1 00:00:00,020 --> 00:00:04,220 Ejaaz: Everyone thinks that the most valuable resource in AI is the GPU. 2 00:00:04,460 --> 00:00:08,400 Ejaaz: It's why NVIDIA is worth $5 trillion, right? But what if I told you you're wrong? 3 00:00:08,520 --> 00:00:12,160 Ejaaz: What if I told you that the most valuable resource in AI is actually energy? 4 00:00:12,340 --> 00:00:14,720 Ejaaz: Without energy, you can't power these things. Without energy, 5 00:00:14,940 --> 00:00:17,860 Ejaaz: the GPU is just a worthless piece of metal junk. 6 00:00:18,140 --> 00:00:22,000 Ejaaz: And in fact, a lot of AI labs are facing this challenge where they just don't 7 00:00:22,000 --> 00:00:27,580 Ejaaz: have enough energy to power the tens of billions of dollars of GPUs that they've spent so far. 8 00:00:27,700 --> 00:00:31,920 Ejaaz: In fact, Satya Nadella of Microsoft recently said that he has hundreds of millions 9 00:00:31,920 --> 00:00:36,040 Ejaaz: of dollars worth of NVIDIA GPUs in his data centers collecting dust. 10 00:00:36,320 --> 00:00:39,420 Ejaaz: Now, there are a bunch of efforts focused on solving this energy crisis. 11 00:00:39,580 --> 00:00:43,520 Ejaaz: You've got nuclear energy, you've got renewable energy, but one particular solution 12 00:00:43,520 --> 00:00:48,900 Ejaaz: that's taking up a lot of mind space with Elon Musk and Google is literally out of this world. 13 00:00:49,040 --> 00:00:50,860 Ejaaz: I've got six words for you. AI 14 00:00:50,860 --> 00:00:56,100 Ejaaz: data centers in outer space. That is a real sentence that I've just said. 15 00:00:56,240 --> 00:01:00,240 Ejaaz: It's the concept of putting GPUs in outer space to harness the energy of the 16 00:01:00,240 --> 00:01:02,760 Ejaaz: sun for AI training and for AI inference. 17 00:01:02,940 --> 00:01:07,060 Ejaaz: And whilst it might sound like a far-fetched idea, it's not when you consider 18 00:01:07,060 --> 00:01:10,880 Ejaaz: that the resources needed on Earth to power up the data centers, 19 00:01:10,980 --> 00:01:16,440 Ejaaz: which by the way, are requiring as much energy as a mid to large city in the US. 20 00:01:16,700 --> 00:01:20,600 Ejaaz: You might need to look at alternative resources for this, the one being the 21 00:01:20,600 --> 00:01:23,380 Ejaaz: sun. And the economics kind of work out in theory. 22 00:01:23,560 --> 00:01:29,240 Ejaaz: You have up to three to five times more energy efficiency if you operate GPUs 23 00:01:29,240 --> 00:01:31,220 Ejaaz: above the Earth's atmosphere than below it. 24 00:01:31,320 --> 00:01:35,120 Ejaaz: So in this episode, we're going to dig into all things AI space data centers. 25 00:01:35,280 --> 00:01:36,700 Ejaaz: Is this real? Is this achievable? 26 00:01:36,880 --> 00:01:39,220 Ejaaz: Or is this kind of like a myth that we're going to laugh about and have pie 27 00:01:39,220 --> 00:01:41,120 Ejaaz: on our face six months from now? 28 00:01:41,720 --> 00:01:46,440 Ejaaz: Josh, you and I have a tumultuous relationship around this topic. 29 00:01:46,940 --> 00:01:53,220 Ejaaz: 50% of us were optimistic. 50% of us, you, were pessimistic about this. 30 00:01:53,440 --> 00:01:54,740 Ejaaz: What's your take on all of this? 31 00:01:54,900 --> 00:01:58,360 Josh: I was the biggest hater in the world. Space was so stupid. Why are we sending 32 00:01:58,360 --> 00:02:00,080 Josh: stuff to space when we have it all here on Earth? 33 00:02:00,420 --> 00:02:04,120 Josh: Well, my mind has changed. I have flipped 180 degrees. I'm pro space. 34 00:02:04,260 --> 00:02:07,760 Josh: I'm ready to send these things to space tomorrow. I want all the data centers 35 00:02:07,760 --> 00:02:09,180 Josh: in outer space immediately. 36 00:02:09,620 --> 00:02:12,200 Josh: And when I was considering why, I did a lot of research, which we're going to 37 00:02:12,200 --> 00:02:12,920 Josh: get into in this episode. 38 00:02:13,140 --> 00:02:18,220 Josh: But there's this combination of things where there is a power grid straining on the U.S. 39 00:02:18,320 --> 00:02:21,260 Josh: Where data centers now are one of the main drivers rising electricity demand. 40 00:02:21,460 --> 00:02:24,040 Josh: We can't fulfill all the electricity demand. 41 00:02:24,260 --> 00:02:26,700 Josh: There are limits on land, water, permitting, cooling. 42 00:02:27,600 --> 00:02:30,800 Josh: And I was thinking, I was like, well, if your house is full of all these gaming 43 00:02:30,800 --> 00:02:33,660 Josh: PCs and the power keeps going out, well, you start to ask yourself, 44 00:02:33,920 --> 00:02:36,380 Josh: well, where else can we put them? And you're like, on the roof? No. 45 00:02:36,640 --> 00:02:38,800 Josh: Well, why don't we just send them off into outer space? And thankfully, 46 00:02:39,120 --> 00:02:42,580 Josh: because of new technologies like SpaceX Starship, we can do this. 47 00:02:42,700 --> 00:02:45,880 Josh: So there are some benefits to space that I want to go over very quickly before 48 00:02:45,880 --> 00:02:49,500 Josh: we get into who's talking about this and who's doing this. One of them being... 49 00:02:50,220 --> 00:02:53,660 Josh: The most important one, unlimited solar energy. A lot of people don't know this, 50 00:02:53,760 --> 00:02:58,720 Josh: but 99.98% of the entire solar system is the sun. 51 00:02:58,940 --> 00:03:01,440 Josh: All of the mass of the solar system, it exists in the sun. 52 00:03:01,680 --> 00:03:06,880 Josh: And the sun is just this huge fusion reactor. And in fact, if you can get access 53 00:03:06,880 --> 00:03:10,060 Josh: to the sun from outside of our atmosphere, if you leave Earth's atmosphere and 54 00:03:10,060 --> 00:03:15,300 Josh: you go into just normal space, it's actually eight times more powerful than on Earth. 55 00:03:15,460 --> 00:03:18,320 Josh: So if you remove the constraint of Earth's atmosphere, immediately the 56 00:03:18,320 --> 00:03:21,060 Josh: solar panels get access to eight times more energy than they would 57 00:03:21,060 --> 00:03:23,740 Josh: have in the case of the around space space is also 58 00:03:23,740 --> 00:03:26,720 Josh: a vacuum which means you can radiate heat directly into this 59 00:03:26,720 --> 00:03:29,520 Josh: empty space where there's nothing to worry about there's no 60 00:03:29,520 --> 00:03:33,440 Josh: land there's security and sovereignty because let me tell you it's a lot harder 61 00:03:33,440 --> 00:03:37,380 Josh: to hack a data center in outer space than it is to do so here on earth so there's 62 00:03:37,380 --> 00:03:39,600 Josh: a lot of perks we're going to go through but i want to hand it back to you to 63 00:03:39,600 --> 00:03:45,040 Josh: talk about who's been pro space yes because there's some notable names that 64 00:03:45,040 --> 00:03:47,860 Josh: are very interested in the space race here yeah. 65 00:03:47,860 --> 00:03:51,260 Ejaaz: There's a lot of people um talking about this josh um before 66 00:03:51,260 --> 00:03:54,080 Ejaaz: we get into that i kind of want to like lay out that this 67 00:03:54,080 --> 00:03:56,780 Ejaaz: is a very real problem and challenge because i'm sure a 68 00:03:56,780 --> 00:03:59,640 Ejaaz: lot of people listen to this and thinking well like i haven't heard any news 69 00:03:59,640 --> 00:04:04,740 Ejaaz: about this let me introduce you to some um there are two data centers that got 70 00:04:04,740 --> 00:04:09,180 Ejaaz: permitted to build these data centers from 2019 by the way josh they're worth 71 00:04:09,180 --> 00:04:13,000 Ejaaz: hundreds of millions i think like book value is like 500 million dollars between 72 00:04:13,000 --> 00:04:15,020 Ejaaz: both of them, they're collecting dust. 73 00:04:15,800 --> 00:04:19,820 Ejaaz: There's no lights on in the data centers because they just can't get the energy 74 00:04:19,820 --> 00:04:21,840 Ejaaz: supply to power these, right? 75 00:04:22,020 --> 00:04:25,280 Ejaaz: And I think there's none other than Satya Nadella of Microsoft, 76 00:04:25,320 --> 00:04:30,080 Ejaaz: which can explain succinctly what exactly this problem is. I'm gonna play a little clip right here. 77 00:04:30,240 --> 00:04:34,120 Ejaaz: The biggest issue we are now having is not a compute glut, but it's a power. 78 00:04:34,340 --> 00:04:39,440 Ejaaz: And it's sort of the ability to get the builds done fast enough close to power. 79 00:04:39,770 --> 00:04:41,690 Ejaaz: So if you can't do that, you may 80 00:04:41,690 --> 00:04:45,450 Ejaaz: actually have a bunch of chips sitting in inventory that I can't plug in. 81 00:04:45,650 --> 00:04:49,910 Ejaaz: In fact, that is my problem today, right? It's not a supply issue of chips. 82 00:04:50,090 --> 00:04:53,870 Ejaaz: It's actually the fact that I don't have warm shells to plug into. 83 00:04:54,110 --> 00:04:58,650 Ejaaz: And so how some supply chain constraints emerge, tough to predict, 84 00:04:58,650 --> 00:05:02,510 Ejaaz: because the demand is just going, you know, is tough to predict. 85 00:05:02,730 --> 00:05:05,830 Ejaaz: So he's not also the only one that's been talking about this, 86 00:05:05,930 --> 00:05:11,010 Ejaaz: Josh. We've got the likes of SpaceX employees saying by 2045, 87 00:05:11,570 --> 00:05:15,470 Ejaaz: We're going to need enough compute to the base power of the entire planet to 88 00:05:15,470 --> 00:05:17,430 Ejaaz: power all kind of like data center efforts. 89 00:05:17,770 --> 00:05:21,530 Ejaaz: Elon Musk says that there'll probably be electricity shortages in about two years time. 90 00:05:21,950 --> 00:05:25,450 Ejaaz: Sam Altman says we need to make a breakthrough, whether it's through fusion or something else. 91 00:05:25,670 --> 00:05:28,790 Ejaaz: Mark Zuckerberg says, listen, we would build a bigger data center, 92 00:05:28,850 --> 00:05:30,330 Ejaaz: but we just don't have the energy to supply. 93 00:05:30,590 --> 00:05:34,410 Ejaaz: So this is obviously something that is captivating some of the brightest entrepreneurial 94 00:05:34,410 --> 00:05:39,650 Ejaaz: minds in the world right now when it comes to scaling AI. 95 00:05:40,050 --> 00:05:42,510 Ejaaz: And, you know, we see this at the infrastructure level as well. 96 00:05:42,670 --> 00:05:45,590 Ejaaz: You just heard Satya talk about something called the PowerShell. 97 00:05:45,950 --> 00:05:50,190 Ejaaz: What he's referring to here is not even just the energy supply lines that you 98 00:05:50,190 --> 00:05:52,290 Ejaaz: need to feed electricity into these data centers. 99 00:05:52,450 --> 00:05:54,810 Ejaaz: There's also like the manual labor, Josh. 100 00:05:55,170 --> 00:05:59,370 Ejaaz: There is the technicians, the specialists that you need to kind of like boost 101 00:05:59,370 --> 00:06:03,010 Ejaaz: this entire thing. There's so many missing gaps here that we just don't meet it. 102 00:06:03,130 --> 00:06:07,730 Josh: It's funny. When I was in school, BOCES and like specialty job training was 103 00:06:07,730 --> 00:06:10,910 Josh: very popular. And that's kind of where you learn to become an electrician or a plumber. 104 00:06:11,270 --> 00:06:13,190 Josh: And a lot of these people, they didn't accrue a lot of debt. 105 00:06:13,310 --> 00:06:15,910 Josh: They left school early and they have like well-paying jobs now. 106 00:06:16,070 --> 00:06:20,590 Josh: And it turns out in a world of post-AI where we have AGI and the robots haven't quite made it here. 107 00:06:20,810 --> 00:06:23,190 Josh: The people who we need the most are the electricians, are the plumbers, 108 00:06:23,250 --> 00:06:26,250 Josh: are the people who can do the cooling, the heating, the electricity wiring. 109 00:06:26,510 --> 00:06:28,790 Josh: Those are the most valuable jobs in the world. and the people are getting paid 110 00:06:28,790 --> 00:06:30,850 Josh: a tremendous amount of money. Now, 111 00:06:31,400 --> 00:06:34,820 Josh: Before we get into this next section, I do want to frame a new question that 112 00:06:34,820 --> 00:06:38,760 Josh: we're going to answer a little bit later, but it's the actual narrative of space 113 00:06:38,760 --> 00:06:41,940 Josh: because a lot of people are trying to raise a tremendous amount of money, 114 00:06:42,060 --> 00:06:45,560 Josh: which requires a tremendous amount of optimism from the investors and the public. 115 00:06:45,820 --> 00:06:51,180 Josh: And you have to ask the question, is this a real narrative or is this kind of 116 00:06:51,180 --> 00:06:56,500 Josh: an exponential kind of like moonshot guess, moonshot like kind of play that 117 00:06:56,500 --> 00:06:59,820 Josh: they're targeting to get people interested to raise more money to build their AI? 118 00:07:00,180 --> 00:07:02,780 Josh: I don't want to answer that yet. We're going to answer that as we go. 119 00:07:02,980 --> 00:07:06,080 Josh: But by answering it, we're going to start with the first company who's actively 120 00:07:06,080 --> 00:07:09,120 Josh: doing this, which is a project called StarCloud. 121 00:07:09,540 --> 00:07:12,360 Josh: Now, StarCloud is the first prototype, as far as I'm aware, EJS, 122 00:07:12,460 --> 00:07:14,220 Josh: right, that's doing this. Can you tell me a little bit more about it? 123 00:07:14,880 --> 00:07:20,040 Ejaaz: For sure, yeah. So StarCloud is kind of what triggered this entire debate online, Josh. 124 00:07:20,360 --> 00:07:24,620 Ejaaz: Prior to StarCloud announcing their prototype of launching a GPU in outer space, 125 00:07:25,020 --> 00:07:27,480 Ejaaz: everyone was kind of like oblivious to this entire thing. 126 00:07:27,560 --> 00:07:31,060 Ejaaz: And then Google and Elon Musk chimed in after it. So these are the pioneers. 127 00:07:31,060 --> 00:07:32,880 Ejaaz: And they kind of pioneered it with 128 00:07:32,880 --> 00:07:35,880 Ejaaz: this prototype, which should already be launched in November, actually, 129 00:07:36,160 --> 00:07:40,620 Ejaaz: which is they launched an NVIDIA H100 GPU, which is technically the most powerful 130 00:07:40,620 --> 00:07:46,800 Ejaaz: computer that's ever existed in outer space, into outer space using Elon Musk's SpaceX shuttle. 131 00:07:47,140 --> 00:07:53,900 Ejaaz: And they ran a Google open source model called Gemma. So it's currently live and up there running. 132 00:07:54,120 --> 00:07:58,080 Ejaaz: And the reason why this is such a big deal is it's mainly a testing phase to 133 00:07:58,080 --> 00:08:02,960 Ejaaz: see if GPUs can survive in outer space. It's running on its own. 134 00:08:03,100 --> 00:08:06,820 Ejaaz: It's training on its own. It's being iterated to on its own outside in space 135 00:08:06,820 --> 00:08:09,480 Ejaaz: whilst communicating to their servers on Earth. 136 00:08:09,580 --> 00:08:11,940 Ejaaz: And what they're really trying to figure out here is, number one, 137 00:08:12,060 --> 00:08:15,200 Ejaaz: is it feasible or will this GPU just die from radiation? We're going to get 138 00:08:15,200 --> 00:08:16,580 Ejaaz: into some of the challenges in a second. 139 00:08:17,050 --> 00:08:20,390 Ejaaz: And number two, does it make economical sense? 140 00:08:20,530 --> 00:08:24,110 Ejaaz: Like if we can do this with one GPU, can we scale this to hundreds of thousands 141 00:08:24,110 --> 00:08:27,830 Ejaaz: of GPUs? So they're the real ones, Josh, that started this trend. 142 00:08:28,190 --> 00:08:33,170 Josh: The initial thing that I found funny about StarCloud was the 16 square kilometer heat diffuser. 143 00:08:33,330 --> 00:08:35,970 Josh: Oh, yeah. That was kind of where I got caught up on, where five gigawatts of 144 00:08:35,970 --> 00:08:40,230 Josh: energy means 16 square kilometers of metal that it took to diffuse this. 145 00:08:40,350 --> 00:08:43,090 Josh: Because again, in space, there is no atmosphere. 146 00:08:43,390 --> 00:08:46,890 Josh: There is no place to cool it besides shooting the radiation off into deep space. 147 00:08:47,050 --> 00:08:51,090 Josh: It's not like you can run like cold air over this because there is no air. 148 00:08:51,430 --> 00:08:54,810 Josh: So I think as I've kind of learned more about the StarCloud project, 149 00:08:54,950 --> 00:08:58,090 Josh: I've started to realize like, OK, that's that's that's kind of within reason. 150 00:08:58,250 --> 00:09:02,530 Josh: And that's an early prototype in how they're going to go about actually dissipating heat. 151 00:09:02,810 --> 00:09:06,210 Josh: So six square miles of solar panels, they fold it into a rocket. 152 00:09:06,410 --> 00:09:08,890 Josh: They could send it out on a starship. It's like it's pretty cool. 153 00:09:09,010 --> 00:09:09,910 Josh: It seems like it makes sense. 154 00:09:10,030 --> 00:09:13,710 Josh: Ejaz, is there anything that I'm missing here that technically doesn't make 155 00:09:13,710 --> 00:09:14,910 Josh: a whole lot of sense based on these posts? 156 00:09:15,450 --> 00:09:18,450 Ejaaz: Yeah i mean you nailed a technical challenge number 157 00:09:18,450 --> 00:09:23,130 Ejaaz: one which is like these deployable structures are not only so large for the 158 00:09:23,130 --> 00:09:26,850 Ejaaz: amount of compute that it's supposedly meant to be supporting like these things 159 00:09:26,850 --> 00:09:31,470 Ejaaz: like you said are 16 square kilometers that is huge to fuel do you know how 160 00:09:31,470 --> 00:09:35,470 Ejaaz: much power it fuels actually josh it's like what's it one gigawatt or is it 161 00:09:35,470 --> 00:09:37,870 Ejaaz: less than that like maybe 100 kilowatts well 162 00:09:37,870 --> 00:09:39,850 Josh: It's showing five gigawatts now for the data center yeah. 163 00:09:40,380 --> 00:09:44,420 Ejaaz: Yeah, okay, right. Four by four. Yeah. So that's technical challenge number one. 164 00:09:44,600 --> 00:09:49,020 Ejaaz: And then technical challenge number two down here is the radiation that you 165 00:09:49,020 --> 00:09:52,360 Ejaaz: mentioned, which is basically, you know, you have all these kind of like nuclear 166 00:09:52,360 --> 00:09:53,940 Ejaaz: and atoms flying around, which 167 00:09:53,940 --> 00:09:58,420 Ejaaz: can really mess with the composure and structure of these data centers. 168 00:09:59,100 --> 00:10:01,760 Ejaaz: Radiation is a very real issue in space. That's why you have all these different 169 00:10:01,760 --> 00:10:03,360 Ejaaz: space suits and insulation. 170 00:10:03,600 --> 00:10:07,020 Ejaaz: So if you expose hardware to this and it kind of messes up with any kind of 171 00:10:07,020 --> 00:10:10,520 Ejaaz: the machinery of it's doing its job, then it becomes more and more implausible. 172 00:10:10,660 --> 00:10:14,260 Ejaaz: But these are things that they were kind of at like the frontier testing in 173 00:10:14,260 --> 00:10:17,300 Ejaaz: terms of research. We haven't really found a resolution just yet. 174 00:10:17,660 --> 00:10:20,020 Ejaaz: There are solutions like kind of like cooling that you mentioned, 175 00:10:20,160 --> 00:10:22,240 Ejaaz: but otherwise there's no clear cut solution. 176 00:10:22,520 --> 00:10:25,800 Josh: Yeah. Okay. But there is another company that we're all familiar with that does 177 00:10:25,800 --> 00:10:29,420 Josh: have a seemingly more clear cut solution and a seemingly more clear roadmap 178 00:10:29,420 --> 00:10:34,180 Josh: which um i believe the ceo goes by the name of elon musk if i'm not mistaken 179 00:10:34,180 --> 00:10:38,700 Josh: and yeah elon and spacex have a strategy uh we have a clip here that i think 180 00:10:38,700 --> 00:10:41,320 Josh: you wanted to share maybe we could start there and we'll get into how exactly 181 00:10:41,320 --> 00:10:42,440 Josh: this is all going to work in. 182 00:10:42,440 --> 00:10:46,960 Ejaaz: Order in order to harness a non-trivial amount of the energy of the sun you 183 00:10:46,960 --> 00:10:51,560 Ejaaz: have to move to solar-powered ai satellites in deep space um 184 00:10:53,140 --> 00:11:01,620 Ejaaz: Somewhat is a confluence of Tesla expertise and SpaceX expertise and XAI on the AI front. 185 00:11:02,280 --> 00:11:05,580 Josh: So there's three critical pillars that Elon mentioned in that little summary. 186 00:11:05,780 --> 00:11:07,980 Josh: There's Tesla, SpaceX, and XAI. 187 00:11:08,520 --> 00:11:12,980 Josh: SpaceX is the single most important pillar in that because without the ability 188 00:11:12,980 --> 00:11:16,380 Josh: to get mass to orbit, it is impossible to build these data centers in the first place. 189 00:11:16,580 --> 00:11:20,600 Josh: And the reason why the window is just barely starting to open is because we're 190 00:11:20,600 --> 00:11:24,360 Josh: able to get the cost per kilogram down low enough to make it economically viable. 191 00:11:24,540 --> 00:11:30,140 Josh: Because when we started with the space shuttle, Ejaz, it cost $60,000 per kilogram 192 00:11:30,140 --> 00:11:31,500 Josh: to get something into orbit. 193 00:11:32,380 --> 00:11:37,620 Josh: Nothing is going to make that viable if it costs $60,000 per kilogram because 194 00:11:37,620 --> 00:11:40,600 Josh: these data centers are very, very heavy and so are the solar panels, 195 00:11:40,660 --> 00:11:41,960 Josh: so are the batteries, so all of that. 196 00:11:42,740 --> 00:11:49,380 Josh: Falcon Heavy, which is SpaceX's newer rocket, got that cost down to $1,500 per kilogram. 197 00:11:49,660 --> 00:11:53,120 Josh: So we've already seen like a few orders of magnitude decrease and Starship, 198 00:11:53,260 --> 00:11:57,020 Josh: which we've covered in previous episodes, that is planning to go to Mars and 199 00:11:57,020 --> 00:12:01,160 Josh: beyond, it's still not working, but they're planning to get it fully reusable by next year. 200 00:12:01,280 --> 00:12:06,560 Josh: That brings the cost per kilogram down to $10 per kilogram, which is the window 201 00:12:06,560 --> 00:12:08,600 Josh: that enables this conversation to happen. 202 00:12:08,740 --> 00:12:12,680 Josh: Because in the absence of Starship, we cannot send mass to orbit at scale. 203 00:12:12,980 --> 00:12:16,560 Josh: With Starship working, we can actually talk about sending these data centers into space. 204 00:12:16,620 --> 00:12:20,220 Josh: And that's a hugely powerful unlock that only exists 205 00:12:20,220 --> 00:12:23,080 Josh: because of spacex which is why i think it's important to pay attention to elon when 206 00:12:23,080 --> 00:12:25,840 Josh: he talks about this stuff because he's the one who's enabled this to 207 00:12:25,840 --> 00:12:29,500 Josh: even be a possible conversation um earth gets 208 00:12:29,500 --> 00:12:32,380 Josh: only one to two billionth of 209 00:12:32,380 --> 00:12:35,680 Josh: the sun's energy so if you go a million times earth's 210 00:12:35,680 --> 00:12:39,880 Josh: capacity you must go into space is kind of thinking where in order to scale 211 00:12:39,880 --> 00:12:44,460 Josh: ai to the size that we want in order for true agi for for ai that scales infinitely 212 00:12:44,460 --> 00:12:49,720 Josh: with us we need to go past earth's ability to generate energy and we need to 213 00:12:49,720 --> 00:12:52,540 Josh: go out into the stars and that's kind of the the thesis with elon and the xai thing. 214 00:12:53,410 --> 00:12:56,570 Ejaaz: Got it. I mean, I was just looking at this other post, Josh, 215 00:12:56,770 --> 00:13:01,550 Ejaaz: where, you know, StarCloud is kind of laying out the economics for their kind of projections. 216 00:13:01,890 --> 00:13:04,830 Ejaaz: And they say something pretty similar, which is like, on Earth, 217 00:13:04,930 --> 00:13:09,950 Ejaaz: it costs near like $200 million to kind of run a 40 megawatt data center for 10 years. 218 00:13:10,490 --> 00:13:14,190 Ejaaz: Technically, or theoretically, in orbit, it would be worth $8.2 million, 219 00:13:14,450 --> 00:13:15,750 Ejaaz: which is 20 times cheaper. 220 00:13:15,870 --> 00:13:20,110 Ejaaz: So what I'm hearing from you and the Elon case is that, well, 221 00:13:20,310 --> 00:13:21,950 Ejaaz: number one, he's going to be the toll master. 222 00:13:22,210 --> 00:13:24,570 Ejaaz: He's going to basically own the highway into space. 223 00:13:24,890 --> 00:13:25,710 Josh: He's the Jensen Huang of space. 224 00:13:25,950 --> 00:13:30,290 Ejaaz: He's the Jensen Huang of space. That's perfect. We have to trademark that term. 225 00:13:30,410 --> 00:13:31,450 Ejaaz: He's the Jensen Huang of space. 226 00:13:31,910 --> 00:13:37,010 Ejaaz: That's crazy good. So he's going to not only own that toll, but the other thing, Josh, 227 00:13:37,550 --> 00:13:42,590 Ejaaz: which I know we both love the idea of and which he actually commented on in this tweet, 228 00:13:42,730 --> 00:13:47,350 Ejaaz: he says, simply scaling up Starlink V3 satellites, which are the main satellites 229 00:13:47,350 --> 00:13:52,950 Ejaaz: that he deploys from his SpaceX launchers will have high speed laser links, which would work. 230 00:13:53,110 --> 00:13:58,450 Ejaaz: And what he's referring to here is work in terms of like a interconnected network. 231 00:13:58,710 --> 00:14:03,190 Ejaaz: So these satellites would beam lasers to each other one terabit per second speeds 232 00:14:03,190 --> 00:14:08,490 Ejaaz: to be able to train or inference AI models up at that level. 233 00:14:08,610 --> 00:14:13,230 Ejaaz: So imagine like a satellite network kind of in space orbiting Earth that you 234 00:14:13,230 --> 00:14:19,330 Ejaaz: can use to beam terabits worth of data at super rapid speeds between each of 235 00:14:19,330 --> 00:14:22,910 Ejaaz: the satellites, train your AI model, and then beam it back down to Earth. 236 00:14:23,450 --> 00:14:28,670 Ejaaz: Just sounds so nuts at this point, Josh, that I can't quite wrap my head around. 237 00:14:28,970 --> 00:14:32,170 Ejaaz: And then the third thing that he's talked about, Josh, I don't know if you saw 238 00:14:32,170 --> 00:14:36,670 Ejaaz: this, he's talking about quantum computing on the dark craters of the moon. 239 00:14:36,790 --> 00:14:41,250 Ejaaz: Can you, sorry, can you help me wrap my head around this? What the hell is this? What's this concept? 240 00:14:41,270 --> 00:14:43,370 Josh: Yeah, I'm pulling a number out of thin air. I believe it's right, 241 00:14:43,470 --> 00:14:46,390 Josh: though, where the craters, there are craters in the moon. 242 00:14:46,550 --> 00:14:49,430 Josh: So there is no dark side of the moon, but there are craters that never see the sun. 243 00:14:49,870 --> 00:14:53,310 Josh: Those craters exist at a perpetual negative 200 degrees Celsius. 244 00:14:53,450 --> 00:14:55,910 Josh: They are literally frozen solid. 245 00:14:56,090 --> 00:14:58,170 Josh: There is no heat that gets in or out. 246 00:14:58,370 --> 00:15:04,770 Josh: And if you can place a gigantic data center in one of those craters at negative 247 00:15:04,770 --> 00:15:09,530 Josh: 200 degrees Celsius or whatever the exact amount is, you have a pretty good cooling system. 248 00:15:09,810 --> 00:15:13,490 Josh: Like it's not that bad. And each as you'll often see with quantum computers, 249 00:15:13,650 --> 00:15:17,770 Josh: there's this huge mechanism on top of them. And then there's this tiny little chip at the bottom. 250 00:15:18,130 --> 00:15:22,150 Josh: The entire mechanism on top of it is just to keep it cool. 251 00:15:22,150 --> 00:15:25,430 Josh: The actual chip is is the size of a normal chip but because 252 00:15:25,430 --> 00:15:28,270 Josh: it's so important and so necessary to get 253 00:15:28,270 --> 00:15:32,510 Josh: quantum computing chips at solid zero so that the qubits don't move around too 254 00:15:32,510 --> 00:15:36,050 Josh: much that's why they need it so there is a world in which quantum computing 255 00:15:36,050 --> 00:15:40,730 Josh: makes sense if we can get it into outer space there is some data and energy 256 00:15:40,730 --> 00:15:44,390 Josh: notes that elon has that i wrote down that i want to share briefly because this 257 00:15:44,390 --> 00:15:45,990 Josh: is also fascinating and he said, 258 00:15:46,600 --> 00:15:50,760 Josh: Starship should be able to deliver around 300 gigawatts per year of solar-powered 259 00:15:50,760 --> 00:15:54,080 Josh: AI satellites to orbit, maybe 500 gigawatts. 260 00:15:54,240 --> 00:15:58,940 Josh: So if this is true, within a few years of full-rate Starship launches, 261 00:15:59,500 --> 00:16:05,460 Josh: orbital AI could rival or exceed entire country's electrical consumption. 262 00:16:06,420 --> 00:16:09,260 Josh: So the scale of this is outrageous. And granted, this is all very optimistic. 263 00:16:09,260 --> 00:16:13,300 Josh: This is all very forward-looking, very sci-fi-esque. 264 00:16:13,300 --> 00:16:16,420 Josh: But in the case that is even directionally correct the 265 00:16:16,420 --> 00:16:19,100 Josh: next 10 to 20 years of these launches the next 10 to 266 00:16:19,100 --> 00:16:21,940 Josh: 20 years of compute can scale in 267 00:16:21,940 --> 00:16:25,300 Josh: a really exciting way and we were talking to to luke who helps produce the show 268 00:16:25,300 --> 00:16:29,300 Josh: right before this and he made a great point that said in order to get exponential 269 00:16:29,300 --> 00:16:33,660 Josh: gains in progress you need exponential gains in technology or something along 270 00:16:33,660 --> 00:16:37,400 Josh: those lines and it's so true where if we really want to build ai that changes 271 00:16:37,400 --> 00:16:40,960 Josh: the trajectory of the universe you need to do something novel, something new. 272 00:16:41,100 --> 00:16:44,720 Josh: And doing this all in space and unlocking this new energy source feels like 273 00:16:44,720 --> 00:16:45,720 Josh: the natural progression given 274 00:16:45,720 --> 00:16:48,540 Josh: where we are in terms of technological capability back here on Earth. 275 00:16:48,720 --> 00:16:52,640 Ejaaz: It's also equally fair to say that Elon sometimes gets ahead of himself. 276 00:16:52,820 --> 00:16:57,180 Ejaaz: Like he can see the vision, right? But his projections are way too near term 277 00:16:57,180 --> 00:16:59,260 Ejaaz: to what it actually requires to get to that point. 278 00:16:59,600 --> 00:17:03,480 Ejaaz: And I wish I could critique him, Josh, except that there's also another person, 279 00:17:03,640 --> 00:17:07,920 Ejaaz: which is bull posting about AI data centers in space. 280 00:17:08,040 --> 00:17:13,300 Ejaaz: And it is Sundar from Google, the CEO of Google, who in pretty much every single 281 00:17:13,300 --> 00:17:16,400 Ejaaz: interview, this clip that I'm showing you on the screen right now is from an 282 00:17:16,400 --> 00:17:17,620 Ejaaz: interview he did literally yesterday, 283 00:17:17,840 --> 00:17:21,460 Ejaaz: which has got nothing to do with Google's AI space efforts specifically, 284 00:17:21,620 --> 00:17:25,260 Ejaaz: but he had to shell Google's project called Project Suncatcher, 285 00:17:25,440 --> 00:17:30,380 Ejaaz: where they are also aiming to put Google's TPUs, which is their GPU equivalent, 286 00:17:30,380 --> 00:17:33,420 Ejaaz: into space so that they can harness the power of the sun. 287 00:17:33,980 --> 00:17:36,680 Ejaaz: There are too many important people talking about these things, 288 00:17:36,740 --> 00:17:40,980 Ejaaz: Josh, where like my inner, kind of like half my brain is kind of fighting against 289 00:17:40,980 --> 00:17:42,160 Ejaaz: this, saying this is absolutely ridiculous. 290 00:17:42,280 --> 00:17:47,340 Ejaaz: But the other half, I think, is overwhelmingly now in favor of this thing might actually work. 291 00:17:48,050 --> 00:17:51,010 Josh: Yeah, it's their moonshot. It's another Google moonshot. 292 00:17:51,150 --> 00:17:55,790 Josh: And again, I'm going to tease the question, is this just a narrative thing to 293 00:17:55,790 --> 00:17:57,810 Josh: get people excited to invest more in their companies? 294 00:17:57,830 --> 00:18:00,670 Josh: Or is this real technology? Maybe we'll see. 295 00:18:01,290 --> 00:18:05,050 Josh: He is partnering with Planet, a company named Planet, and their plan is to launch 296 00:18:05,050 --> 00:18:06,130 Josh: two prototype satellites by 297 00:18:06,130 --> 00:18:10,130 Josh: early 2027 to test TPUs and optical links to orbit and see how that works. 298 00:18:10,610 --> 00:18:13,910 Josh: In the previous example with SpaceX, you mentioned that there is these like 299 00:18:13,910 --> 00:18:15,330 Josh: space lasers that can talk to each other. 300 00:18:15,330 --> 00:18:18,170 Josh: A cool thing is that you can communicate at the 301 00:18:18,170 --> 00:18:21,370 Josh: speed of light there is no kind of interference in space 302 00:18:21,370 --> 00:18:24,370 Josh: so the time that it takes for a photon to get from one laser 303 00:18:24,370 --> 00:18:28,990 Josh: to or one satellite to the next is the latency between the two which is very 304 00:18:28,990 --> 00:18:32,990 Josh: very fast so there is a world in which theoretically if you're looking at this 305 00:18:32,990 --> 00:18:36,430 Josh: from first principles based on there are no laws of physics that are being violated 306 00:18:36,430 --> 00:18:41,890 Josh: in order to do this so therefore it's a matter of technical capability technical called prowess, 307 00:18:41,990 --> 00:18:45,330 Josh: and our intellectual ability to actually figure out and solve these problems. 308 00:18:45,990 --> 00:18:50,690 Josh: It's funny seeing Google and Elon kind of converge on this. I assume, 309 00:18:50,890 --> 00:18:53,470 Josh: like you mentioned earlier, a lot of other companies are not far behind. 310 00:18:53,630 --> 00:18:57,170 Josh: They just don't quite have the infrastructure or the ability to tackle a problem 311 00:18:57,170 --> 00:18:58,810 Josh: like this. But it seems very real. 312 00:18:58,950 --> 00:19:03,310 Josh: And they're planning to do this as soon as the next 24 months with the new TPEs 313 00:19:03,310 --> 00:19:05,310 Josh: that we've been talking about so often. Sundar. 314 00:19:05,310 --> 00:19:10,090 Ejaaz: Said that they're already prototyping a bunch of radiation measures on ground 315 00:19:10,090 --> 00:19:13,650 Ejaaz: here at earth to make sure that their tpus can survive so the next natural step 316 00:19:13,650 --> 00:19:17,650 Ejaaz: there is to launch these babies into space probably using elon musk's spacex 317 00:19:17,650 --> 00:19:19,530 Ejaaz: because he is the thomas he's the jensen and 318 00:19:19,530 --> 00:19:22,350 Josh: To that point actually the uh the economics and the launch cost they actually 319 00:19:22,350 --> 00:19:25,730 Josh: we have a quote here that says uh their modeling says if launch costs get under 320 00:19:25,730 --> 00:19:30,730 Josh: 200 per kilogram to orbit by the mid-2030s then space-based compute could be 321 00:19:30,730 --> 00:19:34,430 Josh: comparable to earth data centers on a per kilowatt per year basis so 322 00:19:34,820 --> 00:19:37,800 Josh: If we can get 20 times more than what Elon's projecting. 323 00:19:38,180 --> 00:19:41,360 Josh: So if the cost is pretty high. So if it's 20 times more expensive than what 324 00:19:41,360 --> 00:19:45,220 Josh: Elon's projecting in a decade from now, then we'll have parity with Earth. 325 00:19:45,380 --> 00:19:49,320 Josh: So it gives you a testament to like the timelines of this. And maybe this is 326 00:19:49,320 --> 00:19:51,820 Josh: a good time to go into the pros and cons of what. 327 00:19:51,820 --> 00:19:53,080 Ejaaz: Is good about this, 328 00:19:53,200 --> 00:19:55,940 Josh: What is bad about this. And maybe humble people a little bit, 329 00:19:56,020 --> 00:20:00,840 Josh: Ejaz, if you want to start with the cons as to why this isn't as rosy as we 330 00:20:00,840 --> 00:20:04,800 Josh: think it might be. like maybe maybe bring us back down to earth in fact. 331 00:20:04,800 --> 00:20:10,180 Ejaaz: I i that's that's where i did did not expect that part josh um okay right so 332 00:20:10,180 --> 00:20:13,540 Ejaaz: i'm going to start off with the bear case here which you've mentioned a few 333 00:20:13,540 --> 00:20:18,560 Ejaaz: times which is this is just a narrative play by some of the biggest companies 334 00:20:18,560 --> 00:20:22,680 Ejaaz: to boost their valuations and let me explain the problem to you right 335 00:20:23,150 --> 00:20:29,750 Ejaaz: Up until today, speaking about these data centers in space, the focus has been on GPUs. 336 00:20:29,810 --> 00:20:32,830 Ejaaz: GPUs is the new gold, right? It's what you need to kind of train your AI. 337 00:20:33,170 --> 00:20:37,210 Ejaaz: And then we found out that they don't have enough energy to kind of like fuel these things. 338 00:20:37,510 --> 00:20:40,610 Ejaaz: Now, the reason why this is so important, as this post explains, 339 00:20:40,850 --> 00:20:45,170 Ejaaz: is a lot of these companies, Sundar's company, Google, Elon Musk's companies, 340 00:20:45,350 --> 00:20:49,190 Ejaaz: are valued based on the amount of GPUs they've been purchasing. 341 00:20:49,190 --> 00:20:53,610 Ejaaz: And the Wall Street analysts have basically assumed, okay, the amount of GPUs 342 00:20:53,610 --> 00:20:58,890 Ejaaz: that this company buys is equal to the economic output. I can do a simple formula, 343 00:20:59,010 --> 00:21:01,390 Ejaaz: right? If they're buying these GPUs, they're going live. 344 00:21:01,550 --> 00:21:05,050 Ejaaz: What they haven't factored in is that these GPUs aren't going live. 345 00:21:05,170 --> 00:21:06,650 Ejaaz: They're just sitting collecting dust. 346 00:21:06,810 --> 00:21:10,470 Ejaaz: And so that's less ROI per quarter than they initially projected, 347 00:21:10,670 --> 00:21:13,670 Ejaaz: which means that the valuations of these companies should probably be lowered. 348 00:21:13,850 --> 00:21:16,350 Ejaaz: So what's the countermeasure to this, Josh? 349 00:21:16,890 --> 00:21:19,810 Ejaaz: Well, obviously it's infinite energy. obviously it's what was 350 00:21:19,810 --> 00:21:22,970 Ejaaz: that 99.98% of the solar system is the 351 00:21:22,970 --> 00:21:26,250 Ejaaz: sun it's all that energy so if you can miraculously wow 352 00:21:26,250 --> 00:21:29,170 Ejaaz: now figure out a way to harness that energy you can 353 00:21:29,170 --> 00:21:32,390 Ejaaz: now maintain that valuation dare I say boost the 354 00:21:32,390 --> 00:21:36,170 Ejaaz: valuations right share prices go up after you you hear Sundar talking about 355 00:21:36,170 --> 00:21:40,470 Ejaaz: this uh google's on an absolute run so that's my skeptics take I think that 356 00:21:40,470 --> 00:21:43,610 Ejaaz: they might be kind of like feeling this narrative which they know is a decade 357 00:21:43,610 --> 00:21:48,850 Ejaaz: out from any kind of real uh prototype being out there that scales um so that's 358 00:21:48,850 --> 00:21:51,810 Ejaaz: my back is do you agree do you disagree with this yeah 359 00:21:51,810 --> 00:21:54,650 Josh: I i think setting expectations on timing is probably 360 00:21:54,650 --> 00:21:59,890 Josh: going to be fairly important because again like satya's saying mid 2030s i mean 361 00:21:59,890 --> 00:22:03,930 Josh: elon's probably saying tomorrow but the reality is it's probably mid 2030s it's 362 00:22:03,930 --> 00:22:06,910 Josh: going to take a while to get all this stuff to orbit and do it at scale that's 363 00:22:06,910 --> 00:22:11,130 Josh: comparable to the united states now when i think about that i think about what 364 00:22:11,130 --> 00:22:13,050 Josh: does the world of ai look like, 365 00:22:13,710 --> 00:22:18,530 Josh: In the mid 2030s. Like, what are we? What's it going to look like here? 366 00:22:18,610 --> 00:22:20,150 Josh: Like, we are moving so fast. 367 00:22:20,370 --> 00:22:25,390 Josh: And to project 10 years into the future, my God, I don't know what the world looks like. 368 00:22:25,570 --> 00:22:28,110 Ejaaz: So our energy bill is going to be 5x, Josh. 369 00:22:28,870 --> 00:22:31,990 Josh: Or maybe we have people like Isaiah Taylor, who was on the show last week, 370 00:22:32,350 --> 00:22:34,830 Josh: solving nuclear reactors right here on home. 371 00:22:34,910 --> 00:22:40,230 Josh: And we actually have energy abundance. That's enough for our midterm goals prior 372 00:22:40,230 --> 00:22:44,450 Josh: to extending out to the long term goals. Maybe this is a fun time to talk briefly 373 00:22:44,450 --> 00:22:47,990 Josh: about the Kardashev scale, which is a sci-fi thing, and there's three tiers to it. 374 00:22:48,050 --> 00:22:51,930 Josh: The first one being you can harness the entire energy of your planet, 375 00:22:51,930 --> 00:22:55,710 Josh: and then it's the energy of your sun, and then the energy of your solar system. 376 00:22:56,170 --> 00:23:01,270 Josh: We have been able to collect perhaps a very small fraction of Kardashev level 377 00:23:01,270 --> 00:23:03,850 Josh: one, which is of the planet, the energy that hits our planet. 378 00:23:04,370 --> 00:23:07,650 Josh: So there's a very long way to go, but I think this is kind of like the grand 379 00:23:07,650 --> 00:23:10,190 Josh: sci-fi vision that gets people excited, gets people motivated. 380 00:23:10,190 --> 00:23:15,230 Josh: If I'm going to work at a company like SpaceX, I'm stoked at the idea of building 381 00:23:15,230 --> 00:23:18,370 Josh: super intelligence in outer space. That's a cool mission to have. 382 00:23:18,630 --> 00:23:21,610 Josh: And like so many of the other missions, it's not physics constrained. 383 00:23:21,770 --> 00:23:25,270 Josh: It is possible. You can do this thing if you can figure out how to create the 384 00:23:25,270 --> 00:23:27,870 Josh: technology to do it. So I love the ambition. 385 00:23:28,070 --> 00:23:30,230 Josh: Is it a narrative play to get more money? 386 00:23:32,030 --> 00:23:37,430 Josh: Maybe, probably. But in the case of Google and Tesla and SpaceX, 387 00:23:37,610 --> 00:23:40,990 Josh: it doesn't seem that way. hey, if this was coming from someone like OpenAI or 388 00:23:40,990 --> 00:23:44,350 Josh: Perplexity or Anthropic, I would feel a little bit differently about it. 389 00:23:44,410 --> 00:23:47,190 Josh: But Google's always had a moonshot division since the beginning of time. 390 00:23:47,530 --> 00:23:50,370 Josh: I mean, that's the reason why AI exists in the current state it does today is 391 00:23:50,370 --> 00:23:52,730 Josh: because they spend so many resources on research and development. 392 00:23:53,510 --> 00:23:58,450 Josh: If you remember from our episode yesterday, Ejaz, they developed the first neural network 15 years ago. 393 00:23:58,610 --> 00:24:01,350 Josh: They did. So it's been under development for a long time. And this very much 394 00:24:01,350 --> 00:24:05,050 Josh: feels like a similar project where these timelines are going to be 10 to 15 395 00:24:05,050 --> 00:24:06,730 Josh: years. But the research starts now. 396 00:24:06,970 --> 00:24:10,430 Josh: And it's very important research because directionally it feels like this is correct. 397 00:24:10,790 --> 00:24:13,390 Josh: Directionally, we're going to need more energy than we can have on Earth, 398 00:24:13,410 --> 00:24:17,570 Josh: even if we solve nuclear power, because of the hope that it's going to require 399 00:24:17,570 --> 00:24:22,030 Josh: so much energy to power superintelligence that will be smarter than the smartest thing on Earth. 400 00:24:22,030 --> 00:24:26,350 Josh: And it just it's a fun science experiment maybe it's a narrative violation because 401 00:24:26,350 --> 00:24:31,650 Josh: they're getting a little ahead of the gun but you know it could it could be and i like this future. 402 00:24:31,650 --> 00:24:34,510 Ejaaz: So do i i think my optimistic take on 403 00:24:34,510 --> 00:24:39,650 Ejaaz: this is all of these things um nuclear fission nuclear power coming into like 404 00:24:39,650 --> 00:24:44,910 Ejaaz: a scalable means renewable energy and you know harnessing energy from the sun 405 00:24:44,910 --> 00:24:50,670 Ejaaz: all comes collectively at a time in the mid 2030s where our energy bills are 406 00:24:50,670 --> 00:24:54,050 Ejaaz: going through the roof and we cannot afford anything else. 407 00:24:54,190 --> 00:24:57,410 Ejaaz: We need to rely on another energy source. And I think it's all going to neatly 408 00:24:57,410 --> 00:24:58,410 Ejaaz: come at the right time, Josh. 409 00:24:58,570 --> 00:25:01,110 Ejaaz: I don't know what makes me think that this might actually happen, 410 00:25:01,230 --> 00:25:02,970 Ejaaz: but I think we are well on our way to do that. 411 00:25:03,090 --> 00:25:05,970 Ejaaz: But to your point, there's a bunch of challenges, right? 412 00:25:06,030 --> 00:25:09,090 Ejaaz: So we've mentioned kind of like some engineering challenges around protecting 413 00:25:09,090 --> 00:25:12,470 Ejaaz: against radiation and building and deploying these infrastructures out in space. 414 00:25:12,570 --> 00:25:13,770 Ejaaz: Right now it's too expensive, right? 415 00:25:13,970 --> 00:25:19,310 Ejaaz: But the idea is presumably using SpaceX's, I'll call it the tunnel to space 416 00:25:19,310 --> 00:25:23,190 Ejaaz: or the highway to space, it'll eventually reduce the cost down by, 417 00:25:23,190 --> 00:25:25,830 Ejaaz: you know, maximum weight. So we're able to kind of achieve this. 418 00:25:26,150 --> 00:25:29,850 Ejaaz: The other thing is kind of like maintenance, right? So it's like a lot of these 419 00:25:29,850 --> 00:25:30,950 Ejaaz: things require a lot of attention. 420 00:25:31,130 --> 00:25:34,410 Ejaaz: Like we mentioned earlier that, you know, supplying energy is just one thing. 421 00:25:34,550 --> 00:25:38,750 Ejaaz: You need to have technicians and experts on there that can like kind of connect 422 00:25:38,750 --> 00:25:41,590 Ejaaz: the right plugs between satellites and make sure that they speak to each other 423 00:25:41,590 --> 00:25:42,750 Ejaaz: or GPUs between each other. 424 00:25:43,070 --> 00:25:45,770 Ejaaz: Are we going to have astronauts roaming around these data centers on space? 425 00:25:45,970 --> 00:25:49,030 Ejaaz: That is kind of harder to wrap my head around because it just seems infeasible. 426 00:25:49,110 --> 00:25:51,430 Ejaaz: We don't do that right now. why would it be in 10 years time? 427 00:25:51,590 --> 00:25:54,410 Ejaaz: Maybe it will be. I'm being optimistic. I have to maintain optimism here, Josh. 428 00:25:54,550 --> 00:25:56,690 Josh: And hey, maybe robots. Robots in 10 years, probably gonna be pretty good. 429 00:25:57,300 --> 00:26:01,100 Ejaaz: Oh, Josh, I need to buy more Tesla. I need to buy more Tesla. 430 00:26:01,300 --> 00:26:02,380 Ejaaz: You've just, the humanoids. 431 00:26:02,700 --> 00:26:05,360 Ejaaz: Oh God, we are getting way too bullish on the con section. 432 00:26:06,460 --> 00:26:08,720 Ejaaz: Okay, hang on, wait, hang on. We got to remain tethered to Earth. 433 00:26:09,340 --> 00:26:14,020 Ejaaz: Speaking of tethering to Earth, actually, how do we get this data and compute 434 00:26:14,020 --> 00:26:17,140 Ejaaz: that we generate in space down back to Earth? 435 00:26:17,360 --> 00:26:19,420 Ejaaz: Presumably the majority of the humans are still going to be on Earth, 436 00:26:19,520 --> 00:26:23,280 Ejaaz: right, Josh? So a question that's in my mind is like, what is the receiving 437 00:26:23,280 --> 00:26:26,320 Ejaaz: infrastructure of this look like? And is that economically feasible? 438 00:26:27,020 --> 00:26:30,180 Ejaaz: Elon seems to claim so in that thread that we showed earlier. 439 00:26:30,360 --> 00:26:33,280 Ejaaz: And Sundar has said multiple times that, yeah, we can just beam data down. 440 00:26:33,420 --> 00:26:35,180 Ejaaz: We can just beam some of the learnings, the compute down. 441 00:26:35,340 --> 00:26:39,120 Ejaaz: I don't know if anyone's actually proven that. And it's because we need to do 442 00:26:39,120 --> 00:26:41,200 Ejaaz: step one first, which is launch the data center. 443 00:26:41,380 --> 00:26:46,100 Ejaaz: So there's so much that this relies on that needs to be kind of solved via frontier 444 00:26:46,100 --> 00:26:47,600 Ejaaz: research that we're kind of 445 00:26:47,600 --> 00:26:51,340 Ejaaz: hoping for, but we haven't got any definitive proof, if that makes sense. 446 00:26:51,420 --> 00:26:55,660 Ejaaz: So if you're investing in this stuff, it's definitely a longer horizon type thing. 447 00:26:56,320 --> 00:26:59,760 Josh: Yeah, conceptually, it all makes sense. We're learning a lot from Starlink, 448 00:26:59,860 --> 00:27:02,020 Josh: actually, in terms of latency and sending things back to Earth, 449 00:27:02,280 --> 00:27:06,020 Josh: where if you're sending data that's timely, it's a little difficult because 450 00:27:06,020 --> 00:27:09,360 Josh: there are tens to hundreds of milliseconds that are added in the latency between the two. 451 00:27:09,580 --> 00:27:13,560 Josh: But if you're batch training, if you're just training mega amounts of AI and 452 00:27:13,560 --> 00:27:16,240 Josh: then you're beaming them in batches down to Earth, there's enough bandwidth, 453 00:27:16,240 --> 00:27:19,960 Josh: if you disregard latency, to send quite a bit of data pretty quickly. 454 00:27:20,160 --> 00:27:23,840 Josh: So I suspect there's some healthy middle ground there where they'll talk to 455 00:27:23,840 --> 00:27:25,860 Josh: each other very fast and they'll beam data down kind of slow. 456 00:27:25,860 --> 00:27:30,180 Josh: And that will be okay um there are pros there's lots of pros we talked about 457 00:27:30,180 --> 00:27:33,480 Josh: a few energy abundance being one cooling being one where you can just create 458 00:27:33,480 --> 00:27:36,440 Josh: mega radiators you could send things on the dark side of the moon put them in 459 00:27:36,440 --> 00:27:40,920 Josh: craters where it's really cold um the environmental thing there's you don't 460 00:27:40,920 --> 00:27:44,600 Josh: need permits or you don't use water or you don't need energy we're just, 461 00:27:45,380 --> 00:27:49,380 Josh: unobstructed by ai which is kind of an interesting thing because currently ai 462 00:27:49,380 --> 00:27:54,520 Josh: is very intrusive um it uses a tremendous amount of data and resources and energy and, 463 00:27:55,180 --> 00:28:00,420 Josh: human labor, which maybe is a good thing to some extent, but that's another thing. 464 00:28:00,600 --> 00:28:03,380 Josh: And then we have just like the sovereignty and resilience point where when stuff 465 00:28:03,380 --> 00:28:06,440 Josh: is in space, it is pretty anti-fragile. No one's really going up there to mess 466 00:28:06,440 --> 00:28:07,920 Josh: with it. It is very secure. 467 00:28:08,160 --> 00:28:10,800 Josh: You kind of know what you're getting when it's up there. You set it, 468 00:28:10,880 --> 00:28:14,460 Josh: you forget it. It's a nice thing. So that's kind of how I think we would weigh the pros and cons. 469 00:28:14,820 --> 00:28:19,020 Ejaaz: Okay. So if those are the pros and those are the cons, the next natural question for me, Josh, 470 00:28:19,260 --> 00:28:22,920 Ejaaz: is like, if I was an analyst looking at this over the next say like five to 471 00:28:22,920 --> 00:28:26,960 Ejaaz: ten years yeah what kind of milestones should i be looking for like do you have 472 00:28:26,960 --> 00:28:31,600 Ejaaz: any idea of what or like what's like what's sensor i guess is it like an entire 473 00:28:31,600 --> 00:28:34,780 Ejaaz: data center out there or is it a couple of gpus what is it yeah 474 00:28:34,780 --> 00:28:37,960 Josh: So because there are a few companies that are actively working on this problem 475 00:28:37,960 --> 00:28:41,300 Josh: we can just kind of evaluate their progress as i think a way of tracking things 476 00:28:41,300 --> 00:28:46,000 Josh: so we could set a few milestones starting with the suncatcher prototypes who 477 00:28:46,000 --> 00:28:48,020 Josh: are planned to go up in 2027. 478 00:28:48,280 --> 00:28:51,540 Josh: So if you're trying to map out trajectories, we have Project Suncatcher. 479 00:28:51,740 --> 00:28:56,800 Josh: We could ask the question, did the 2027 Suncatcher prototypes launch on time? 480 00:28:57,000 --> 00:29:00,160 Josh: Did they actually even make it to outer space? Is Starship working? 481 00:29:00,380 --> 00:29:05,600 Josh: Are they able to carry the correct payloads? Is the payload even able to get into outer space? 482 00:29:05,840 --> 00:29:08,280 Josh: Then once it's in outer space, we can ask the question, well, 483 00:29:08,440 --> 00:29:10,880 Josh: does it actually work? Is it able to shield itself from radiation? 484 00:29:11,140 --> 00:29:14,440 Josh: Is it able to power itself on? Is it able to cool itself down? 485 00:29:14,900 --> 00:29:19,280 Josh: Do they show working optical links and TPUs in orbit? Can they actually make that work? 486 00:29:19,440 --> 00:29:23,620 Josh: So we have the prototypes of Suncatcher. Then we want to look at Starship cadence. 487 00:29:23,860 --> 00:29:27,020 Josh: Is Starship actually getting the cost to orbit low enough to make this economically 488 00:29:27,020 --> 00:29:29,680 Josh: viable? That, I think, actually is the single most important thing to look out 489 00:29:29,680 --> 00:29:31,600 Josh: for, is the Starship program. 490 00:29:31,800 --> 00:29:34,560 Josh: If you are not watching every Starship launch with us on this show, 491 00:29:34,580 --> 00:29:36,900 Josh: you are missing out, because we're going to cover every single one. 492 00:29:37,020 --> 00:29:39,920 Josh: It's the most important thing to the progress of this project. 493 00:29:39,920 --> 00:29:44,660 Josh: And we are going to closely monitor how the cost per kilogram to orbit is going. 494 00:29:44,900 --> 00:29:47,060 Josh: As long as that number is going down, as long as Starship's 495 00:29:47,400 --> 00:29:52,100 Josh: start blowing up less, that's going to be a huge win. And that's going to make it at least plausible. 496 00:29:52,380 --> 00:29:55,040 Josh: And then the third thing I guess to monitor is the power crisis on Earth. 497 00:29:55,160 --> 00:30:00,420 Josh: We want to see how much power we're able to generate outside of these moonshot ideas. 498 00:30:00,580 --> 00:30:04,480 Josh: So are people like Isaiah Taylor of Valor able to create modular nuclear reactors 499 00:30:04,480 --> 00:30:07,360 Josh: to power gigawatts of these data centers and eventually terawatts? 500 00:30:07,460 --> 00:30:10,940 Josh: Are we able to use natural gas turbines and other forms of renewable energy 501 00:30:10,940 --> 00:30:13,680 Josh: to not only power data centers, but our day to day lives? 502 00:30:13,820 --> 00:30:16,520 Josh: Can we do so without making the electricity bills run too high. 503 00:30:16,740 --> 00:30:20,000 Josh: So I think the convergence of these three things are things to look out for 504 00:30:20,000 --> 00:30:23,920 Josh: that we can monitor, that will probably give us a better idea of the roadmap, 505 00:30:24,180 --> 00:30:25,320 Josh: how we're going in the trajectory. 506 00:30:25,480 --> 00:30:27,920 Josh: And again, 10 years from now, who knows what the world looks like. 507 00:30:28,140 --> 00:30:29,920 Josh: But that's what I'm going to be looking out for, at least. 508 00:30:30,360 --> 00:30:33,920 Ejaaz: Yeah, I think that last point is actually one of the most important things, 509 00:30:34,020 --> 00:30:35,720 Ejaaz: because maybe this is a hot take. 510 00:30:35,880 --> 00:30:42,580 Ejaaz: But I think the AI data centers in space hype dies down if things like nuclear 511 00:30:42,580 --> 00:30:47,100 Ejaaz: and renewables can scale on Earth. because people aren't going to focus on things 512 00:30:47,100 --> 00:30:49,980 Ejaaz: in outer space if they can just kind of fix the thing on problem on Earth. 513 00:30:50,040 --> 00:30:55,000 Ejaaz: I think this is more of a philosophical semantic thing due to human nature versus 514 00:30:55,000 --> 00:30:58,260 Ejaaz: kind of doing something in the sky where they haven't spent that much time out there, right? 515 00:30:58,460 --> 00:31:01,640 Ejaaz: And then the other thing is like, to your point, we've got to reduce the cost 516 00:31:01,640 --> 00:31:05,140 Ejaaz: of space travel, which I fully believe Elon's going to do, but he has to prove 517 00:31:05,140 --> 00:31:08,960 Ejaaz: it before people get really bullish and confident that this can be a thing. 518 00:31:09,440 --> 00:31:13,320 Ejaaz: But that's it. Those are the pros and cons. And that is the thesis and the pick, 519 00:31:13,320 --> 00:31:17,920 Ejaaz: the for and against for AI data centers in outer space 520 00:31:18,430 --> 00:31:22,390 Ejaaz: It's so funny seeing our evolution and reaction to this, Josh, 521 00:31:22,770 --> 00:31:26,550 Ejaaz: given that six weeks ago, we were absolutely laughing at this when StarCloud 522 00:31:26,550 --> 00:31:29,510 Ejaaz: announced that they were launching an H100 GPU in space. 523 00:31:29,710 --> 00:31:34,290 Ejaaz: And now here we are opining about, you know, Elon being the toll master, 524 00:31:34,770 --> 00:31:38,090 Ejaaz: it being cheap enough to launch data centers in space, and then figuring out 525 00:31:38,090 --> 00:31:40,490 Ejaaz: how to beam down terabits of data to Earth. 526 00:31:40,710 --> 00:31:43,190 Ejaaz: Just an insane thing that I can't quite wrap my head around. 527 00:31:43,430 --> 00:31:44,950 Ejaaz: But it's been a super fun episode. 528 00:31:45,750 --> 00:31:51,250 Ejaaz: And just like the trajectory of putting data centers in outer orbit, 529 00:31:51,450 --> 00:31:55,430 Ejaaz: there's something else that is reaching the levels and that is exiting the stratosphere. 530 00:31:55,810 --> 00:32:01,590 Ejaaz: Josh, it is the limitless follow-up page. It is our social pages. It is our YouTube pages. 531 00:32:01,770 --> 00:32:07,990 Ejaaz: We have gained, I think, what was it? 3,000 subscribers over the last couple of... More? Tell me. 532 00:32:08,450 --> 00:32:13,910 Josh: Over 4K. And that's not enough because 80% of those people are still not subscribed 533 00:32:13,910 --> 00:32:15,130 Josh: who are listening to the episode. 534 00:32:15,490 --> 00:32:18,950 Josh: So if you're listening on YouTube, please make sure not only click the subscribe 535 00:32:18,950 --> 00:32:21,470 Josh: button, but there's a little bell next to it, which means you could turn on 536 00:32:21,470 --> 00:32:23,690 Josh: notifications to get notified whenever a new episode comes out. 537 00:32:24,010 --> 00:32:27,730 Josh: Happens three times a week. They're all pretty good. They normally range from 25 to 35 minutes. 538 00:32:27,970 --> 00:32:30,290 Josh: And if you enjoy them with your friends, you should share them along because 539 00:32:30,290 --> 00:32:33,150 Josh: it really makes a big difference to the growth of this podcast. 540 00:32:33,250 --> 00:32:34,230 Josh: And we're trying to grow. 541 00:32:34,370 --> 00:32:36,590 Josh: We're trying to make this the biggest thing ever. 542 00:32:36,810 --> 00:32:40,350 Josh: We've been stalling out a little on the audio front. So if you do prefer to 543 00:32:40,350 --> 00:32:41,590 Josh: listen to it as a podcast, 544 00:32:41,830 --> 00:32:45,370 Josh: perhaps go find it on your favorite podcast player, like apple podcast or pocket 545 00:32:45,370 --> 00:32:49,370 Josh: cast is my preferred one those are good places to to find out and hang out with 546 00:32:49,370 --> 00:32:52,730 Josh: us but yeah i think that's mostly it jazz do you have any parting words because 547 00:32:52,730 --> 00:32:55,210 Josh: i actually i have one parting story so you go first. 548 00:32:56,030 --> 00:33:01,130 Ejaaz: Okay. One final thing is we also have a newsletter and we have about 70,000 549 00:33:01,130 --> 00:33:03,210 Ejaaz: of you tuning in every single week. 550 00:33:03,610 --> 00:33:07,930 Ejaaz: Either Josh and I write an essay or a thesis on what we think the future of AI is going to look like. 551 00:33:08,170 --> 00:33:10,990 Ejaaz: Our last one is a juicy bull case on Anthropik. You'll go check it out. 552 00:33:11,190 --> 00:33:14,890 Ejaaz: And we give you the five highlights of the week. So we are anywhere and everywhere 553 00:33:14,890 --> 00:33:18,470 Ejaaz: that you could possibly ingest information about AI and frontier tech. 554 00:33:18,790 --> 00:33:22,430 Ejaaz: Follow, subscribe, give us your email list. Josh, what is your parting words? 555 00:33:22,610 --> 00:33:25,410 Josh: Yeah, there's actually a lot of alpha in there. Google, Ejaz, 556 00:33:25,530 --> 00:33:28,470 Josh: you wrote the Google Bowl case like four or five weeks ago, and the price has 557 00:33:28,470 --> 00:33:29,750 Josh: gone off like 20% since then. 558 00:33:29,850 --> 00:33:33,750 Josh: So if you're not listening to or reading the newsletter, I would advise. 559 00:33:33,950 --> 00:33:37,110 Josh: I wanted to leave with one little bit of sci-fi because when I do episodes like 560 00:33:37,110 --> 00:33:37,730 Josh: this, I'm always reminded... 561 00:33:38,360 --> 00:33:41,460 Josh: Of sci-fi. Like when I look at the future, I'm always reminded of sci-fi books 562 00:33:41,460 --> 00:33:42,440 Josh: and a lot of them get it wrong. 563 00:33:42,460 --> 00:33:46,220 Josh: But I think you could almost always find a sci-fi story that will map to the 564 00:33:46,220 --> 00:33:48,880 Josh: reality that we're having currently. And you kind of see this with Black Mirror. 565 00:33:49,180 --> 00:33:52,420 Josh: One of them that I would encourage everyone to read, it's called The Last Question by Isaac Asimov. 566 00:33:52,660 --> 00:33:56,760 Josh: And it's where I see this going as we move AI into space, where he kind of like 567 00:33:56,760 --> 00:33:59,100 Josh: tracks the progress of humanity over millions of years. 568 00:33:59,260 --> 00:34:02,300 Josh: And each time there is a smarter and smarter and smarter AI. 569 00:34:02,520 --> 00:34:05,920 Josh: And it is an all knowing form of intelligence. And every time it gets smarter, 570 00:34:06,100 --> 00:34:08,880 Josh: because they've been able to capture more and more energy and eventually it leads 571 00:34:08,880 --> 00:34:11,700 Josh: to the last question which is the last question humanity will 572 00:34:11,700 --> 00:34:14,880 Josh: ever need to answer um which i will leave that for you guys 573 00:34:14,880 --> 00:34:18,140 Josh: to explore and to uncover for yourself but it 574 00:34:18,140 --> 00:34:20,860 Josh: is like it's so exciting that we get to live 575 00:34:20,860 --> 00:34:23,820 Josh: in this world that feels like a sci-fi novel where 576 00:34:23,820 --> 00:34:27,000 Josh: we're actually talking about sending ai into outer 577 00:34:27,000 --> 00:34:30,260 Josh: space building artificial general intelligence doing so using space 578 00:34:30,260 --> 00:34:33,320 Josh: rockets and satellites it's like it's really cool and it's 579 00:34:33,320 --> 00:34:36,000 Josh: not here today but there is a very clear trajectory to getting there in the 580 00:34:36,000 --> 00:34:39,500 Josh: future so i think that's mostly what we are excited 581 00:34:39,500 --> 00:34:42,420 Josh: about covering and i think the people who have been here with us 582 00:34:42,420 --> 00:34:45,400 Josh: long enough understand that and they're excited to be on the journey with us so for 583 00:34:45,400 --> 00:34:48,260 Josh: all of you who have made it this far through our fairly long episode today thank 584 00:34:48,260 --> 00:34:52,500 Josh: you i hope you enjoyed the space race um our agi in space episode i hope you 585 00:34:52,500 --> 00:34:57,240 Josh: left informed or excited optimistic or just downright frustrated that we are 586 00:34:57,240 --> 00:35:01,140 Josh: missing the mark so much and ai in space is stupid but whichever one it is let 587 00:35:01,140 --> 00:35:05,220 Josh: us know in the comments down below and we will see you guys in the next one thank you so much for.