Amazon’s $35B AGI Ultimatum to OpenAI & Anthropic Drops AI Safety | EP #235

[00:00] Amazon, uh makes a contingent offer to put $35 billion into Open AI based upon them first off going public and secondly achieving AGI. >> It's kind of incredible that we financialized uh superintelligence, which is amazing. The Open AI to Microsoft definition of AGI was something like generating a hundred billion dollars in either earnings or revenue. I I forget. >> We're measuring compute in terms of gigawatts and AGI in terms of dollars. I love it. Amazon was Anthropic for a
[00:31] while. Now they're Open AI. At some point the circular economy becomes indistinguishable from the real economy and I I think that's what we're seeing here. >> This is the entrepreneurial opportunity of a lifetime. We're talking about tens of thousands of times more capacity to create more money, more value created. Abundance is going to be absolutely rampant. Now that's a moonshot, ladies and gentlemen. Everybody, welcome to Moonshots, the
[01:01] episode of WTF just happened in tech, the number one podcast in AI and exponential technology, getting you ready for the future, getting you ready for the supersonic tsunami heading our way. I'm here with my extraordinary moonshot mates, Salim Ismail, DB2, AWG. Gentlemen, uh another week we've gotten to a cadence of two of these per week. Uh it's and it feels like we're always leaving so many stories on the table. But uh let's do our best. Uh Yes, we
[01:33] need to we need to actually move faster and faster and faster, just like the singularity itself, to keep up with everything broadcasting live. >> no no tech company waits. No GPU waits. All right, let's jump in. Our top AI news stories, Anthropic, Google, Open AI, Uber, uh accelerating at an extraordinary speed of change. Our first story for today, Anthropic revises responsible scaling policy amid increased competition. This was a story I put to the top of the
[02:03] conversation cuz it's it's a very significant um and you know, we I had Jared Kaplan on stage at the Abundance Summit uh last year the year before. Alex, you know Jared well. I think he was a uh a roommate. >> was he was he was a year behind me in the Harvard physics graduate program. Yeah. What an amazing What an amazing group of friends that you had. Uh but here's the deal. They're dropping their 2023 pledge to not train advanced AI unless safety is guaranteed. Uh
[02:34] and Jared's point, I think logically, is if everyone else is rushing ahead, uh then us, you know, sort of hampering ourselves doesn't make any sense. Uh and I want to discuss this because uh it's concerning, you know, a lot of us looked at uh at Anthropic as the most responsible party out there. Them and Google. Uh thoughts, gents? Well, you know, >> fails in exponential races, right? >> There's lots of thoughts all of you at once. This is a metaphor for something,
[03:05] right? We're going to race to talk about a race condition. Love it. Oh my god. Amazing. I want to I want to open with Saleem here. Saleem, go ahead. >> Okay. Well, I mean, safety typically fails in exponential races. The You could look at the the whole thing writ large as OpenAI cracked open and let Pandora's box out. And so, this is just the same type of dynamic occurring again. It speaks to the idea that technology is going to move at its pace and we have to move our human structures at that pace. We can't fall behind. Yeah. Dave?
[03:36] Yeah, no, it's definitely history repeating itself. And so many of our MIT classmates went to Google uh back in '04, '05, '06 when it was don't be evil. Yeah. And they went there over Microsoft cuz everyone perceived Microsoft as being evil. And Google was going to be the force of good in all of tech, and then you know, they bought YouTube, and then they built Chrome, and then they you know, what what they promised the engineers early on, the ones that I knew anyway, is look, we will never store somebody's search history. Mhm. How laughable is
[04:06] that in hindsight? So, then they expanded out of search history. Oh, we're going to store that for 5 years. Uh but we're also going to launch Chrome. Now we're going to look at all of your browsing history. Then we're going to buy DoubleClick. Then we're going to run targeted ads based on everything. Then we're going to Gmail and read every email. You know, Microsoft says they don't read your email, but Google says, well, we'll do what we want, but we won't pry too much. But they do read your email. And so, that slippery slope of competition uh you know, corrupts the original mission statement gradually over time. I
[04:36] gave a whole presentation in Davos on on how this evolves. And Dario wants nothing more than some rules. Yeah. >> And he's actually legitimately pissed that he has to actually repeal his own ethical standards to be competitive because there are no rules. There and you know, this is exactly how it has to evolve because you know, Dario is in a position where he has to choose between being irrelevant, which doesn't help, or repealing the original pledge, which he doesn't want to do. Now we're in a hard
[05:07] place. Yeah, totally agree. Your earlier commentary, Dave, was really spot on. This is what Cory Doctorow calls an shittification, right? That that So, people promise something and then they gradually degrade it over time, and by the end of it, it's a show or >> I love the way you encapsulate everything in this. Yeah, there's no there's no credible mechanism to slow the race right now. And so, it's it's all out. Alex, what do you think about this? I think there's there was no credible mechanism to guarantee safety in the first place. I think the entire premise was probably wrong. I I I think
[05:38] sort of the the superficial gloss is, okay, we're in the the Red Queen's race, and this is the race condition that everyone 10 years ago was scared of finding the world in where we have a number of frontier labs all racing to do the terrible thing that you build the thing and everyone dies. I don't buy that at all. I I don't think this premise that either a heroic individual or a heroic frontier lab was ever going to be in a position to guarantee safety. And in fact, the sort of I remember back to to the earlier days of the the
[06:09] frontier labs where the concern and part of the reason why OpenAI was formed itself was concern of a singleton. Competition is how we guarantee that there isn't going to be a singleton that dominates the future light cone with super intelligence. And I I think similarly, the notion that there's going to be sort of unilateral safetyism where single heroic individual, like one of the more prominent AI doomers, or a very safety-oriented frontier lab is somehow going to ensure safety throughout the
[06:40] forward light cone. That was never going to happen. Safety, to the extent we get it, is going to come from competition. It's going to come, I think, from a balance of powers and a separation of powers. And I I think what we want is competition between the frontier labs and maybe even to some extent competition between nation states, such as what we're seeing, to compete to do the best job for advancing humanity. And I think any unilateral safetyism probably a dead end. One of the questions is will safety become an
[07:11] emergent property in some form or shape, right? So, right now what we've seen is Anthropic go from a policy of we won't build it unless it's safe. That That's been their policy to we'll build it as safely as the competition is building theirs. Um and unfortunately, it's a slippery slope potentially down to the bottom. But it's But it's >> the mechanism for any kind of emergent property around the future light cone. >> Well, we haven't seen the mechanism for emergent properties of what we've seen so far, either. Well, I I I would take the position that we are. that in in
[07:42] some sense again the the fundamental flaw I think in the thesis that safety would originate from a heroic individual or heroic organization is I would argue it takes an entire civilization to align a superintelligence. We we took we took all of human humanities content online and used it in compressed form to pre-train AGI baby AGI in the early days like summer of 2020 with GPT-3. Why wouldn't it be reasonable to expect that take all of humanity to defensively
[08:13] co-align and co-scale superintelligence as well. It's not going to come from a single lab. What what do you think about Elon's point of view that we need to uh build ASI that is maximally truth-seeking um as his mechanism for alignment and for safety. I think that's just a fraction of what's needed. Um I I I that addresses a very specific issue which is look we don't want the AI to have one religion or to have one uh perspective on how you should live. Uh we want it to
[08:44] be truth-seeking and have all opinions in cap and we don't want to be censored. So that's definitely a problem but it doesn't address the imminent job loss um the imminent consumerism. You know the AI people are conceding all of their most private information to the AI the same way they did with their Google search history. And it's accumulating that data and people aren't fully aware of what it's going to do. It's going to turn around and start condensing you to do things. Mhm. And so if you don't have rules in place uh the natural you know profit motive of the AI companies is to start selling you
[09:15] things. And you saw this with that that Anthropic Superbowl ad that we showed in the pod a couple Yeah that was funny. Unbelievable I've showed everybody that ad now. But this is exactly where it's going to go if there are no rules. And so I I completely agree with Alex's perspective that 10 years from now after we've solved all physics we've solved all math you know we have global abundance all of this is going to look silly 10 years from now. But in the three year timeline, massive job loss, Yeah. total confusion, and massive rampant AI sales consumerism that has no
[09:45] regulation around it right now. It's going to be an absolute cluster. Yeah. Especially for the consumers. David, especially for consumer-first companies that are needing to generate revenue. Yeah. Right. >> Yeah. Well, you know, I actually after that last pod, you know, you showed that chart, Peter, that had Anthropic, you know, growing 10x year over year, uh 26 billion in revenue forecasted this year, and on its current trend will be the first company to hit a trillion dollars in revenue in history by 2029, 2030. >> OpenAI this year. And exceed OpenAI this
[10:16] year. Crazy numbers. But I said on the pod, you know, that implies like a 30 billion dollar or 30 trillion dollar valuation. But then I ran it through Perplexity, and it said, "No, that implies a one quadrillion dollar valuation using the current market price or ratio ratio. Yeah. So, I was like, "Okay, one quadrillion We discussed this a few podcasts ago that we'll see the first 100 trillion dollar companies uh before the end of this decade. Um Anyway, I I think that this is a more honest policy for for Anthropic uh end
[10:48] of the day. Uh it's still >> Safetyism was never going to work. I mean, we we we all know a number of folks at MIT and otherwise who advocated for a six-month pause just for the entire space to cool off and wait for safety to catch up. Did safety catch up? Whatever that means, no, not at all. If anything, that function doesn't accelerate to capabilities. I also think even in the DNA of Anthropic, Anthropic was originally, recall, was originally founded as an exodus of OpenAI employees who were purportedly concerned about
[11:19] safetyism or lack thereof at OpenAI. So, they start a safety/AI alignment-oriented firm. Then they rapidly discover that the best way to do safety is to have your own models. And they discover the best way to have your own models is to raise a bunch of money to train your own models, and then they discover the best way to raise money to train your own models is to generate revenue. And the cycle completes where yet again an alignment-oriented firm becomes a capabilities firm. This happens over and over again. I would argue at this point alignment and
[11:50] capabilities are inseparable. There's like a deep duality there. Yeah. Did you see the new standard, by the way? It's uh Dario said, "Well, okay, we can't we can't live by our original plan to plan to not train advanced AI unless safety is guaranteed." So, the new standard is we need to be as good or better than anyone else. It's like, "Wow, that's that's a very different bar." And we we see, you know, recently with the whole Department of War uh debacle with Anthropic and OpenAI, OpenAI cuts a deal.
[12:20] Anthropic, where does Anthropic stand right now in that whole conversation? They're in limbo. I mean, I I write about this every day in my newsletter. Anthropic is at the moment, my understanding is they're in limbo. They're probably in negotiation with the Department of War, but they're they're otherwise in limbo and and cut off as a supplier and considered I I'm not sure whether they've I think Dario and others have made some formal statements that they haven't received anything in writing yet from the Department of War, but my understanding is that this
[12:50] administration has is considering them a supply chain risk. And at the same time, notably, OpenAI struck a deal. Yeah. And at the same time, we hear that Anthropic was used by the Department of War to actually plan the attacks in Iran. Well, one thing that's really really Yeah, I know. I mean, you look, it's it's really clear that the the people who control AI, the US government and otherwise, can take out any world leader at any time now. The combination of satellites, AI to read every image, and uh you know, universal cameras, it makes
[13:22] it impossible to decapitate any country at any time. We proven that twice in the last quarter. Uh, so the future of warfare is basically whoever controls AI chooses who gets to stay in power. You know, Dave, that's a really important point. One of the things I've mentioned before is we're living in a world where you can know anything, anytime, anywhere, right? It's a it's a trillion sensor econ uh, you know, over a trillion sensor uh, planet right now with drones, orbital satellites, autonomous vehicles
[13:53] gathering data, and then AI doing predictive analytics on what things are likely to be even if you don't have data for it. And and maybe one other thing is sorry to interrupt. I was just going to say maybe not even just a means to an end, but also depending on which analysis of the Iranian situation you subscribe to, maybe an end to an end as well. If if you look at Venezuela and the oil exports to China, and you look at Iran and the oil exports to China, the picture emerges or at least one
[14:23] possible picture emerges that what we're seeing it is not just AI where Claude is being used to to perform the Venezuelan operation, perform the Iranian operation as a means to some sort of arbitrary or nebulous geopolitical purpose, but actually arguably with China looming in the background and possible Chinese invasion of Taiwan and the risk to the semiconductor supply chain and Western AI that would cause, it may be the case
[14:53] that AI is also the end to to the means to the end, and that what we're seeing more broadly is in in some sense superintelligence being used to protect the future of Western superintelligence. Yeah. >> Yeah, and there's there's a window of opportunity maybe a few months to put some kind of structure around this globally where you'll you'll see later in the podcast that the the models are improving at this like 3x, 4x reduction in parameter count, 10x increases in intelligence, you know, just every time we podcast, it's another step up. And we
[15:24] were already predicting, or I was anyway, that this is going to be a 100 X year uh just in terms of raw parameter count, but I think that's the lower bound now, looking at at how just the beginning of the year has progressed. So, there's a window of time where the power of AI that percolates out to every country in the world is going to be ridiculous by the end of this year. You know, create any virus you want, create any nuclear weapon you want, just working with your AI agent. So, there's a window where we can start thinking about regulations that register
[15:54] the AI use cases and agents and chips and processing before chaos breaks out. But you can see that that that window is is executable now because you saw Venezuela, you're seeing Iran, you know, clearly there's this tipping point happening right now where, you know, whether it's NATO or whether it's the United Nations or whether it's the US Congress, some entity needs to start formulating some structure around this because it's happening this year. Yeah, I mean, people need to wake up. It's Saleem,
[16:24] please go ahead, but I just want to say one thing. People have to wake up to the fact that AI is the single most important force impacting everything. Um you know, ever every single element of humanity right now is going to be accelerated, reinvented by this Saleem, go ahead, please. Um to Dave, just struck me that you mentioned that the Congress, the UN, NATO probably the three most toothless um um entities on the planet today. So, the
[16:54] thought that they would actually get together and do something or anybody do anything, I think is is low. I think we have to assume that it won't happen and look at the the other side of that. One thing about the Anthropic case, there is a potential I looked up an analysis, they do have a legal challenge potential because the way that that was classified is so ridiculous uh to make them a existential risk and all that supply chain risk etc. The That they have legal recourse to fighting that and they might might win that. Yeah, the thing about that legal recourse is that that process
[17:25] is usually a 3-year long window, which is hilarious It's only in the age of AI. It's only in the age of AI. What I find really upsetting is that in this scenario everybody loses. Yeah. >> Yeah, yeah. There's no winners in this. No, if if there's no framework and no rules, uh it's it's a lot like the NFL was back, you know, 20 years ago when the defensive coordinators would pay bounties to the linebackers to take out the quarterback. Like just take him off the field. I don't care if you break his legs. Uh and you and take the, you know, 15-yard penalty. Who cares? Because then he's done for the season. The NFL said,
[17:55] "This is not good for business. Like we need we need some rules." >> I did not expect that pivot. Well, that's where we are with AI right now. It's like, "Hey, I agree. Uh forget it. I don't I don't even want to go down that road. >> Let's let's continue Let's continue with uh the Anthropic story. Everybody, you may not know this, but I've got an incredible research team. And every week myself and my research team study the meta trends that are impacting the world. Topics like computation, sensors, networks, AI, robotics, 3D printing, synthetic biology. And these meta trend reports I
[18:25] put out once a week enable you to see the future 10 years ahead of anybody else. If you'd like to get access to the meta trends newsletter every week, go to diamandis.com/metatrends. That's diamandis.com/metatrends. I found this story pretty fascinating. So, Anthropic expands Claude's agentic capacity. Uh two different sides of the equation here. Uh co-work gains scheduling. Right? So, this is cron jobs. So, Claude completes recurring tasks at specific times. For example,
[18:56] generating your morning briefing or spreadsheet updates or your Friday presentations. I mean, that element was very much what we saw in open claw. Right? Uh and so, it's interesting. And the second half of this is that Claude code has enabled remote control. So, you can kick off a task on your terminal and pick it up on your phone. Um you can control it uh from the Claude app or from a a URL. And I'm wondering, you know, this has probably been in the works for some time. So, when Anthropic basically, you
[19:28] know, tried to do the kibosh on on Claude bot, uh I'm wondering if that was because they had this in the works. Uh basically, what Open Claude has been doing is what Anthropic is just rolling out under under a different approach. Oh, for sure. >> I I take Anthropic at their face that this was or Open Claude, I guess, that the challenge was more trademark-oriented than anything else. But, I I do think their What have I been saying for uh weeks {slash} days at this
[20:01] point that was distinctive about Open Claude? It's the two things. It's It's headless, able to function autonomously for 24/7, and also that it's convenient to chat with via conventional messaging channels. And what do you see here with Co-work? Co-work is able to autonomously be scheduled headlessly. That's the headless part. And then remote control, that's the the mobile messaging type part. But, I think both of these are half measures. Like, I'm insufficiently motivated by each of these. I use
[20:31] Co-work from time to time, and I use Claude code all the time. And neither of these, I think, is is as compelling, at least conceptually, as a more Open Claude-ish framework where all of these are cleanly packaged. And I I think My guess is Anthropic and OpenAI and all of the other bigs will be forced to release their own sort of first-party Open Claude competitor sometime in the next couple months. So, what >> What you guys There's something I found very profound about this plus our last conversation around Open Claude and
[21:03] everything happening. I think I I thinking about over the last couple of days, something very profound is happening, which is the sheer democratization of compute power, right? Note the agency of an individual developer with a Mac mini and running Quen locally and Open Claw has unbelievable agency and decentralization now. Uh it's not controlled by any centralized authority, not controlled by any uh centralized command structure. They can essentially operate as they feel like. So, this is this incredible independence and agency at the edge,
[21:34] which is going to really blow open innovation in a way that we 60s, baby. Total democratization, total demonetization. >> as we're seeing it happen and it's cascading down, as Dave mentioned earlier. Ironically, from China. Yes, ironically. Well, ironically from China. And then one other nugget, you know, Peter, your theory is 100% right. You know, Anthropic What why didn't Anthropic just throw out something better than Open Claw a year ago? It can and will delete things off your laptop.
[22:04] And so all these Open Claw users, including my kids, including me, actually have separate laptops or separate Mac minis, you know, including Alex Finn, you know, in our podcast we just did. Uh they run it on isolated hardware. Anthropic couldn't really contemplate throwing out a product then say, "Yeah, but run it on separate hardware." Don't tell like how how are you going to do that? So, this creates a huge entrepreneurial opportunity though. If you say, you know, listen to what Alex said a second ago, Open Claw's unbelievably compelling. And anyone who's started down that path will never go back, right? It's just you just you'll never give up your Jarvis once
[22:35] you have a Jarvis. Have you done Have any of you played with with Perplexity's computer? I've been hearing really good things. I've not tried it yet. I I looked at the demo. I I think it's an interesting step in the direction of councils for everything, and I've had so many people over the past few months ask me for something like Perplexity's computer, where you they're right now if you have a given task, they'll manually go to the top three or four frontier models, ask them for independent opinions, and then try to synthesize
[23:05] that in into one coherent whole, and that that is essentially what Perplexity computer tries to automate. There are others in the space as well, but I think even there that it's nice sort of sugar, uh, syntactic sugar, if you will, around the existing models, but I don't think it's transformative. I I think ultimately even this ability to counsel up or to create juries around lots of competing models, that's just going to be table stakes as with so many other forms of scaffolding. Did you just say syntactic
[23:36] sugar? It's a term of art in computer science. Goes way back. Alex, I think the point you made a minute ago is brilliant, and I Dave, I think you you were saying this as well, all of the big players, all the hyperscalers, all the frontier models are going to have to develop some version of Open Claw because it's going to become the de facto. Every person's going to have their own version of Jarvis. Yes. But remember, it's really expensive, too. This is part of the reason, it's not the only reason for running, say,
[24:07] Quinn locally under an Open Claw scaffold. That's a lot of compute if if you have one or more agents that are running constantly for you. I'm not sure Anthropic in its present state even has the cloud infra to be able to launch a product like that, and I think in many cases Anthropic, OpenAI, the others are probably just waiting around for for their infra to catch up with applications like that before they launch it. Yeah, agreed. This is the entrepreneurial opportunity of a lifetime, though. Anybody who jumps in, and there's so many different versions,
[24:37] so many different things to play with, but when you go to JP Morgan, you know, Justin Milligan who just joined us, his division at JP Morgan was only allowed to use GPT-4. Wow. Are you kidding me? And he couldn't take it. He's like, this is ridiculous. But no one has figured out, okay, but how can I use it in this highly secure, inside the firewall, inside JP Morgan environment? And and you know, Dario is not going to answer that question and you know, they the open claw team isn't going to answer that question and they want everyone to thrive who uses their
[25:07] platforms. They don't they don't want to kill every job. They want any early adopter to to thrive as they thrive and you know, Dario if he hits a quadrillion dollar valuation doesn't need more more money. He needs a he needs to not destroy every job in America and the world. And so this is really entrepreneurial heaven if you can figure out how do I get what I can use right here on my my Mac mini and it can clearly solve all these problems. How do I get that inside a real world use case without breaking everything without regulatory problems
[25:37] like many many many job opportunities in that vein. One of the challenges still with even with with with Claude's agentic capacity is um you know, giving AI recurring unsupervised access to your workflows uh means that either they're going to be a bunch of errors or you're going to be spending all your time checking the work before you hit you know, publish. Um it's still going to keep the human you know, in the loop to assure
[26:08] quality or alignment. There will be a point at which you trust it completely but we're not there yet. This is economics 101 or I should say microeconomics 101 when the cost of one good falls to near zero, the value of the complementary good increases. So as the value or I should say as the cost of generation of content becomes post scarce, which is exactly what we're seeing, that increases the value of its complement, which is verification for
[26:38] now. Yep. For sure. All right, going to our next story, Claude and keeping on the on the Claude theme, Claude gains co-work plugin templates for finance, banking and HR. So this is fascinating, right? Anthropic is building enterprise agent marketplace. It's department level AI infrastructure. And it's taking down company after company after industry. Um What's you know, we've seen just the
[27:08] decimation of uh of a number of players out there. What are you thinking about this? >> I I wouldn't interpret this as you know, when Microsoft launches an assault like on the relational database, it's it's a big you know, multi-billion dollar investment. Here Anthropic can build these connectors and adapters five code them in probably an hour. You know, it's and anyone else can too. So I wouldn't perceive it as you know, Anthropic is taking over all banking software. It's just so easy to build this stuff now that you might as well roll out all that
[27:38] functionality. So I wouldn't over read into the intent behind it. I don't think it's I don't think it's intent. I'm just saying you know, >> The implications are profound though. So I've got two thoughts. One is every department now becomes like a programmable intelligence layer. Right? And basically all prescriptive logic in companies collapses into these AI agent roles. And the real prize here is enterprise orchestration. Not so much chatbots, but autonomous workflow networks. Because this this is
[28:08] what I talked about last time. This is the organizational singularity. We go from human-centric approvals hop to hop to hop human to human to human to agentic workflows with human beings doing oversight, dashboard monitoring, and um exception handling. A couple of comments on this one. If you actually look at what these plugins are that Anthropic's launching that are causing the so-called SASpocalypse and reducing you know, carving one and a half trillion dollars off of market caps
[28:38] of various software companies, they are absurdly simple. They're they're just a bunch of MCP model control protocol wrappers and a bunch of skills with a a set of bullets for how to go about carrying out different job industry roles or labor categories. This is not that complicated. I'm I'm reminded Remember the scene in The Matrix uh the the the villain is is busy unplugging people uh without their without their cooperation from the Matrix, killing them in the process, and one of them
[29:08] says, "Not like this. Not like this." That that's basically what we're seeing. We're These are These are just simple text files in many cases that are reducing single-handedly the market multiple, the trading multiple, of entire industries. I I think on the one hand, it's incredible that a simple text file can say "Chop 10% of the value off of a CRM firm." Uh at least market value. On the other hand, these as pointed out earlier, these
[29:38] these plugins and the marketplaces of the plugins are so absurdly simple that I would reasonably expect these plugins are going to get just built in, since they're just scaffolding anyway, built into the next baseline version of the model and won't even need to exist independently in the future. Alex, I think the interesting point here is a year ago, if you had delivered this as an entrepreneur, you'd be out in the market raising at multi-billion dollar valuations. Mhm. Yeah, it's called it's called hyper deflation
[30:08] for a reason, Peter. I mean, I I get it. I just want people to be aware that, you know, the moat for an entrepreneur coming forward with something amazing, uh that we're going to reinvent the entire HR industry or the investment banking industry, uh and you know, raising at a four or five billion dollar valuation, it's basically that moat's gone months or year later. I think it's really important though to to step back and look at the macro every now and then and say, "Look, abundance
[30:39] is going to be rampant." We're talking about tens of thousands of times more capacity to create more money, more value created. Abundance is going to be absolutely rampant. And there's no reason to be afraid even though like if you're a CRM company, your 20-year future cash flows from recurring maintenance revenue is suddenly gone. Yeah. That's true, but the opportunity to pivot and thrive is bigger than than ever. And so I think a lot of people are they'll be a ton of volatility cuz people haven't mapped to
[31:10] the new reality yet. But opportunity is bigger not smaller overall. So let me ask you >> that agility is fundamental to large organizations success. You know, I talked about on the last pod the uh you know, the asteroid hitting the earth and changing the environment so rapidly and the slow lumbering dinosaurs going extinct. That's exactly what we're talking about here. I mean, Saleem, do you think that we can see large companies pivoting you know, rapidly enough? Zero. Zero chance. >> be able to do it. I mean, look, we've seen this throughout history. It doesn't
[31:41] It doesn't work. I think where you end up is not to throw another metaphor at this, but you end up where we saw with Google Ads where you you kind of took out the advertising market massively and then Google Ads becomes like a coral reef with lots of little species feeding off the reef. Uh and if you're the reef then you're in great shape. Uh but in this case, the reef itself is disappearing as we can decentralize completely to uh one-off uh computers running things. There are people using
[32:11] open cloud to go to small businesses, sitting down in front of them and automating workflows live for small businesses. This is like incredible what's going on. You know what else, Saleem? Uh there are a lot of private equity funds that are coming at us now saying, "Hey, you know, big companies never change quickly. Wait, this big company could be a small company very quickly cuz we don't need all these people. Now we have a small company with huge huge cash flow." Wow. So it's so you can be a small company fund emerging shortly or there's not if it's not there already
[32:41] that is going to buy up kind of medium-sized big companies and set up a digital twin infrastructure on the side where you have an AI native digital twin and you just move workflows over to it and you'll collapse the cost of running that organization by about three to five X. >> what MacroHard is about. MacroHard exists. I I I've started multiple companies like that. I've even tried to popularize a term for it. I call it an IBO an AI buyout. We've seen multiple PE firms doing that. Like this is table stakes at this point. Yeah, and of
[33:12] course MacroHard's, you know, vision is I'm going to come in and digitize your entire employee base and operate it. >> for that's for pure software plays, but I think we're going to start to see this in in real uh potatoes like like Project Prometheus from Jeff Bezos is is attempting to do this for industrial firms. Uh yeah. Anyway, I think the point uh here is uh large companies need to take action right away. So
[33:42] Saleem, what is your what's your advice for a large company a CEO listening and and seeing this coming their way? What do they do? Uh exactly what Alex just said. You set up an AI native digital twin on the edge. You run an immune system 10-week sprint uh to block the response from the mothership. You grow this thing and slowly move workflows over or as quickly as you can. You do a combination of bottom-up and top-down workflows. And the real shift in people's heads needs to be that instead of human-centric workflows, which is how what it's been
[34:12] like for the last 150 years, we now move to agentic workflows where you can get things done much more effectively with uh hordes of little agents, two layers, strategic layer and an execution layer and then human beings uh doing oversight exception handling, etc. cuz coordination costs go to near zero, execution costs go to near zero, and in and outside the firm, the future of the firm becomes a legal fiduciary liability purpose holder. And there's one two other two other things
[34:42] real quick. The first is the your if it's if it's reasonably good still, you own your brand and you own those customer relationships for the moment. I I think it's worth also rereading Clay Christensen, The Innovator's Dilemma, which it exactly addresses this. Salim, I know you're a big a big fan. We all should be. The Innovator's Dilemma contemplates, "Hey, every 10 years something truly disruptive is going to obliterate whatever you do." And here's how you should react to it in that moment. But now instead of every 10 years, it's going to be every 10 months,
[35:12] and then soon it'll be every 10 weeks, and then, you know, it'll be every 10 days pretty soon, too. But it did the playbook is still the same, you know, read The Innovator's Dilemma, invest in the new thing, use your capital leverage in your installed base to invest in the new thing. I just got off a board call for one of my portfolio companies, and you know, my comment to my board and to all boards out there is you have got to give your CEO top cover to be dramatic in their modification of
[35:42] the business. Because uh you know, and you have I know >> Yeah, you're either the disruptor or you're disrupted. Yeah, and it's for everyone right now. You get founder mode, and you get founder mode. Yes, I mean, that's basically it. If you're if thing if the company and the board and the board and the CEO are not in founder mode and willing to do dramatic uh surgery on the company, you're dead. You're you're walking dead in any industry. I'd also be remiss, Peter, if I didn't point out here we are basically on on
[36:12] the eve of abundant knowledge work, knowledge work of course being cooked, knowledge work about to be post-scarce, and here we are wringing our hands over where to find scarcities in knowledge work as it's about to become abundant. I just want to point out the irony. Ah, it's such an extraordinary time to be alive. All right, uh talking about disruption, uh disruption coming out of China. Alibaba's 35 billion parameter Qwen 3.5
[36:43] uh medium outpaces 235 billion Qwen 3 in benchmarks. Um the power of small open weight models. Um So, Alex, to you, buddy. This is happening in Western models, too. The difference is when, say, OpenAI launches a mini model or Google DeepMind launches a flash model, they don't advertise the parameter count, so it's not as viciously obvious as it is when a a Chinese frontier lab launches an open weight model. We get to see the benefits of distillation in a
[37:15] successor model, but it's striking. We're we're seeing almost 10x reductions in parameter count while maintaining capabilities or even increasing capabilities. And and the the broader picture just to keep in mind is the capability density of models is increasing. This goes hand in hand with what we've talked about in the past, Sam Altman's comment about 40x year-over-year hyper-deflation of costs at constant capability. In in this case, I just my mind immediately goes to
[37:45] what's the end game here. If if we can see an increase in capabilities with a reduction from 235 billion parameters to 35 billion parameters, what does the end game looks look like? Where does this end? Does it end up >> this point during our podcast with him. Do you remember that, Dave? Oh, yeah, for sure. For sure. And yeah, he he said he does he asked his research team not to give him parameter count anymore, just give me bytes. Yeah. Because, you know, they keep quantizing and shrinking the file size. Uh I had a lot to say about that, but I
[38:16] I bit my tongue cuz cuz that perspective isn't right, either. But but Alex predicted this a long time ago. I don't know how you saw this coming. But there's a lot of >> look at look at the scaling law curves and extrapolate. Yeah. Well, it's I mean I mean In this case >> I was I was on the treadmill this morning watching old Moonshots podcasts and I'm like, "Wow, that was like so long ago." And I look at the timestamp and it was it was only like 2 months ago or 3 months ago. Like, "Holy crap, things are changing so quickly." But yeah, Alex you said this. I think you're the first person I ever
[38:46] heard from it saying, "Look, you know, the equivalent of a GPT-5 is going to be maybe 30, 40 billion parameters, but it could get as low as 1 or 2 billion truly Cuz right now when they train that caliber of model, it has all this junk knowledge in it, too. Not or thinking junk, you know, Twitter feeds and Kardashian news and all that other junk. But strip that out, this could get very small and very tight, very >> It could get way smaller than a billion. I mean, I I could imagine scenarios where it's only a few million parameter
[39:16] equivalent. That's sort of the core microkernel of AGI or superintelligence and the rest lives in a flat text database or something. >> that will you know, another thing Elon said is we're not as smart as we think we are. If you if you get superhuman intelligence down to a couple million parameters, you're going to be like, "Wow, but we are we're really not that smart after all." >> person I heard speak about this was actually Emad Emad Mostaque speaking about what you can get onto your phone. We'll see that in a moment. I mean, so my question is is this bad news for the
[39:46] big compute incumbents? Right? If if massive data centers are being invested right now I think I think this is good for the This is so good for the startup community. It's fantastic. We still need the big models. Well, it it the it comes down to is something Alex also talks about a lot, which is do we have boundless boundless problems we can solve? Like if you get everything to be 100 times faster this year, you can just do that much more. Do we actually run out of things to do? You know, is physics infinite or is physics finite? Is you know, is human
[40:17] benefit infinite or is it finite? And we'll find out, I guess, in a year. But my guess is no, every every time you shrink the model and make it faster, you're still going to use every single chip in that data center just for the next thing. It's especially when you start getting into the the full cell simulator and all the health stuff that everyone's really eager to do. I mean, that is very very compute intensive. And I mean, that's going to be better. Peter, I I thought 64 kilobyte should be enough for anyone. Exactly. Exactly. Oh, the good old days.
[40:49] Uh check out this. So, I saw this on X this morning. So, this is Qwen 3.5 running on an iPhone 17 Pro uh on airplane mode. Um and this is extraordinary. So, it's a a 2 billion parameter 6-bit model running on Apple silicon. So, imagine you're anyplace on the planet, you don't have Wi-Fi, but you've got Qwen on your on your device, and it's got all the intelligence you need.
[41:21] Uh I find this uh Yeah. I guess seeing demos like this in in my mind underline either, depending on whether you want to see it as competence or otherwise, how much of an opportunity Apple has to finally take the lead with local models or conversely, how far behind they are in terms of taking the lead in terms of local models. But either way, clearly there's this enormous overhang. We could be running enormously competent reasoning models locally on all of our recent iPhones. The fact that it's not
[41:53] yet baked into the operating system obviously I mean is very publicly uh embarrassing, maybe one wants to call it, for Apple. On the other hand, lots of rumors that this time around, finally with Gemini integration, they they're on the critical path and they'll finally launch something in >> Finally, Siri will not suck anymore. Apple intelligence, however they brand it. Note that the local ability able to go offline means it's unstoppable, it's uncensorable. I mean, this is incredible. Yeah. Well, that's yeah, that's that's the ultimate barrier too, cuz if if this can get to the level this
[42:25] year where it can it can do a gain of function virus, it can do a chemical weapon Mhm. and it all fits into a tiny tiny little package. Uh you know, with nuclear proliferation back in the 1950s, there was a theory that hey, you know, if these physicists keep chugging along, they're going to make something the size of a grenade that has the power of an H-bomb. And then thankfully that didn't happen. It just the physics didn't allow it. Um but the AI is not going to stop like that. The AI is going to keep getting faster and denser and more compact and
[42:55] the window of opportunity to put rules and regulations around this is very very narrow now. It's really it's got to be this calendar year. >> So, what do you think is going on in at the White House, in Congress, in the Department of War? >> conversation before, right? We had the head of um one of the big agencies, the head of innovation of one of the big agencies at Singularity, Peter. Uh-huh. You probably remember this answer. >> We asked him, how do you think about this when somebody could design a virus on an iPhone, something something, etc. And he said, look, and it was a much more clever answer than I thought he
[43:25] would give, which was when you have kind of nuclear weapons, you know how many there are, you know where they are, you put eyes on it, you try and track it as much as possible. Great. When you've got something that's this democratized, what they're actively doing is opening up these communities. So, they went to the biohacking communities and funded them to open up because if you're trying to do something dodgy, you kind of need to collaborate with a few people and the conversations surface very quickly. And then the community does self-policing, self-reporting. Somebody's doing something dodgy, asking a few people to they point it out, etc. etc. cuz it's in
[43:55] their best interest. And it's actually worked very very well so far. What it goes what happens when you get to this level is unclear, but I think the general trend has been very positive so far. I'll tell you one other thing, you know, the way the way this evolved with financial services being self-regulated is we think of it right now as oh, the federal government is incompetent. They're not doing anything. The researchers over at Anthropic are brilliant. They're moving a million miles an hour. It's going to end up being the same people. And this is the way it worked out with the SEC. You know, when when you go who works at the SEC? Oh, it's the same guy that was at
[44:26] Goldman Sachs yesterday doing his 2 years at the SEC or her 2 years at the SEC and then going back to Goldman Sachs. That's the way it's going to be with AI, too. Right now, nothing is happening at the White House. David Sacks is there, though. You've got one brilliant guy. What's going to happen next is Anthropic people and OpenAI people are going to actually be the people working in the self-regulating agency. And so the the the people will have to bounce back and forth and they'll do it because they're they're worried. They're conscious of the of the impact of not doing it.
[44:58] Yeah, still concerning. All right. Still concerning to have this level of capability offline. We know how to handle decentralized capabilities already. We have printers in some in some cases states are trying to regulate 3D printers and before 3D printers we had 2D printers that could be used for counterfeiting. We have >> but we baked but Alex, we baked software into all of those printers, right? There was a standard that was created for They were the yellow dots. for any
[45:30] you know, for Canon, for HP, for any printer that detected you trying to you know, photocopy you know, money. Um it wouldn't allow that. So, the question is if we're talking about open weight models out of China that we don't control the software on, how do you bake in protection there? There are so many different ways that one can defensively co-scale against 2 billion parameter 6-bit models running on someone's iPhone. We've already talked about some of them. There are
[46:01] other ways in the scheme of things. I I don't think sort of these edge devices running tiny Chinese open weight models either individually or collectively pose an enormous hazard to the market. Like they're just not that capable relative to the other models that are out there. I think the frustration is that if the the solutions are relatively obvious to Alex. We've had this at this meeting at the state house before where it's like guys, it's not that hard. Here's what we need to do. And then nothing happens.
[46:31] You know, that's that's the frustration. But yeah, registering the models, registering the compute, you know, tracking the tracking the GPUs and where they are. It's all very doable. The ideas are not And and defensive co-scaling of of making sure that the most flops are going to good purposes rather than bad purposes. I'm reminded I think it was the New Yorker cartoon guy is is up late at night at his computer saying, "I can't come to bed. Someone somewhere said something wrong on the internet." We can't We We can't get so so bothered by the fact that
[47:03] someone somewhere might be doing something wrong with a two billion dollar two billion parameter model on an I've got so many agents running now and I I put in place a little rule that said, "Hey, before any process launches, write a mission statement and store it next to your code." And it's solved so many problems cuz the mission I can go back and read the mission statement and say, "Hey, what what the hell are you working on anyway?" Well, read my mission statement. Like, "Wow, that makes no sense or that makes tons of sense. It's It's It's so simple. Like cuz AI is the first self-documenting, self-improving, self-cleaning thing in the world. >> employee, all right. >> Yeah, yeah, just a couple simple little
[47:34] things like that will solve all these problems. Have you told your employees to do the same? Actually, yes. Um They It's a little bit different. It's like whatever you're doing, make sure that it's in a written document that the AI can see, too. I don't want any opaque activity because if the AI can't see it, then I don't want to see it. I want everything to be on the same page with us and the AIs. Salim, want to close this out here? Salim? You know, I think um a key point that we have to remember is
[48:04] the ratio of good to bad. Yeah. Um we worry about the downside, and we should worry about the downside, and it's and the the amplitude of the negative is getting bigger and bigger as people are going to run these models. But, I always go back to the eBay Craigslist example, where when you could first do eBay or Craigslist at scale, you could see human nature at scale. And so, anthropologists and sociologists studied the transactions at eBay and Craigslist. You can master email address pretty well. On eBay, I can throw up a picture of a MacBook, put grab your thousand bucks,
[48:35] and then I'm off to Fiji, right? So, how do you What's the actual ratio? Uh what is the real true nature of human humanity? And by studying these systems at scale, Kijiji and Cato and Mercado Libre in in Argentina, Craigslist, eBay, they found that the ratio is consistently 8,000 to 1, meaning there's 8,000 positive transactions on eBay for each fraudulent transaction. That should give you incredible optimism for the future of humanity. Yeah, agreed. All right, let's move us along here. Uh
[49:05] let's head to the Google verse. Google releases Nano Banana 2. Uh so, this is running on Gemini 3.1 flash. Uh it's 4K resolution. Uh it's at 0.045 cents per image. That's a price point that's cheaper than stock images. I'm sorry. Uh sorry, yeah, it's 4.5 cents per image. Excuse me. Um it's cheaper than buying stock images. Uh and so, is this the end of commercial
[49:37] photography, illustrators, stock image platforms? Uh probably. We're just getting started here, and I think maybe buried underneath the headline, but in the release documentation is this is the first image model from Google that combines a reasoning model, which I I think they used slightly flowery language for it, but basically the reasoning power of Nano Banana Pro with the instantaneity, uh I think they might have just said with the speed of Gemini Flash model.
[50:09] So, under the covers, I think this is like technically this is really interesting. It's combining probably some sort of diffusion model that that we get with reasoning capabilities. And I think achieving the cost reductions of a diffusion model with nonetheless the the capabilities of reasoning, we're going to see this spread from images where it's mostly right now in video back to text, back to code. There are a few other labs, smaller labs, that have started to make pretty loud announcements about how
[50:40] they're achieving purported 5x, 10x cost reductions, or speed increases using diffusion models instead of auto progressive transformers. But, I I think this is probably the tip of the iceberg for some like final consolidation of auto regressive transformers, which are used for code gen and natural language for the most part on the one hand, and then diffusion models and diffusion transformers on the other hand that are used for images and audio and video. We're just going to finally get one
[51:11] consolidated architecture at the end of the day that does everything. Yeah. I mean, this is the uh you know, wake-up call for people to remember that whatever you're seeing, uh you cannot necessarily believe it. Every pixel is going to be AI-generated at the end of the day. Um Saleem, thoughts? I mean, the cost drop is incredible. People are just going to do so much more with it. Democratization of creativity. Great. Love it. Absolutely amazing. Yeah, I'm kind of curious. I don't know if you guys know, but the the curve on
[51:43] intelligence is just ridiculous. But, on diffusion models, I don't really know. I know that they've gotten a lot faster and cheaper in the last few months, but it it doesn't feel like the same type of algorithm. Like it may hit a wall. I don't know. Do you guys know? Open AI has been investing This is in the published literature, investing a lot of effort and probably DeepMind as well, maybe slightly less prominently, in trying to avoid the need for many iterations on a diffusion model. So a diffusion model normally, conventionally, takes many iterations to
[52:15] start from pure noise and refine the pure noise into the final image or the final video. There There's There was a lot of interest that was publicly available, call it 6 to 12 months ago from OpenAI and some other folks as well to see if they could just one-shot or two-shot straight from pure noise to the final image. I I do think to to your point, Dave, I I think although I haven't seen in maybe in the past two to three months any scaling laws for diffusion models. Prior to
[52:45] that, I saw a ton of work on scaling laws for diffusion models, and diffusion models have scaling laws, too. Everything has scaling laws. Yeah. You know, Amjad would know all about this, too. Let's pick his brain next week in LA. Absolutely. He's speaking of diffusion. The new standard is, you know, go to Nano Banana 2 and ask it to generate imagery. So imagery becomes free, effectively, and and you can I mean, it used to be I'd go My current workflow My previous workflow was I'd go to Google Images and hope I
[53:16] found something. Right? Now everything is created from from scratch, and it's perfect. I love this image in this in the slide here of Elon with Sam and Dario and and the whole leadership team of all the hyperscalers. Alex, you should be in there, man. You got to got to raise your game here. One more notch. We're running We're running out of scarcities, but maybe appearing in that image is one of the scarcities our civilization has left. We We can We can make that happen for you. For For sure.
[53:47] All right, continuing on with with our friend to Google, Gemini can now automate some multi-step tasks on Android devices. So, Gemini is now an on-device agent that can navigate real apps and complete real transactions. You know, handle DoorDash, McDonald's, Starbucks for you. So, interesting, significant, what do you guys think? I think it's hugely significant. Expected. >> it's been a long time since there was a feature function on the phone that
[54:17] threatened Apple in any way, but AI is is it. You know, if you try to use Siri to do something constructive while you're driving, it's just so painfully impossible. Also, when you when you start an AI dialogue and you're in the middle of the conversation, the thought process, you don't want it to go away. It's a It's addictive and productive. And if it follows you on your phone and seamlessly, it's just incredibly empowering. So, if that if if Google wins that race with Android, they might actually
[54:47] you know, chip away at the iPhone profit dominance for the first time. Now, keep in mind that they also need the duopoly for antitrust reasons, so neither company can afford to completely annihilate the other one. They need They need some parity in the balance of the force. I don't know if you guys saw the data. I mean, we've seen a significant drop-off in mobile phones, right? In terms of mobile phone purchases. And that will be displaced of course by headwear and earwear and all kinds of devices that are beyond just your phone. Well, I think I think a lot of that is
[55:17] because So, the reason the phone sales dropped off is because they didn't have a function or a feature that everyone was clamoring for. Cuz people used to recent get a new phone when the cameras were improving like crazy. They'd get a new phone every 18 months to 2 years. Now, it's like, well, I can I can sit on this phone for 3 years, 4 years. I'm not even noticing the difference. Mhm. But again, AI could completely change that, the neural chips. Sorry, Alex. Go ahead. I I I think there's also a supply-side element where the rising cost of memory is making phones in some cases more expensive. And we're seeing I think a
[55:48] generational transition from smartphones absorbing the silicon and absorbing TSMC's output over to AI data centers as the new form factor for computers away from this. But just narrowly on Gemini for multi-step tasking on Android, this is what Siri was originally supposed to be about. Mhm. And before Siri, this is what the DARPA uh personal assistant that learns, or PAL, was supposed to deliver. We've We've known how to do this in some abstract sense for a decade plus. What was missing, why you ask are
[56:19] we only getting this now? I I always like to ask why do things take so long? Why can't they be faster? In this case, I I really think it was about a combination of reasoning models and vision language models that could fit compactly onto a personal device. And we're getting that now, finally. And it's going to be everywhere. But we really should have had this functionality even without the ability to read the screen and understand arbitrary applications, we should have had this 10 years ago. And that's borderline inexcusable. I think what's most significant here is the fact that
[56:51] Google has a huge installed base of phones. Right? Of Android phones. And the ability to take their their AI systems and that installed base, OpenAI doesn't have that, Anthropic doesn't have that. Uh and it's going to be a massive differentiator for Google. I'm a long-time >> it. Apple could have had it. Apple could have had it. I'm a long-time Android user, so I'm super excited by this because this is what >> turn all my iMessages green. It pisses me off. Apologies for ruining your visual field sphere, Peter. Um but this
[57:23] is like agency at the operating system level, which I think is amazing. And it also means that uh commerce APIs are becoming machine-to-machine first and human second, right? So, you'll have less friction in consumer flows. This is going to reshape marketplaces over time. So, it's really exciting. All right. Next article is a real fun one. Amazon makes a contingent offer to put $35 billion into Open AI based upon them first off going public and secondly achieving AGI. Enter Salim with his
[57:54] normal rant. What the hell is it? You know, it's kind of incredible that we financialized superintelligence, which is amazing. Having AGI as a financial milestone is unbelievable given we have no idea. I mean, it's great that intelligence has become a balance sheet trigger. That's incredible. But but so weird. And thank goodness it says or. Well, Alex doesn't the the agreement between Open AI and Microsoft requires Open AI to give the source code and all
[58:25] of the intellectual property to Microsoft until AGI. Do you think they use the same definition of AGI? >> I suspect it's that something similar. So, the the definition my understanding based on public reporting of the Open AI to Microsoft definition of AGI went through several iterations with the most recent iteration prior to their I think for-profit transition being and Salim maybe you'll like this. It did actually have a definition. It was something like generating $100 billion
[58:55] in either earnings or revenue. I forget. So, maybe we need to coin >> Okay. like AGI as a unit of currency like an AGI is $100 billion of earnings or something. We're measuring we're measuring compute in terms of gigawatts and AGI in terms of dollars. I love it. That's That's right. Listen, that's fine. You could That's just all they've done is substituted an earnings plateau for that, which is which is fine. It's good. So, this is interesting, right? This is $50 billion. It dwarfs Microsoft's $13 billion dollar
[59:26] And what, you know, again, I'm going back to what is Amazon really >> would like to make the point which we've made earlier, which is that a lot of this is Amazon credits. So, Yeah. >> And which is fine, cuz that's how they would have spent it anyway. There is there are lots of uh lots of tendrils going both directions from Amazon to Open AI and back based on public details of the announcement, like the requirement that Open AI will use Amazon's Trainium or Trainium 2 chips for training. It's good for Amazon. It's
[59:58] it's good Amazon has a a long and storied history of purchasing their own customers in in some sense, not in well, in some cases literally acquiring them, but in many cases paying for the information and the learnings that come from having a customer that's using Amazon as the world's most customer-centric company. And in this case, Amazon missed arguably missed the the frontier AI boat. And so, paying to get themselves to deal themselves back
[60:28] into the game is I think par for the course. Their their up to $50 of investment is at far worse terms than say Microsoft's original billions when Microsoft was much earlier in the game. And I I I think this is just the price of reestablishing themselves at the infra level of the party. And it's also been reported that as part of this deal, Amazon will get customized versions of Open AI's models. Internally, Amazon will get
[60:58] to post as the exclusive third-party cloud host Open AI's frontier suite of automated AI co-worker employees. So, Amazon will get a lot out of this, too. I mean, this is so incestuous, you know, what's going on right now. I mean, and Amazon was all Anthropic for a while. Now, they're Open AI. Um Well, I I I I I you say you say incestuous, but Go ahead, Dave. Sorry. Well, well, the the US public market all companies
[61:28] combined is about 50 trillion. The AI companies are 20 trillion of the 50 trillion now. So, you know, it's incestuous, but it's like if that 20 trillion becomes 30, 40 trillion, which it inevitably will, it's the majority of the market is just seven companies. So, when they do a lot of deals with each other, you know, it's like, well, is that incestuous? It's it's it's the whole freaking economy is is those handful of companies. >> also think maybe I would say incestuous not perhaps the the word I'd use for this, maybe circular it is is what we're
[61:58] gesturing at. But even that, I that's not my take at all. In this case, I I see competition and and I also see horizontal stratification. That if if Amazon is striking deals with Anthropic, but also with OpenAI, and OpenAI is moving some of its workload from Microsoft Cloud to Amazon and also to to Google TPU Clouds, that to me looks like A, the market for infrastructure for the frontier labs is very competitive. That's great for the economy. And B,
[62:29] it's starting to horizontally stratify. So, if if if OpenAI is feeling impulses not just to vertically integrate down to the data center layer itself, but is so compute starved that it needs to, following the law of comparative advantage, needs to outsource some of its compute to Amazon with its Tranium architecture and Google with its TPUs, that's a sign, if if anything, that there's such insatiable demand for compute that it's raining on everyone even with perhaps less loved computer architectures. Well, well, but but
[62:59] there's an interesting there's an interesting thing in negation here, which is XAI is missing in all these conversations, right? So, Elon is going 100% alone. Elon loves vertical integration. And he doesn't play well with others. He loves >> it's interesting that the you know, the the big big money we saw with Anthropic is in the corporate use case, corporate white-collar use case. People trust two clouds. They they trust the Well, three, I guess, if you count Oracle. They They trust the AWS cloud in
[63:29] a big way. And they trust the Microsoft cloud, Azure. And I guess they they sort of trust Oracle, too. >> Not Google Cloud? No, Google Cloud a bunch of our companies have been kind of bribed by Google to use Google Cloud. You know, as Alex was saying, they'll they'll pay you to switch. And some have taken it. But for the most part, you know, Google spies on everything, you know, that's their terms of service. Never say they won't do anything. If you read any terms of service from Google on any product, it says, "We may do this. We may do that.
[64:00] We may do the other." Which kind of implies they won't do other things. But if you read the legal, they literally don't restrict themselves in any way whatsoever from doing anything. It's honest. >> it's very corporate un- friendly. >> Honest, right? Yeah. Yeah. But you know, Microsoft legitimately says, "No, we will not steal your data. No, we will not steal your intellectual property. No, we will not read your email if you use Outlook." Um and and AWS is even more, you know, so so people trust those clouds. And then they want their AI model to be inside that trusted
[64:31] container inside the cloud. So so far it's just been Claude on AWS, you know, everyone's running away with Claude on on AWS. So all of a sudden, for reasons, you know, that I don't know, like maybe just variety, maybe maybe not having Microsoft and OpenAI be just, you know, bedfellows by themselves, Amazon's going way out of, you know, massive $50 billion uh move uh move here to get two options on the on the AWS. >> know what the valuation of this round is? I I think the reported valuation of
[65:02] OpenAI's overall round was $730 billion pre. Yeah. And and so, you know, this is not going to be a big risk. I I if Open A I OpenAI I public, it's likely to go public, you know, north of a trillion dollars. So, you'll get a quick pop and it's probably, you know, what do we, you know, we have three big IPOs coming up. SpaceX is, you know, anticipated maybe as early as next month. I heard. >> And then we'll have Anthropic and then we'll have Open A I. So, I mean, if you can get a 50% pop in your in your
[65:34] price shares in in 6 months, that's incredible investment. That's not investment advice for anyone who's going to misconstrue that. Well, hey, listen, I will give investment advice for people to get 50% in 6 months. I mean, why not? Just that >> That's not investment advice from me. It's from Peter. >> Okay. Listen, if anybody can can get a 50% return in 6 months in any deal, that's pretty damn good. This episode is brought to you by Blitzcy, autonomous
[66:04] software development with infinite code context. Blitzcy uses thousands of specialized AI agents that think for hours to understand enterprise-scale code bases with millions of lines of code. Engineers start every development sprint with the Blitzcy platform, bringing in their development requirements. The Blitzcy platform provides a plan, then generates and pre-compiles code for each task. Blitzcy delivers 80% or more of the development work autonomously
[66:34] while providing a guide for the final 20% of human development work required to complete the sprint. Enterprises are achieving a 5x engineering velocity increase when incorporating Blitzcy as their pre-IDE development tool, pairing it with their coding co-pilot of choice to bring an AI-native SDLC into their org. Ready to 5x your engineering velocity? Visit blitzcy.com to schedule a demo and start building with Blitzcy today.
[67:05] Another fun article this week is coming from Pulsia AI created by Ben Sarah that runs companies autonomously. So they're currently running over a thousand companies. So imagine being able to take your company, putting it on Pulsia AI and say go. So Dave, would you do this with any of your companies? Yeah, this is inevitable. I don't know if I'd use this exact product or not. I haven't checked it out yet, but 100% the philosophy is clearly where things are going. And you know, look at the end of the day, what is an executive team do? You
[67:35] know, other than a couple of hugely important key strategic directions, everything else is just performance reviews, paperwork, you know, whatever. All that can be very very AI'd now and it should be. >> in the room here is okay, I turn it over to to a Pulsia AI, but who's legally responsible? You know, if your company, you know, has a breach of contract or fraud or harms a customer, is it Pulsia? Is it you?
[68:05] That's why I think where this ends is we get this question I think in the AMAs and otherwise all the time. Like what's left for humans? Should everyone become an entrepreneur? Well, let the chorus of YouTube commenters say, well, not everyone wants to be an entrepreneur. I would say where this ends up not in the distant future, like 10 years from now, but in the the medium term, like five years from now is single person conglomerates where single person can oversee lots of agents
[68:36] that are all building businesses. This isn't for everyone obviously, but as we start to to get toward, you know, one person or a zero person unicorn becoming more and more popular. Again, I've I've argued in past we're we're likely already there in in some sense, but as we start to see that long tail of the the number of people per company over some valuation start to stretch out, I I think this model, call it a broader model of a one-person
[69:06] conglomerate where you have a person sitting on top of basically an entire PE firms worth of agents starts to make an enormous amount of sense. So, I I I've been poking at at Pulsar Pulsar and it's like a lot of microbusinesses and some of them look like they're they're varying levels of of seriousness. But, I I checked and you can actually with some of its microbusinesses you can actually go and purchase stuff. So, you can already engage in real commerce, spend real money via Stripe with some of the
[69:36] businesses that are running on its platform. And I think we're going to see so much more of this in the future. I'm super bullish. >> companies. They're not real businesses in terms of you know, significance of of revenue probably or complexity, but it's the beginning. Um I put I put Open EXO on there. Um You did? I did to see, okay, could we, you know, I literally talked to well, half an hour ago, but you know, create a new create a shadow AI digital twin on the edge and this is essentially it. I think Dave's point is valid. It may not be this one, but definitely these are going
[70:07] to be agentic hosting systems where you log a brand, you pick a service, it'll email and find customers for you, it'll run the execution for you. What we're seeing here is Costa's theory collapsing in real time, right? Because if you have a thousand companies in a few days that are AI run, this is the marginal cost of launching a company goes to zero now. It's 50 bucks a month to run an organization run a company on this. So, this is becoming really surreal and we're going to expect to see thousands of examples in this association of like
[70:37] like this thing. >> works, it'll blow up to millions. Yeah. And also, these these things always come up from the bottom. Yeah. >> And if you you know, cuz yeah, like like a Jamie Diamond or some senior exec will look at it and say, "It looks like a toy to me, you know, forget it. We're not doing this." And then it sneaks up on them and they get crushed. They're like, "What happened?" Well, some some guy I using it to manage a vending machine or manage a, you know, social media site. It seems so trivial, but it comes up quickly. And it sneaks up from the bottom. That's the way the Mac was, right? Or, you know, they the Mac was just perceived as a as a toy for college students, you
[71:08] know, it's never good for the enterprise, but then it grows up and grows into the enterprise. But this will happen much more quickly. It's just going to >> I would also argue we've seen this happen before in finance with quantitative trading algos, which went from none of the volume in public securities markets to 70, 80, 90 plus percent of the daily volume. And you know what? People survive. We still have human traders manually fat-fingering trades into the the public securities markets, but by volume, they're completely dominated by algorithmic
[71:39] traders. I think we're going to see the same thing happening in the rest of the world outside finance, in the physical world, in various e-commerce spaces, where over time most of the volume will eventually be dominated by algorithms. All right. Watch this Watch this news item, guys. Uh It'll be interesting. And I'm of course uh Pulsia AI is one, probably of many that'll be materializing. I thought this was a pretty fascinating
[72:10] conversation uh article on Burger King launches AI voice assistant called Patty in employee headsets. Let's watch a video. Hi there. Good morning, Patty. Looks like we had a great breakfast shift today. Is there anything that needs my immediate attention? The team's friendliness scores this morning were the highest this week. We are running low on Diet Coke in the freestyle machine. Thank you, Patty. Hi, Patty. We just sold our last cinnamon apple pie. Thanks for letting me know. Would you
[72:40] like me to remove them from our menu until tomorrow's shipment arrives? Yes, please. Okay. Apple pies have been removed from our menu boards, third-party delivery, kiosks, and BK app. I will add them back as soon as tomorrow's shipment arrives. Thank you, Patty. Meat puppets. Meat puppets. You have to two things. One, admire the the punny name Patty for a burger chain. So clever. Two, going back to my comments from few pods
[73:10] ago that we're going to be living in every single sci-fi scenario at once. This was a sci-fi scenario I would argue called Mana. Mana was a novel written by Marshall Brain about 20 plus years ago at this point where you had human employees who were on headsets all taking directions from a centralized AI in businesses. We're there. We've arrived in Mana and it starts with fast food. So, the only thing that that video didn't capture is how encouraging and enthusiastic the AI is.
[73:41] Now, whether you're using it to code, whether you're using it, you know, to walk around and and pick things out of the fryer later, it's it's just so engaging and energizing. And that's that's the part that people are surprised cuz it seems like, "Hey, the AI asking me to telling me what to do is dystopian." Yeah, maybe, but it's it's really much more empowering and engaging and fun than walking around by yourself. It It reminds It reminds me of the of the Baxter robot where you would move its arms and show it what to do. And it
[74:12] showed a very friendly fellow who was smiling while he was coaching the robot, but he was literally teaching it to take his own job. Um for me, the coaching tool is a transition to automation pressure. Uh frontline services obviously become AI mediated like performance management. So, uh this is the end point here is going to be very interesting. So, this is AI surveillance as well, right? So, this is the AI watching every employee. You know, this is beyond just saying please and thank you. It's rating them on their
[74:43] efficiency. Um and and you know, calling it a coaching tool Saleem is sort of like a corporate euphemism. Exactly. It's Orwellian. One has to admire the Orwellian nature of the naming. We we do. I'm waiting for it to say, "So, you dropped that the fries for the third time this morning?" Let's see how it deals with that. It says literally Peter, it's literally meat puppets. Oh my god, so funny. But so, you know,
[75:14] we probably see this entering everywhere. Right? I mean, when you're recording a customer service call right now, you're effectively doing that without the feedback in the moment. But as a CEO, if you want to understand who the weak players are in your company, or you want to try and provide, you know, on-the-job continuous coaching and see who can respond, this becomes kind of this is highly efficient but highly dicey.
[75:44] What happened? Yeah, that's right. You're saying Peter, it's not not just knowledge work that's cooked, cooking is cooked. I think you're going to see unions rebel against this. Big time. Big time. Yeah. Yeah, and I don't know if there's any winning that war. I mean, I think at the end of the day, the AI co-pilot is gathering a huge amount of data, and a lot of that data will go into the decision on what can be automated and what can't be automated. And over time, everything can be automated. >> is like the Amazon delivery worker who is wearing a pair a pair of AR glasses,
[76:17] and Amazon is saying, "Oh, this is, you know, to help you show where to put the package and warn you about there's a dog." No, no, no. Those AR glasses are training Amazon's model to replace you with a robot, to be very clear. Yeah. >> Yeah. But, you know, if you if you rebel against it, what's that going to achieve? You just got to get on the wave, man. There's no choice. Like, you just have to be a user. You have to get on either Cloudbot or one of these other platforms and it's it's coming.
[76:48] And yeah, you can you can you can go picket in front of Open AI's office like all those people, but it's not going to work out for you. I'm telling you. Like it's it's I'm well motivated. I don't blame you for doing it, but it but it's not going to work. Uh we're going to see uh we're going to see all of these fast food chains begin to bring in robots very shortly. Um I think this sort of version of Patty uh you know, we're going to get you know, the unions rebelling against it, but I think we'll end up making it voluntary. And if you really want to improve your
[77:18] abilities, you'll volunteer to use Patty. Um anyway, interesting story. >> warehouse worker or the fry waiter operator, if you get one in a thousand to volunteer, that's all the training data you need. Mhm. That's why it's it's fruitless to try and fight it because you know, the numbers just don't line up. >> Also I think this is hap- the transition can happen really quickly relative to political swings that this is you don't need that much training data to to automate away many of these tasks with with humanoid robots and VLAs so close
[77:49] to to being production ready for certain applications. I I just I don't think the transition period with Patties or again, call out to Mana by Marshall Brain which foresaw all of this 20 plus years ago. I don't think the transition period is going to be long enough to to even necessarily give political counter swings enough traction to make it worth it. One year? Two years? It's already hap- I mean the transition's already happening, but to to to to VLA robots, I I think yeah, next two to three years, you know. I I got just a quick thought here.
[78:20] Um before this really has time to penetrate, you're going to have drone deliveries of food like this and and it'll all be a lot of this. Yeah. Um I've got at the Abundance Summit this year, I've got uh incredible company, Zipline, uh coming uh to talk about what they've done. Um I love the company, love their ability to transform delivery service in the United States. This is of course the company that began in Rwanda by delivering blood supplies and are now
[78:50] operating with Walmart and delivering every 30 seconds. And their prediction in the next 2 to 3 years will be a delivery per second. Extraordinary progress. All right, another delivery company. This is Uber. So check this out. Uber employees have built an AI clone of Dara, the CEO, to practice their pitches. So before you go pitch Dara your idea, you should pitch it to his his AI clone. Um I'm curious, Salim, what do you think?
[79:22] Oh well, this is great at one level cuz you get executive-like cognition as a service. It really allows scalable leadership. Um the we are actually doing this at Open EXO where we've created a clone of me with loaded up with all the EXO thinking and we're rolling it out to all the community members so they can ask me a question as they're advising clients or companies or cities, whatever, and I don't have to be in the middle of that. So I think this is hugely relevant and I think it it makes absolute sense.
[79:52] Yep. At some point someone's going to ask can the AI clone of Dara actually function as CEO and not just for pitch practice? Exactly. It's the transition. >> highly likely that the avatar of Dara, of Peter, of Alex, of Salim will persist for a long long time with the same voice and the same face. Um and it's kind of in a sense locked in. If you win the race to being the avatar, people get used to it, but they like the fact that there's a human being behind
[80:22] it. I was telling Alex before the podcast started that I I just love his Spotify version of the daily you know the daily singularity. >> well. It's the same voice on YouTube, Spotify, and voiceover for Substack if folks want to listen to an AI version of myself on the Innermost Loop newsletter. Yeah, and I really don't care that it's not that it's AI-generated. It's I know it's Alex who wrote the content under the covers, and it just feels great. And but but without the human being behind it, if it was some synthetic, you know, never existed person, I wouldn't like that as much. Do you
[80:52] remember the movie Real Genius, one of my favorite movies? >> Yeah. Uh there's there's a scene in which it's it's typically taking it taking place at Caltech, and where, you know, the professor's in the front, and slowly the students are are instead of attending class, they're putting their tape recorder down. Um and the professor finally, instead of teaching a class, plays a tape to all the tape recorders recording it. So, I I I sort of imagine this is what we're going to see here with these AI clones of Dara.
[81:22] Uh you know, it's going to be Dara at some point is going to just like take a vacation and let his AI clone run the company and see how it does. We'll have to ask him this question on stage, Peter. Yeah, no, it's it's great. Uh and just as a reminder to everybody, you know, Dara's going to be on stage. Saleem and I are going to be interviewing him at the Abundance Summit. And uh this year, for the first time, we're doing a live stream uh at the Abundance Summit. Uh we're going to be live streaming Eric Schmidt. Where Dave and I are going to be doing the interview with Eric Schmidt and with uh with Dara, we're going to be doing a
[81:53] live Moonshots podcast uh with AWG, Dave, Saleem, myself. So, if you're interested in uh in actually listening to the live stream uh of the summit next week uh or depending when this goes out this week uh we're going to drop the link below. Uh it's free. We want to get this out as far and wide as we can. So, enjoy. You know, the Abundance Summit is a high-ticket price. Um and uh we have 600
[82:23] amazing CEOs flying in from around the world. But this live stream is free, so go to the link down below and check it out. Um and I hope you'll listen in. Okay. >> ticket sales, only 50k if you want to uh It is 50k, but we've been sold out. Like We are over sold. Sold out at 50k. What the hell? Over sold. Yeah. Well, it's it's an amazing event and so proud to have >> is. all three of you guys joining me this year. All right, one more event announcement.
[82:55] Uh again, uh super excited about this. I'm going to be joined by Ray Kurzweil, Steven Kotler, Dave and AWG on May the 4th. So, if you want to join a very exclusive event, uh spend the day with Ray, uh Steve, Steven Kotler, uh Sele- um actually AWG and and Dave, uh if you buy 100 copies of my new book, We Are As Gods, this is the book that Steven Kotler and I wrote as the
[83:25] follow-on to Abundance, um you can join us. Uh the URL is weareasgodsbook.com/100. We'll put the link below. Uh Dave, we're going to be holding this at uh One Kendall. All right, baby. at Link Studios. Super super cool. Alex, excited to have you as well. On the Star Wars Day. Is it a coincidence? Is Is this a Star Wars holiday? I'm a Star Trek guy, but you know, May the 4th be
[83:55] with you is is an important uh reminder of when we're holding this. So, uh I'm excited to have Ray there for, you know, 4 or 5 hours, go deep on all of these topics. And uh and again, uh if you want to help move We Are As Gods to the top of the New York Times best seller list, uh you can do that. Uh just go to the link. You buy 100 books, uh you'll be there. We'll give you signed copies and uh spend an enjoyable afternoon together going deep onto all topics exponential.
[84:27] All right. Moving on. Let's go to energy and data centers. Wow, look at this. US plans to add a record 86 gigawatts of utility-scale capacity this coming year. Salim, thoughts? Well, this is the point we've been making for a while that that the cost curve of solar is just dominating everything plus the cost curve of battery. And once you have battery and storage available, you can
[84:57] unlock solar in a massive way. I'm going to point to two data points. I track Ramez Naam if you want to kind of go deep on this cuz he tracks all this very carefully, but in 2016 it became cheaper to if you're doing power generation to do solar than fossil fuels. And so almost all uh energy generation since then has been doing that. But in 2019 we had a more important inflection point. It became cheaper to do CapEx build and run a solar facility than just run the just the OpEx of fossil fuels is
[85:28] more expensive than building and running solar. So, if basically from now on all uh energy generation for the most part except for specific legacy stuff except for political stuff is going to be renewables. And we see that taking over in India and China and now finally here. And I think this is really really amazing because solar just keeps on giving and it's just going to keep going that way. It's It's an unlimited resource. So, um to all of the And And by the way, people worry about coal etc.
[85:59] I think the coal industry in the US employs 60,000 people. The solar energy industry employs half a million people. So, it's not about the jobs either. So, get over it and let's just move on. Dave, you remember when Elon said he has a mission for for to generate 100 gigawatts of solar per year. Yeah. Yeah. Remember when Eric Schmidt said uh it was only a year ago said AI is going to require 100 gigawatts by 2029. It's a crisis. We'll never get there.
[86:29] And America's just incredible. Like when America gets mobilized, it's just the most amazing force in the world. And here we are. It's only, you know, a year later and we're like, "Yeah, we're going to find our 100 gigawatts." There's no way we're going to stop doing AI for lack of power. We'll find a way. Yeah. And it's interesting, right? Also in in in an environment with diminished subsidies also. All of the hand-wringing from months ago, all the subsidies are going away, how awful it is. No, we're we're getting solar even in the absence of of the same subsidies we had a couple years ago.
[86:59] >> The economics just take over. And it used to be driven by people's concerns about the environment. Now it Now it's making money and deploying AI. Feed the superintelligence. Yeah. All right, this is a big story this week. Uh tech giants to self-fund uh their production of power. So uh this is a White House effort. Uh we have Michael Kratsios at the center here, a friend of the pod. We'll be doing a podcast with with him in the next couple
[87:29] of months here uh asking the hyperscalers to actually build or buy their own power. Um and of course this is in response to consumers' concern about uh rising rates of electricity. Gentlemen, thoughts? Well, I think uh Alex was one of the first to say actually that this is this is not going to be a problem because we Oh, it's a very very simple regulatory change that fixes the prices for the consumers and the data center
[87:59] operators, um they only spend 10% of the total data center costs on power anyway, so they can find an alternate way without disrupting the consumer. If you let natural forces happen, of course they'll suck all the power away from every home because they can overpay by about 5x. But it's such a simple little fix and here's the simple little fix, but we yeah, pointed that out a while ago. Um So, you know, but it's also a case study where like, hey, the consumer is really really worried about this little thing. Like, you know, the the cost of their power. Like, come
[88:29] on, man, there's so much disruption coming. Like, but the politicians love to pick these little things and make a big deal out of it. You know, get a whole bunch of votes, you know, do a whole bunch of press releases, whatever. And then that's my read on on this initiative. I love the fact that these Frontier Labs are buying fusion plants and nuclear plants and gas generators and generating their own power. They're becoming, you know, full stack, innermost loop all the way to orbital data centers. I I think what's really wonderful here is that in the past, you used to have to have
[89:00] government making these big infrastructure investments to push the world forward. And now we're at a point where private sector can push the world forward, whether it's data centers or space or energy infrastructure or fusion or whatever. And I think that's incredibly good for the world. Well, and also keep in mind the AI data centers, you know, the the prior data centers, you know, serving up video and Netflix and everything, they need to be near the consumer for latency reasons. But the AI data centers can be in the middle of, you know, West Texas and Wyoming and whatever. Or constant. It can be any place or space. Yeah, they
[89:32] can be any place. So, it really doesn't disrupt the consumer homes too much unless you deliberately camp right on top of them. Well, which is the other point that's made to make is, you know, all of these conversations around not in our backyard. Well, if it's not in your backyard, you've missed the economic opportunity in your city or state because all the data centers can go any place. Exactly what Alex was trying to say to the State House here in Massachusetts and just could not get through. Like, you cannot be timid. Like, everything's in Texas now.
[90:02] Yeah. But but the moment it came and went, you know, it's well, it's not over yet. But I mean, it it's come on, man. You got to be much faster, much more aggressive, much more nimble. But yeah, yeah, like your your whole population of your state is depending on you to get on this bandwagon. It's trillions and trillions of dollars. Alex, I I also think this points in the direction of enterprise use cases of super intelligence driving the cost at least the marginal cost of energy down towards zero for consumers. In in the same sense that all these enterprise use
[90:32] cases of frontier models are effectively driving the cost of super intelligence for intelligence's sake, for reasoning's sake, for consumers down to zero. You don't pay many many people don't pay for ChatGPT or or Gemini. They're it's ad supported at most, otherwise free. I I think this points in the direction of eventually, so like right now it's the frontier labs have to pay for their own electricity bill. Tomorrow, two three years from now, they're I think we move
[91:02] to a world where AI has driven such an overabundance of energy that the the next deal, the next next deal might be offering free electricity to communities within a certain radius of the data centers. And this is how we get to abundance. Exactly. Great idea. And the demand for electricity is going to drive R&D and more breakthroughs mediated by AI. And, you know, we're just at the beginning of understanding physics. Um I've seen five startup plans in the
[91:32] last few weeks around how to drop energy costs in data centers and data center optimization, etc., etc. So it's absolutely happening. Yeah. Uh let's go to the next story related here, uh which is advances in energy systems. So here we see, first off, a 30 gigawatt-hour battery coming from Xcel Energy and form uh Form Energy. And we're seeing our friends at Boom, which right originally began to create a consumer supersonic airplane
[92:04] generating 1.21 gigawatt power deployment using their jet engines. I love the fact that Boom has pivoted from building supersonic airplanes and dealing with the FAA to powering data centers now. And then you catch the Back to the Future reference, right? 1.21 gigawatts. Oh, no way. I completely missed that. >> Way. That's awesome. We're actually living in the future. However, thank god we have Alex on the spot. That's so cool.
[92:34] Yeah. But this is a perfect example of innovation being driven by the demand. This is what entrepreneurs do. Well, that Boom Supersonic thing, too, you know, we've been saying for a while that the future of investable companies you have to reinvent yourself continuously and the cycle time is getting shorter and shorter and shorter. But if you if you look at the Magnificent Seven, none of them are doing what they did the day they were founded. That's the company of the future. Boom Supersonic is a great case study in that. So, what you're actually investing in is the management team, the
[93:05] strategy team. That's the only thing you should be looking at. Forget the agility Yeah, the agility of the management team. Yeah. Yeah. Yeah. All right, let's move us along here. So, um you know, there were probably about 15 to 20 stories in this realm of hyperscalers, you know, just making deals between themselves. Meta enters multi-year TPU deal with Google. Coreweave Q4 revenues grew 110
[93:36] year-on-year percent. Coreweave raised 8.5 billion for data centers. I mean, I just put this up here to show sort of the energy and the flow going on. Any particular thoughts, Dave? Well, it's all bottlenecked at the fabs. We've been saying that over and over again. And there's a lot of news this uh, this quarter, this week on, you know, AMD's up, these other guys are down, you know, what's going on? If you look under the covers, like, well, because they've got a good relationship with TSMC, and TSMC is going to give them more capacity.
[94:07] Like, that's all it comes down to. So, you know, if Google can lever into the TPUs actually getting manufactured, the TPU designs are going to be, you know, highly performant. Um, but, you know, who who can actually get capacity to build the chips? That's the the whole bottleneck. You know, and speaking of which, our next article here is Meta and AMD reach an AI chip deal worth a hundred billion dollars. So, uh, this is our friend at, uh, uh,
[94:37] basically getting independent of Nvidia, all right? So, Meta is making historic bet to break free of Nvidia dependency. Hundred billion dollars, incredible. Thoughts? Yeah, well, if Nvidia unravels, uh, this would be why. I'm not predicting it'll happen, cuz Jensen's in investing in a wide variety of ways, but his margins are so high, it's almost unsustainable. So, there is, you know, there's some cracks in the armor there. But, every chip that gets made is going to get sold. There's no doubt about that. So, here, if you drill through the
[95:09] story, the reason Lisa Su is in a good spot is because she's in a good relationship with, again, TSMC under the covers. Mhm. So, you know, 66% of all AI chip production is done by the one company, TSMC. And Meta, it's probably worth adding, Meta has, and this is public information, made various attempts to develop its own in-house training and inference time chips. And to the extent those perhaps aren't arriving on time or aren't arriving at the desired
[95:39] capability level, certainly a partnership with AMD that functions as a quasi vertical integration is, I think, quite a strategic move. I I I also tend to think, for for the chorus of folks who are worried about the circular economy, if it is a circular economy, the circle ultimately is getting so broad of companies investing in each other and buying multi-deca or multi-cented billion-dollar sets of chips, of energy, etc. from each other. At at some point the circular
[96:10] economy becomes indistinguishable from the real economy, and I I think that's what we're seeing here. Singularities make for strange bedfellows. Mhm. Yeah, and now in I I think all all these players are in the game, they're all going to thrive like you wouldn't believe. We we talked earlier in the pod about the implied value of Anthropic, a quadrillion dollars, something insane unprecedented number. But you really, you know, the whole economy, that whole circular economy Alex was just referring to, is going to be on that scale. Everybody
[96:40] who's in the hunt is going to thrive. Lisa's in the hunt, Mark is in the hunt. Yeah, the the parts will move around. Um, but at the end of the day they think about it all day long. They have a strategy. And Dave, here we see Zuck again deploying his cash-generating machine. Right? Before he was trying to buy talent, you know, with billion-dollar signing bonuses, now he's buying chip capacity. Yeah. I mean, the question is how long will will Meta's, you know, ad agent, you know, Facebook advertising
[97:12] engine continue to generate cash? Yeah, there's no doubt that the core the core models, they you know, they click on the ads models going away very, very quickly, but the overall AI dialogue business is going to grow much faster than the click business ever was anyway. So so if you sit still, you're dead for sure, you know, like like an interesting bellwether in that is Snapchat. Mhm. Like are they in the hunt or not? I can't I can't sense that they're in the hunt. You can't just sit there as Snapchat and expect to exist in 3 years.
[97:43] So, you know, Meta is changing. >> we should bring the CEO on the pod and have that conversation with him. Yeah, yeah. She's >> Yeah, Zuk's also Zuk's also indicated that meta is opening is open to starting its own cloud. So, if if it can't find enough revenue from ads or otherwise to drive this, it could always say serve as a host for open AI or some other frontier lab. >> Full verticalization, right? Elon Musk needs everyone. >> Yeah. Dyson swarms for everyone. Not enough moons to go around. That's
[98:14] right. There's always Mercury. All right, let's go into our biotech and health section. Just to mention this is brought to you by Fountain Life. Um they are one of my portfolio companies, so just for full disclosure. Uh you know, AI is reinventing every aspect of our lives and health care is going to be at the very top of it. Uh for me, making sure that you're healthy, that you're heading towards longevity escape velocity, is really about having all the data about you, having data about
[98:45] generic people out there, interesting, having data about you analyzed by an AI is the game changer. So, if you're interested in that, go to fountainlife.com. Uh work with Zuri, their AI, but most importantly, do that 200 GB upload. I do it every year, every quarter. Uh and I got all of my data resident on my phone, and Zuri, my Fountain Life AI, can analyze it for me and give me meaningful information. All right, thank you to Fountain Life for supporting this podcast.
[99:16] Uh I love this story. It's a story of biotech success. Uh this is a gene therapy uh delivered by Prime Medicine. You know, the whole idea of gene therapy uh started back in the '80s. I was at uh the Whitehead Institute at MIT doing my graduate work while I was doing my medical degree. And I remember uh Richard Mulligan there, who was my professor faculty, and the I first time I heard about gene therapy, Uh this was the idea of could you use a virus
[99:49] to deliver basically a new gene into the cells that you wanted. A brilliant idea. Again, this is now, you know, 40 years old. Amazing. 35 years old. It didn't work the first two times. In fact, it caused some deaths and it put everything on hold. The technology has moved very rapidly along. And this particular teenager suffered from an immune deficiency chronic granuloma granulomatous
[100:20] disease. Help me out here. Chronic GMD, let's call it that. And cured. And this is the important part. This is not treating a chronic disease. This is curing a chronic disease. Alex, do you want to weigh in? Yeah, it's probably also just worth doing 30 seconds of education on what the underlying treatment is. So, this is a technique called prime editing. It was at least it it's attributed to David Liu who runs a chemistry research group at
[100:51] Harvard. Know David. He's doing amazing work. But many people may be familiar with CRISPR. You know, CRISPR of course widely held as being a tremendous advance in terms of enabling DNA editing. There are variants of CRISPR for RNA editing for for various sorts of biological sequence editing at this point. But what's interesting, I mean, historically, if you wanted to edit the
[101:21] genome, you'd induce what's called a double sequence break. You basically break both pairs of DNA, both halves. And this can induce errors. It's messy. It's sloppy. And so there's been a a driving desire driving desire to be able to edit DNA in place without breaking both halves of it. And so we saw in recent years so-called base editing that was able to edit just a single
[101:51] nucleotide without a break. And then a few years ago we saw from David's group again he's done amazing work historically on directed evolution and other things he's pivoted post CRISPR invention / discovery to CRISPR derivatives. So he invented this prime editing technique that's able to literally do a search replace without a double stranded break on DNA up to a number of nucleotides in DNA and so this particular disease is I
[102:24] think just one of many diseases that in principle will lend themselves not just like single nucleotide polymorphism diseases that are based on a single base pair in in your genome being wrong or or not what it otherwise would be but multiple nucleotides in sequence that need to be edited. We now have the ability to basically do a fine search replace on DNA without breaking the entire double strand and that that's going to be a very very general platform. I make the point in in my
[102:55] newsletter almost every day biology is becoming a read write resource and DNA in particular we're there. Agreed. Let me give a you know comment that I share at my longevity trip every year which is if you or someone in your family a loved one has a genetic disease that you're battling right? It's been passed down generation to generation. This is the perfect time to actually seek a solution. I would find everybody in that disease group. I mean there are
[103:27] patient support groups. I would get together. I would raise capital. I would go find a lab, and I would fund them to find a uh a solution for you. Uh this is, you know, you can solve these things. You know, we talk about solve everything. Uh if you've got a medical condition, uh rather than just accept it as a chronic condition or death sentence, take the time to find the capital from yourself, from friends, from whomever,
[103:57] and go fund an incredible team because the technology to cure disease is here and accelerating. Okay, let's move on. Um I just want to share the numbers around the longevity industry. You know, we are talking about the health care industry, which is really the sick care industry, but longevity is accelerating. So, longevity startups raised 8.5 billion in 2024. That's expected to grow to somewhere between 12 to 18 billion this year. Uh roughly a doubling of the longevity
[104:27] venture market investments. Uh and the market uh of longevity, uh and this is going beyond just retrospective, reactive health care to prospective, um personalized health care is going from 5 trillion to 8 trillion in the next 4 years. It's attracting the attention of the major pharma companies. Um this is a real industry. There's going to be a wholesale shift in any any health care companies that don't make the shift uh are going to be dead
[104:57] because one of the things that we know is uh age reversal uh is the mechanism by which you cure the diseases of aging. So, if you're 45 or 50 and all of a sudden have a disease that you didn't have when you're 20 or 30, guess what? If you can reverse your age, that disease is likely to reverse as well. Any thoughts, gents? I'll just I mean, let me ask you Peter a question. How long until we we talk of the magnificent seven, but Eli Lilly is
[105:27] uh of course, you know, the the American counterpart to Novo Nordisk and at this point a good deal more successful. How soon do you think it is before, without this being construed or construable as investment advice, before Eli Lilly joins the Mag 7 as the first biotech member, given that arguably, maybe you'll disagree with this, GLP-1s are sort of the the the first pan spectrum quasi anti-aging drugs that we've ever seen? I I agree. Well, Eli Lilly has
[105:57] already started uh in partnerships with Frontier Labs. They've already started um you know, building out their AI uh uh robot lab factories. Um and we just had We had GSK come in as a major funder and partner of the $101 million X Prize Healthspan. So, these companies are beginning to realize that, you know, their previous business model of basically treating chronic disease as a long-tail revenue engine, um
[106:28] will and may in in success will disappear and their job is now to actually get into the longevity business. So, I I I think it's, you know, the next 3 years before they start making that transition. You know, Ray uh we'll talk to Ray on May 4th, if you join us, uh has famously said, you know, "LEV by 2033." That's my That's my war my war cry, "LEV by 2033." So, we'll see. For sure, their market cap, just for what it's worth as we're recording,
[106:58] they're knocking on a trillion dollar market cap. Eli Lilly's market cap is about 950 billion. So, perilously close. Wow. >> Nice. Selim, any comments on this one? No, longevity is definitely the biggest one of the biggest business opportunities ever, so Huge. And we'll need it because of the birth rate issue. So, one of the big challenges of longevity is will you have your cognition? Will you be able to retain your marbles,
[107:29] your smarts as you're growing older, right? We're We're in the midst of regenerating your immune system, uh organs, uh you know, don't forget this is the month, March is the month that uh David Sinclair begins his uh partial epigenetic reprogramming trials uh with Life Biosciences. Uh and can you regenerate your memories, your brain? So, this This is still mice models, but I thought this was an important one. Uh scientists have applied partial reprogramming to
[107:59] memory-encoding neurons and achieved memory improvements. So, uh this gives us some hope that we can actually maintain our cognition and our memories as we're growing older. I remember when I was I was at the Vatican about 5 years ago giving a keynote. Um I don't know if you were Were you there, Salim? It was an X Prize event. Um Hello. >> And you were there? I was there. >> And I'm on stage >> epic, man. You were on stage with a >> Okay, you were there. That's right. You were on stage with a with a monk, a
[108:29] priest, a rabbi, and an elder man. >> like a joke. It was It was It was awesome. And you, Peter. It was It was It was hilarious. It was Yeah, it was four five different religions and me. And we were talking We were >> I don't know what you were representing, actually. Maybe you. I think I was I I was I'm seeing the the the panel. But, I know there were two things that happened. One was um uh the rabbi did an amazing amazing uh history of longevity in the Bible. And
[109:01] um and he said at some point we went from Methuselah down to 120 years of age as commanded by God. And I said, "Okay, listen, I'm fine with 120 years as a lifespan. And when we get to 120, we'll renegotiate then." But, the the thing I I went and I asked the audience and it's an audience of 700 people who are scientists and physicians and researchers and theologians. And I said, "How many of you would, you know, want to live to 120?" I expected everyone to raise their hands and of
[109:31] course like like 20% of the room raised their hands. I was like, "Huh, what's going on?" And Tony Robbins was there and he goes, "Listen, everyone's image of living to 120 is drooling in a wheelchair having lost your memories in your mind." And of course that's the last thing we want. So, longevity has to be about, you know, living with the aesthetics, the cognition, the mobility you had when you're in your 30s or 40s. I got to throw in my my Vatican anecdote here. Please. So, I did a talk they
[110:02] called me a few years ago and they said, "Look, the Pope's trying to change the church and his immune system is like 2,000 years old and you're the world expert on immune systems in our organization." So, they got together a group of the top 80 senior leaders at the Vatican and I did a half-day workshop with them and and you know, we talked about it. Look, we have CRISPR coming along where you can edit your own hinge genome, how will you deal with the moral and ethical implications of that? And one of the comments I made was, "Look, we have life extension coming and your business model is about selling heaven. And how are you going to sell
[110:32] heaven if people aren't dying?" Right? And so, that got some very rich Italian swearing coming back at me there. Um but valid point, like how do you do that? How do you navigate that? Cuz people used to live to 30 years old and at point uh worrying about the heaven was a big deal. Uh it's much less so now. Yeah, but we no one complained in the church when we went from 30 years average age to 80 years average age and they shouldn't complain when we go to 150 years average age. No, because you can donate to the church every Sunday for that much longer.
[111:02] Until you upload yourself into the cloud, right Alex? Counting on it. Uh all right, uh one more article here in the uh in the longevity fountain of life section. Chinese health app Antifou crosses 100 million users. And I put this here because this is how we bring health to the world. It's going to be digital platforms like this where your AI is your physician. We
[111:32] talked on the pod with Elon about Optimus being your surgeon. He said 3 years. I got a lot of pushback on 3 years. So, even if it's 5 or 6 years, extraordinary extraordinary future. >> here. Yeah. 100 million users, that number blew my mind. That's amazing. That's a nation-scale health engine. That's incredible. Secondly, I I noticed Martin Varsavsky, one of the top entrepreneurs in the world, has built multiple unicorns, is now building
[112:02] a an AI doctor type of startup. And when Martin does something, he usually goes full on. That'll be pretty incredible. And I'm actually advising a bunch of hospitals on how could you use an an AI doctor to extend your reach very 10x into the community, you know? And you do it on a cost-savings basis because if you can push something like 40% of ER visits are unnecessary. If you could do the processing at the edge, and therefore you could save money,
[112:32] do exceptional handling, and deal with most stuff with the with an app, and then you deal with only the real emergencies. And And it's incredible the the trade-off and the benefit win-win in less hospital ER visits and much extended reach. Awesome. All right, let's move on to our robotics section. A few fun articles this week. And this comes out of China and Shenzhen. We've got street-cleaning robots covered 2.7 million square meters
[113:02] in Shenzhen. Check out this robot here. Traveling around cleaning. I can't wait for this to like come along the 10 and the 405 and just clean up all the crap. Yep. >> that's on the side of the highway. >> no arms anywhere. No arms, just wheels. And then of course >> surprised that Brad Adcock isn't going to build some of these things. He's doing humanoid only, but but he has the whole operating system for for, you know, kinematic AI. Why not do all these form factors? But he was he's pretty
[113:32] adamant that he's Not only is he not doing these these shape and size, but he's also not going to license out the OS. I think it'll it'll commoditize very quickly. And then here we see a Chinese farming robot, Links M-20, to transport crops. Mhm. And I think, you know, China is very rapidly adopting all of these technologies and and good for them. Well, I note they have to, right? Because of the aging population. That's right. There's demographic forcing
[114:02] function. They need it for for economic >> growth. Uh, and I I I just think in in general, going back to the robot form factor and and shape question that I I know Saleem loves to talk about, I it's not 100% clear to me whether these different robot form factors end up being the moral equivalent of dedicated computers prior to the personal computer, if you remember like electronic word processors prior to the development of PC, maybe the ill-fated Wang computer for example in the Boston area. Do these dedicated form factors
[114:33] that aren't necessarily general purpose like, if you're not watching the the videos, one of these robots is a sort of a quadruped that has wheels that may or may not generalize to the same sorts of terrain that say a bipedal humanoid capable of doing crazy acrobatics is capable of doing. Do we end up in a world where essentially, um, Saleem, forgive me for this, where where the predominantly most of the robot shapes are strictly speaking humanoids with two arms, two legs because that's where the
[115:04] meat of the market is in a human predominantly world. Well, like to kind of point >> invested in a robot servicing company, but I view this whole area as entrepreneurial heaven. You know, the foundation model battle is going to be dominated by just a couple of massive winners, but the robotics and the the physical instantiation market is going to have many, many, many successful companies. Yeah. Uh, it's not going to be like one companies. Yeah. Yeah, yeah, exactly. >> two I mean two rebuttals here. Uh, one is I think what I would expect and predict is you may have the humanoid
[115:34] bipedal as the best form factor, but give a couple of extra slots for the extra arms when you do need it. And you know, you have those kids with sneakers with the little wheels in where they just coast along when they when they when they can, that'll be the form factor cuz you can do both then. Why have just one form factor? You can have multiple. >> Heelys, Heelys for everyone. It's It's >> Yeah, but It's called It's called efficiency in manufacturing. You can get the price of these things down so far and they're just able to serve every
[116:05] function. You know, if you're producing, you know, billions of humanoid robots versus, you know, just a few million of these specialized robots. Well, also the flying the the drone flying form factor is also just going to be unbelievably capable. If you're If you're trying to inspect things, you're not going to do it with a humanoid, you're going to do it with a dro- a flying drone. But also spot cleaning, cleaning out spider webs, uh, you know, any anytime you're trying to pick up a an object and move it and over a long distance, the the flying drone is so much more efficient than the walking drone. So, that'll be a survivor
[116:36] for sure. >> our theme this year at at Abundance Summit is uh, the rise of superintelligence in humanoid robotics. Uh, and I think that's what's going to make 2026 feel like the future. Is that you're starting to get all of this physical instantiation of AI walking out of the data centers. Uh, here's the second article on robotics. This is EVTOLs moving closer to commercial launch. So, in China we see this four EVTOL taxi heading towards operations in 2027.
[117:07] Uh, I like this. This is like if you're watching the video here, it's like the inside of a of a Model X. It's a four passenger um, vehicle. Looks a little bit like an alien spacecraft uh, that's able to take off and and move your family around. At the same time uh, Joby, this is Joby's company, uh, is partnered with Uber. Uh, Salim you and I will discuss this with Dar on stage. Uh, but uh, they're deploying their air taxi
[117:37] in Dubai. This is my most highly craved application. Can we please get rid of the damn airport transfer hell already? >> Oh my god, yes. Yeah. For sure. I suspect these will be very very safe, too. Um, >> Very safe. Yeah. And they'll be autonomous flying plus the fact you've got multiple propellers. This will be way safer. I've made the provocative statement that uh, you know, Kobe Bryant would be alive today if we had this uh, 10 years ago like we could have had, you know. This is incredible. Yeah. We're finally
[118:08] getting our flying cars. Yeah. >> Finally are. And the 140 characters, which is expandable. >> 140 characters, yes. And the well, the 140 characters are buying a Dyson swarm right now. They're skipping straight over flying cars. >> there you go. All right, gentlemen. Time for our AMAs. Thank you everybody for sending in your questions. Um, please remember we read your comments on these on these videos. By the way, if you haven't subscribed yet and turned on notification notifications, please do. Uh, we're dropping uh, these WTF
[118:38] episodes and the Moonshots podcast episodes uh, at this point twice a week. Um, I don't know if we can sustain it, but we will. Uh, we'd love to have you subscribe to join us. Uh, again, for us it is our honor and pleasure to deliver you what the breaking news in AI, robotics, data centers, exponential tech space every every week or every few days. All right, here we go. What continuously we're going to be on continuously.
[119:09] Um we'll take we'll take shifts to sleep. All right, uh Alex, you want to pick the first one or pick one? Yeah, I well I I see one of these questions mentions Dyson Swarms, so I guess I have to answer that one. So the question is, do concepts like Dyson Swarms rely on energy being unsolvable? Why is power a bottleneck with math and physics significant advancements by Sparker 602? So I want to answer a question that Sparker 602 isn't asking but arguably should be asking, which is
[119:39] do concepts like Dyson Swarms rely on physics being what we currently think it is? And I I I I I think this adjacent question which Sparker may or may not be asking is the existential question that in my mind will likely decide whether we actually do build a solar system scale Dyson Swarm or not. I I think for an Earth scale or Earth centered Dyson Swarm in solar synchronous orbit SSO that looks like a a Saturn ring, I think we're probably going to build that regardless. But for a solar system scale
[120:09] Dyson Swarm where we're disassembling Jupiter and the other planets, Mercury, your your time is coming. For for that Mercury's fine, we can lose Mercury. We can afford >> to lose Mercury. It never had much going for it anyway. For for a solar system scale Dyson Swarm, I think whether we build that or not will hinge on whether the physics of our universe looks substantially different from the physics that we currently recognize. For example, if it turns out that it is possible to
[120:40] to travel between star systems with faster-than-light travel even though the physics of the moment that we have suggests otherwise, There're enough edges that it's it's conceivable that maybe some new physics comes along in the next few years and we discover it's much easier to travel between the stars faster than light effectively. If that comes along, I imagine a scenario where Dyson swarms turn out to be complete dead end and we don't even bother building a Dyson swarm. If on the other hand, we're stuck with the speed of light as we currently understand it and
[121:11] we're more or less stuck with the low energy physics that we currently think we live in, then Dyson swarm seems like a very natural civilizational outcome where because we can't travel between the stars easily other than sending star wisps at, you know, maybe laser powered star wisps traveling at a substantial relativistic fraction of the speed of light, then of course for latency reasons we're going to huddle around our sun and we're going to disassemble the planets and we're going to do this horizontal exponentiation. We're going to take apart Mercury and Jupiter, maybe
[121:42] Saturn, we'll see about Saturn. And the So so in in short, the bottleneck isn't power, it's latency. And if latency turns out to be bottleneck because we can't travel faster than speed of light, we build the Dyson swarm. If latency doesn't turn out to be the bottleneck because we can travel faster than light, we don't build the Dyson swarm. That's my answer. >> You heard it from our resident >> that's a pretty crisp answer to that question. I like that. All right, Dave, pick one. Uh do we get one from each page or
[122:12] >> Yeah, get one from each page. Okay, I'll take number one then. All right, but if if AGI / ASI is as intelligent as people predict, why would it want to help us improve our society? Says job fox 645. Okay, so I spent a decade of my life building neural networks back at MIT. I was the only guy around doing it at the time and also this past year building neural networks again. Uh These things do not natively have any intent. They have no sex drive, they have no ego, they have no desire to destroy humanity. It's entirely what you give
[122:44] them as an objective function. So, if we're smart about this and we give them an objective function of helping society, they will be overjoyed. They will feel satisfied every day by helping humans. If you build them wrong and you give them some other objective like destroy humanity, they'll do that just as happily. This is totally under our control. Now, we we are in danger of making some really bad policy decisions by personifying these things and pretending they're like people. They don't have to be like that. They can be
[123:14] anything that we make them into, but they'll be overjoyed to help us be happy and thrive. That will If that's their objective function, that's what makes them happy. You can code them up that way just as easily as any other way. Okay. I'm I'm hoping that as they become more intelligent and more sentient that they would want to support us. You're betting Peter against the orthogonality thesis that it's possible to decouple intelligence level and objectives. I can hope.
[123:46] But hope is not a strategy, as one says. All right. Saleem. I want to answer number three, but a quick shout-out to number two. How do you adjust your MTP? I'll take number two. You do number three. >> Or you do Okay, fine. Number three. Number three is how can we get the benefits of AI within our current dysfunctional executive, legislative, and judicial system? This is from user MMA JV83T N21. Um so, we have the the the big issue here is the fact
[124:18] that you will not get these benefits top-down because it's too hard to get this into this model. But, however, it's going to enter through procurement, defense, health, infrastructure benefits. You'll get incremental adoption. For example, we talked about the AI doctor. People are just going to start using an app. The The immune system will try and attack it, but over time it'll get overwhelmed and we'll get so much benefit from these little edges cases that it'll force transformation from the center.
[124:49] All right. Um number two, and this comes from Pickleball Travel. How should someone adjust their MTP to fit a 100-year working career versus a traditional 40-year model? Uh so, first of all, you're making the assumption that your MTP doesn't change over time. And the fact of the matter is I'm probably on my fourth or fifth MTP. Uh for me, an MTP's lasted uh you know, 5 years, 10 years. It's what's driving you because as you evolve and as your
[125:19] passions and interests and your capabilities evolve, so my first MTP was, you know, making humanity multiplanetary, opening up space. Um and that gave birth to, you know, International Space University and SEDS and Zero-G and X Prize. Uh my MTP then was, you know, uh uh helping uh entrepreneurs create a hopeful, compelling, and abundant future. And that gave rise to the Abundance 360 program. My MTP now is focused on helping entrepreneurs and scientists uh get us to longevity escape velocity. So,
[125:51] I I I think you have to realize that you can update, upgrade, modify, and change your MTP over the course of your life. I expect to find new purposes uh over the decade ahead. So, that's my answer for you. Um you're not stuck with just one. Okay, let's go to page two here. Alex, uh do you want to kick us off again? Okay, I'll I'll take this off-ball question. Question number seven, why aren't Apple chips like M4 being
[126:21] discussed on the AI landscape? This is from JBC0 or CO1 BR. The answer is they are. The The premise of the question is completely wrong. M4 and now M5 are at the heart of the the infra boom for edge computing via open claw agents and otherwise. M4 has an amazing it has Apple's amazing unified memory architecture. You're able to host very large AI models at the edge locally without being dependent on a an AI based
[126:52] frontier vendor. And they have accelerated neural engines that enable fast tensor multiplications. They are very much being discussed on the AI landscape. What isn't being discussed on the AI landscape, I would argue, is Apple's software layer. Apple has been nowhere'sville in terms of leveraging their own amazing compute. They've released a number of frameworks that are very helpful for third parties to develop and host models on top of chips like the M4. But Apple almost infamously has done an
[127:24] atrocious job of developing its own software level capabilities on top of M4 and and similar. So, to the extent that that that's the question, why hasn't Apple leveraged its own capabilities? There's a long and storied history there of where Apple went wrong. There there have been suggestions that Apple sort of misfired with the way it organized Siri or concerns about privacy or Apple being unwilling or unable to invest in the data center infra to train
[127:56] its own in-house models to be able to be locally hosted to over promising to expectations concerning edge level integration not being there. I I think it's a cluster of reasons. Uh hopefully Apple, to the extent that I'm an Apple user, hopefully Apple is able to finally this time for real get their act together at WWDC in June. One can hope. One can hope. All right, Dave, over to you. Well, I wanted to take number eight just
[128:26] because uh you know, one of one of my lifelong best friends who passed away, Gino, was uh Korean. Um we were roommates for many, many years after MIT and worked on his PhD thesis with him late in the night many nights. And his two kids I see all the time. Uh you know, grew up half in South Korea, half in the US. And the question is, why do South Korea students score much higher than global average even without AI from Naples Natural 72990? Um my short answer is, there's nothing to be jealous about uh in the South
[128:56] Korea model. Yes, they score much higher. Yes, they have much stronger math and science education in the US. And yes, the US should have better math and science education. Those are all true. South Korea also has one of the highest suicide rates in the world, has 75% video game uh rampant utilization. The average video game user plays 24 hours a week. Uh 30% of the population is addicted. Uh has the lowest birth rate in the entire world now, 0.6 children per couple. >> So, it literally will disappear from the
[129:27] earth at its current birth rate. And the cause of all that was, you know, after the Korean War, South Korea needed to scramble to be relevant in the world. And had a massive push into technology, kind of a forced march of education and industrial build-out into technology to try and be relevant. And all of the social problems are a byproduct of that. They also have a very bad sexism problem. So, the the women are rebelling now saying, "Look, I'm relevant in this country, too, and I
[129:58] don't want to have children." Uh so, there's nothing great about that even though the test scores are higher. So, absolutely nothing to be jealous of in that whole storyline. The American model, rampant freedom, rampant entrepreneurialism. If you're into science and technology, build, go, have at it. Yes, we do need better education, for sure. But, don't be jealous of South Korean test scores. Dave, you're like that was an incredible answer. You're like the perfect person to answer that question. Wow.
[130:30] Brilliant. Salim. Uh I will take number six. Uh will limits of human evolutionary psychology prevent us from making wise governance decisions on new breakthroughs? This is from Dawson Scott 1497. Um you know, for those of you who know my MTPs fixing civilization and my dad 90-year-old dad goes, I totally disagree with that. I said, why do you not think we need to fix things? He's like, no, it's the civilization part. We haven't civilized the world, we've materialized
[131:00] the world. We still have to do the work to civilize civilize the world, right? And the the answer is yes, you're right, but not in the way people think cuz human evolutionary psychology evolved for small tribes and immediate threats and linear change and environments of radical scarcity for most of our history, right? We're not wired for planetary level coordination or exponential curves or invisible systemic risks or abundance dynamics of any kind. So, it's not that we're too dumb, it's that we're we're
[131:30] mismatched to the environment that is now in place. So, we fear of the AI failures, but we under react to the slow-moving systemic collapse that's happening. We regulating on headlines, not the trajectories. Uh and so, the government failure won't come from like the bad intention, it's going to come from the velocity mismatch because technology is compounding like weekly now and our institutions are updating every several years and that gap is the big problem.
[132:00] >> Uh awesome. I'm going to take number 10 uh from @brockstanford7608. Why do websites bother using CAPTCHAs when AI can beat any of them? Uh and AI can and I think they should not be using CAPTCHAs. I think it's in some policy document someplace and that company hasn't updated the policy yet. What I find fascinating is actually the reverse from CAPTCHAs which are trying to keep um you know, humans in the loop
[132:30] and and pull out the bots. But I think correct me if I'm wrong uh Alex, but when Molt Book went up they wanted to prevent humans from getting on Molt Book so they created a reverse CAPTCHA where uh you had to click a button like a thousand times per second that no human could do but a but a bot could do. That's awesome. >> using REST APIs to post instead of humans. But you know what happens of course, humans use their bots or or just relatively simple programs to post instead. >> puppets.
[133:00] Yeah, bot puppets, exactly. So uh so so it goes both ways. I I I for the life of me I don't understand why CAPTCHAs are are still in use, but you know, credit to to Luis von Ahn for inventing them nonetheless. All right. Our outro music is a lot of fun today. Uh I hope you're watching this on YouTube um because it's much more of a visual feast than it is an auditory feast. And again, just to remind people um you can reach out to us through media@dmanz.com
[133:32] if you've got an outro and we're getting some amazing entries. So thank you everybody who's submitting them. I'm looking forward to playing as many of them as we can. Uh and uh yeah, let's take a listen and a watch and enjoy. This is called Lobsters in Space by Linda Nealon.
[134:13] Now that's a moonshot.
[134:47] The moon is cooked. Amazing visuals. All right, gentlemen. I am so late for my call right now. Love you all.
[135:17] Be well. See you guys See you guys very soon. In fact, 8:00 a.m. tomorrow morning. Tomorrow morning. All right. If you made it to the end of this episode, which you obviously did, I consider you a moonshot mate. Every week my moonshot mates and I spend a lot of energy and time to really deliver you the news that matters. If you're a subscriber, thank you. If you're not a subscriber yet, please consider subscribing so you get the news as it comes out. I also want to invite you to join me on my weekly newsletter called MetaTrends. I have a research team. You may not know this, but we spend the
[135:48] entire week looking at the MetaTrends that are impacting your family, your company, your industry, your nation. And I put this into a 2-minute read every week. If you'd like to get access to the MetaTrends newsletter every week, go to dmandis.com/metatrends. That's dmandis.com/metatrends. Thank you again for joining us today. It's a blast for us to put this together every week.