: Composer 2.5 on its own, which is Cursor's own model that they wrapped a harness 467 00:29:06,220 --> 00:29:10,980 Ejaaz: over, is comparable and on par with 4.8 and 5.5. 468 00:29:11,200 --> 00:29:14,700 Ejaaz: So if you wrap it around Grok 4.3, I can imagine the same is the case. 469 00:29:14,900 --> 00:29:17,580 Josh: That's pretty amazing. They have made some unbelievable progress. 470 00:29:17,720 --> 00:29:20,480 Josh: And I think, again, it's a testament to never count Elon out. 471 00:29:20,680 --> 00:29:22,480 Josh: They are making incredible progress. 472 00:29:22,780 --> 00:29:25,660 Josh: And as we know, they have the training models. They have the hardware. 473 00:29:26,080 --> 00:29:29,800 Josh: They've currently given it out to Anthropic. Elon restated publicly this week that 474 00:29:30,150 --> 00:29:34,690 Josh: It is more of a shorter term deal is kind of the idea that they both have the 475 00:29:34,690 --> 00:29:36,270 Josh: option to opt out at an inhuman moment. 476 00:29:36,450 --> 00:29:39,290 Josh: And I think the terms were 90 days that he said or something like that. 477 00:29:39,410 --> 00:29:43,230 Josh: But it is cool to see another competitor in the race. It's great to see another 478 00:29:43,230 --> 00:29:45,470 Josh: harness that people are excited about. I mentioned there was two topics. 479 00:29:45,630 --> 00:29:50,710 Josh: The last one is actually Google's $80 billion race that we covered yesterday in the episode. 480 00:29:50,870 --> 00:29:54,630 Josh: So unless there's net new information, I think I'm going to tease it a little bit. 481 00:29:54,770 --> 00:29:58,890 Josh: And then if you want to see the full thing, we could go back to our episode from yesterday. Yeah. 482 00:29:59,240 --> 00:29:59,580 Josh: Yeah. 483 00:29:59,680 --> 00:30:01,040 Ejaaz: And the only quick. 484 00:30:01,320 --> 00:30:02,980 Josh: Oh, yeah, what do you have? Do you have any updates? 485 00:30:03,320 --> 00:30:07,860 Ejaaz: Well, yeah, the only quick update is they doubled down and raised it to $85 billion. 486 00:30:08,240 --> 00:30:11,620 Ejaaz: So there's an extra $5 billion that they raised at the last minute. 487 00:30:11,800 --> 00:30:16,220 Ejaaz: And I just wanted to spend like 10 seconds saying a lot of people's takes on 488 00:30:16,220 --> 00:30:20,220 Ejaaz: this raise is it's bearish that Google selling stock, they're not confident 489 00:30:20,220 --> 00:30:23,040 Ejaaz: in the future of CapEx. I actually think it's the opposite. 490 00:30:23,340 --> 00:30:28,060 Ejaaz: I think Google has $126 billion on their balance sheet, and they can use that 491 00:30:28,060 --> 00:30:31,440 Ejaaz: to raise higher debt in the future. So they're being strategic. 492 00:30:31,620 --> 00:30:35,320 Ejaaz: This is financial bookkeeping, creative financial bookkeeping to make sure that 493 00:30:35,320 --> 00:30:39,260 Ejaaz: they have a healthy balance sheet in the future for when they need to raise even more money. 494 00:30:39,380 --> 00:30:43,480 Ejaaz: So selling a bit of stocks now won't get punished if their stock value ends 495 00:30:43,480 --> 00:30:46,760 Ejaaz: up going up more for the fewer shares that they end up having. 496 00:30:46,860 --> 00:30:49,860 Ejaaz: So I think this is ultimately a good move. And I just want to go on record saying 497 00:30:49,860 --> 00:30:51,500 Ejaaz: that for when it eventually is proven. 498 00:30:51,940 --> 00:30:54,860 Josh: And there you have it so if you have made it to this part of 499 00:30:54,860 --> 00:30:57,680 Josh: the episode congrats you're totally caught up on the week you can go touch 500 00:30:57,680 --> 00:31:00,560 Josh: grass go enjoy your weekend you have been exposed and up 501 00:31:00,560 --> 00:31:03,720 Josh: to date on everything as for the next few weeks buckle up 502 00:31:03,720 --> 00:31:07,000 Josh: it's going to get exciting next week we have wwdc from 503 00:31:07,000 --> 00:31:11,080 Josh: apple early in the week which is huge this is make or break for them if they 504 00:31:11,080 --> 00:31:14,360 Josh: cannot deliver again apple we're gonna have to write them off to zero we need 505 00:31:14,360 --> 00:31:19,180 Josh: the following week is seeming like it's going to be spacex ipo week that's going 506 00:31:19,180 --> 00:31:22,300 Josh: to be huge biggest ipo in history and i'm sure along the way, 507 00:31:22,380 --> 00:31:23,760 Josh: there will be many other releases. 508 00:31:24,000 --> 00:31:28,000 Josh: So buckle up. Very excited. Ejaz, any final parting thoughts before we head into our weekend? 509 00:31:28,380 --> 00:31:32,920 Ejaaz: No, that's it. I'm curious if any of you were out at any of the major Microsoft 510 00:31:32,920 --> 00:31:34,560 Ejaaz: Build or Computex events. 511 00:31:34,700 --> 00:31:37,060 Ejaaz: Let us know if there are any different takes that we missed. 512 00:31:37,200 --> 00:31:39,860 Ejaaz: But aside from that, wherever you listen to us on YouTube, Spotify, 513 00:31:40,100 --> 00:31:41,940 Ejaaz: Apple Music, please subscribe if you haven't. 514 00:31:42,060 --> 00:31:44,580 Ejaaz: Please give us a rating if you haven't. Please leave us a comment if you haven't. 515 00:31:44,800 --> 00:31:47,360 Ejaaz: All of these are incredibly valuable in helping us grow our channel. 516 00:31:47,480 --> 00:31:48,320 Ejaaz: And we have grown massively. 517 00:31:48,420 --> 00:31:51,460 Ejaaz: I think it's like 12,000 people over the last month, which is just absolutely 518 00:31:51,460 --> 00:31:53,160 Ejaaz: insane. It's lovely to meet you guys. 519 00:31:53,620 --> 00:31:56,140 Ejaaz: And yeah, we will see you on the next one next week. 520 00:31:56,180 --> 00:31:56,620 Josh: Thanks so much. 521 00:31:56,800 --> 00:31:57,240 Ejaaz: See you guys.