Go back

The NVIDIA Scandal Goes Deeper Than You Think | Prof St Onge

1h 11m 37s

The NVIDIA Scandal Goes Deeper Than You Think | Prof St Onge

Peter St. Onge joins the podcast to discuss the resignation of NVIDIA CEO Jensen Huang and its implications for the AI industry. St. Onge expresses concern that Huang's departure threatens the push for open-source AI, which he views as essential for preserving liberty against both government censorship and corporate oligarchy. He warns that without competitive, government-independent AI, voters could be brainwashed and freedom could be lost. The conversation explores two central debates: open-source versus closed-source AI, and whether to sell chips to China. St. Onge notes that Huang advocated for open models as an alternative to the dominance of Anthropic, ChatGPT, and Google. On China, he acknowledges the tradeoff between denying chips to slow Chinese development versus incentivizing their domestic industry, comparing it to the Cisco-Huawei dynamic. St. Onge, a libertarian economist, argues that government is a gang of criminals that destroys lives and wealth. He advises young people to pursue skilled trades, human-to-human services, and entrepreneurship in the near term, as AI commoditizes expertise but not human connection. Regarding AI's economic impact, he believes this time differs from the dot-com bubble because the appetite for intelligence is nearly unlimited, though model commoditization limits lock-in. He recommends Bitcoin as a hedge against inflation and government overreach in an era of massive debt and potential monetary crisis.

Transcription

13237 Words, 72035 Characters

English
Speaker 1The people, the voters, are going to be brainwashed. They're going to be sheep being let off the cliff. That's what terrifies me about AI. If China takes over AI, give it 20 years, and we don't have a military. They're floating Tim Cook taking over NVIDIA. It's extremely important that you can have competitive AI that is free from government control, and that's what Huang was pushing for, where you are not locked into the oligarchy of Anthropic, ChatGPT, Google. You gotta play along with the regime. I view the government as a gang of criminals, destroys lives, it makes us poor, it makes us into slaves. You are in a system that is inflating and is starting to flirt with hyperinflation. If you think that it's going to be a massive explosion, then you will want to hold something that governments absolutely cannot control, which is the premise on Bitcoin.
Speaker 2All right, Peter St. Onge, thank you so much for joining us today. Thank you for having me on. Absolutely. It is a monumental day, historical day in markets. We just saw Jensen Huang, the CEO, founder of NVIDIA, resign. From the $4 trillion company that he started himself. What's your initial gut reaction to this news?
Speaker 1Okay, I'm biased. I like Huang because he has been the biggest voice pushing for open source AI, right? Open models where you are not locked into the oligarchy of Anthropic, ChatGPT, Google. I trust Elon. I don't trust the other three. And Gronk is like 5% market share, right? So realistically, it's essentially an oligarchy. And Huang had been fighting that really hard. He'd been building chips, specifically aimed at the open weight part of the industry, LLMs. He'd been funding open source. So when it first came out, my first reaction was, this is BS. The media cooked this up. Some anonymous whatnot. Because I wanted the support, bro. And, you know, it turns out it was not made up very much. I think there were a lot of people surprised that the board kicked him out. Or anyway, I guess they allowed him to resign. And so, you know, people on one side are like, it was too, you know, it was too early. Tech bros sleep with women all the time. Who cares? And then there are guys on the other side saying like, you know, they shouldn't have let him resign. They should have, you know, beat him on the way out the door. The big question, you know, there's two questions at this point. Right. One of them is NVIDIA is, I think it's the most valuable company in the world. More important, it is very, very important in AI. Right. They are by far the leading, like they have the best technology. That's the reason they're so valuable. And shockingly, even though he's Taiwanese American, he, you know, was, I mean, if you hear him speak English, he's completely American. He's a citizen. And he is, I think, a big reason why, uh, the semiconductor side of AI, uh, I mean, really the entire supply chain on the, you know, chips that run AI on the data centers that's dominated by America, which is pretty shocking because America is so weak in manufacturing in general, uh, that it's been really impressive. And then we combine that with the fact that we have so well capitalized LLMs, right. The models like chat GPT and Claude, you put that together. And we've, I think sort of against all odds, America has been dominant. In AI, China or, um, Europe was, I mean, they're, they may as well be Mexico. They're they're done. Um, but China, China, um, you know, you would expect that a technology that's based on semiconductors is going to be dominated by China. In fact, it's not, uh, now that's where Nvidia comes in, right. Because apparently they were selling to a company that where, you know, one of the senior executives was the spy that Wong was sleeping with. But more important, they were channeling, uh, that company. It's a Singaporean company. It was channeling seven and a half billion in chips that was reselling into China. This is the state of software. Thank you, Apple for the thumbs up. And the issue there is, so I started my career working in a big, um, telecoms infrastructure company. And you as a marketer, you know, how much demand there should be for your product in each country. Okay. And if Singapore has an extra seven and a half billion dollars of chip demand, you know, damn well where those chips are going. So I think that's the larger issue here is, you know, Wong has been, um, very friendly with Elon, uh, he has been friendly with Trump because Trump's overarching goal is that he wants the jobs that AI bring, but he also wants the US to be dominant in AI because he correctly believes that if we lose dominance to China. It is an economic disaster is also a military disaster. We will be like Zulu fighting the British, right? If China takes over AI, if we're shut out, if we're like Europe, give it 20 years and we don't have a military. So I think it is critically important to be in control of AI. So at this point, it's almost like if you've seen the movie, Lord of war with Nick cage, all right. And you know, he gets pinched and then he's sitting there with Ethan Hawk. And he says, look, yeah, you have me dead to earth. Dead the rights to all these crimes, but it doesn't matter. All right. Somebody is going to make a phone call and I'm going to, you know, I'm going to be walking in five minutes. That's the question for me is that it is so critical, not only to the economy where you don't have those kinds of phone calls, but in the military, you absolutely do have those phone calls. So that's, that's my question is that I love NVIDIA as a company. I love what it's done. I love Huang as an individual. Okay. But there is a, I think it's pretty clear that NVIDIA knew what was happening. I don't want to ascribe any motives to any individual partly because of libel laws. But anyway, it seems pretty, it seems a live possibility. But on the other hand, the question is, do the TLA's intervene, you know, just the department of defense doesn't come in and say, look, we're gonna have to let this one
Speaker 2slide. There's so many levels of intrigue here. And it seems as if there's three fighting factions, right? There's the military industrial complex, there's the finance side and the tech side. And when it comes down to the debates, there's really two fundamental debates that we're dealing with here. There's open source versus closed source, Huang being a huge advocate for open source versus the labs that want to keep their secret sauce. And then there's the selling chips to China versus not selling chips to China. So let's break down these two different sides. And you started off saying that you're a big advocate for open source. So let's talk about that. What is the difference? What is Huang's philosophy when it comes to open source versus say, Dario?
Speaker 1Yeah. So I'm a tech retard. So anybody listening should discount half of what I say. Okay. However, my understanding is that what open source does is that a company, they can buy the chips, they can essentially have the AI in house, and then they can decide what the parameters are, what the biases are, if you will, on that AI. This, this strikes me as existentially important for not just for the economy, but for liberty, you know, thank God for the sensors during COVID, because they did a lot of things that normies thought was kind of hyper. Hyperventilating black helicopter fantasies. And they did it, dude, we were there, right? We were, well, we were giving pregnant women untested vaccine. I mean, just, we were doing crazy shit that I think people, um, didn't think would happen in our lifetimes. So thankfully that gave us a very, you know, valuable, uh, wake up call. And you're already seeing it where people are not thinking anymore, right? They're outsourcing at the AI now, 90% of the time, that's an improvement. Okay. Because AI is pretty smart. Like I love watching. Now you have all this AI slop on YouTube of like, you know, what did people eat in the medieval era? And guess what? It's a hell of a lot better researched than people did. So I enjoy it. I know it's slop. I don't care. It's my guilty pleasure. So that's fine. The issue is on politics, right? So you remember, uh, when Google, okay. For a minute there, when Gemini came out and, you know, you would search for like Pope's and every single Pope was black, and then he searched for a certain military during the 1940s and everybody apparently in the Vermont was not white, which, I mean, that was a surprise to me at the time. So, you know, that's, that's what we're up against. And currently I think the AIs are playing nice because there's a cop driving the rear view veer, uh, named Trump, but once he's out of power, which given the midterms looks increasingly likely that, um, you know, JD or Ruby are not going to be president next, uh, at that point, the sensors are going to town. The reason being that they are, they have to, right? Like Mark Zuckerberg, I think is a great example. Um, or, uh, what's his name with, um, Jack, um, Dorsey, okay. At Twitter, both of those guys early on, they were free speech absolutists, right? They were nut job and cap, let them speak hate speech. What is hate speech? Right? It was beautiful. It was a golden age of free speech and what happened, of course, the activists got to them, right? So the so-called deep state, which includes regulators, you know, they can come. I mean, they were doing that at a Tesla, for example, right? They would show up. I remember they had some lawsuit because Elon had built a glass house in one of his factories and then he was sleeping in it because Elon, and they actually came out of, and they were like, well, you know, you gotta be permitted for a residence. And, you know, is this, is this, um, expropriation of corporate assets for personal gain? And that's the shit they pull, right? So in that environment, OK, and well, that's part of it. And then they were like literally firebombing corporate helicopters, you know, corporate headquarters during the BLM riots. So in that environment, even if you're Mark Zuckerberg, who loves free speech, you got to look out for your workers, for your shareholders. I mean, it's something for, you know, your own legal jeopardy. You got to play along with it, with the regime. So once Trump is out of power, if we have a Democrat president next, it doesn't even matter if Congress flips back because the president Trump has shown us can do quite a bit. So all of those guys are going to flip back to censorship and, you know, free the social media is going to flip back. But more important in this context, AI is going to flip back and we're going to be back to, you know, everybody in the Wehrmacht is from Congo. And in that kind of environment, the people, the voters are going to breathe. They're going to be brainwashed. They're going to be sheet being let off the cliff. So I can't shit on Twitter because I sort of knee jerk defend AI. I mean, partly I like it because I think it's going to make us immortal. I think that we're hitting the Ray Kurzweil singularity. But aside from that, that's what terrifies me is that if we don't pull off AI right. OK, if either China gets control of it or if the, you know, progressive censors get control of it. I think we are. I think we are doomed. I think liberty is doomed.
Speaker 2And these open weight models would allow it so that not all the answers come from one of two sources, right? Either chat GPT or anthropic.
Speaker 1Bingo. Exactly. So at least you have an alternative voice out there. You know, one of I think a nice example. So China censors the Internet, of course, and they have their own version of Twitter called Weibo and they censor that, obviously. But there was this fun moment where so somebody put out a tweet on Weibo. And it was President Xi as Winnie the Pooh. So in the grand scheme of the universe, this is not this is this is not regime threatening. But the thing is, it got like 200 million views in like 10 minutes or something. And of course, what what that was really doing is it was telling people you're not alone. And that's why it's so terrifying for the regime, right? The regime wants all Chinese users to think that there's a couple of nut jobs who, you know, barefoot live in the woods. Who oppose the regime. But everybody else, you know, is a supporter. And so moments like that get it out. That's what I'm interested in. Right. So as long even if it's, you know, a crushed voice where you have like the equivalent of pirate radio, I think it's extremely important that you can have a competent, competitive AI that is free from government control. And that's what Huang was pushing for. Now, of course, there's going to be the Candace Owen. I don't know that she knows much about AI, but anyway, there's going to whatever that. An analogy is an AI. And they're going to say, no, no, dude, dude came out for open source. And next thing you know, you know, cooked up a spy. You know, that's coming because it's the Internet. But at any rate, I do think it's extremely important. You know, whoever picks up the torch next, they're floating. Tim Cook taking over NVIDIA. Tim has not been a huge supporter of free speech in the past couple of decades. So we'll see what comes next. I wish it hadn't happened. But, you know, give. Given the legal questions, what did Huang know? There's speculation. He made a phone call asking DOJ to lay off this woman. So it looks pretty ugly.
Speaker 2Yeah. And just to keep it on the open source, closed source debate. So, I mean, it benefits his business model, correct? To be open source, because that way he can sell as much compute as possible. Whereas with these closed source models, they don't really care about the sales of compute. They just want to make sure that they control it.
Speaker 1Which is a great point, which suggests that whoever takes over NVIDIA next, is going to be, you know, similarly enthusiastic about open source.
Speaker 2And then the other side of the debate is whether we should sell these chips to China at all or not, right? So right after this news broke, we got a letter from a senator saying that we got to reassess everything that we do in terms of chips selling to China. We need to shut it down. Some people say that that gives an advantage to China because it's going to force them to learn how to do it by themselves. And when we sell them a little bit, then we can keep them on the hook. Where do you land in this debate?
Speaker 1Yeah, it's an interesting debate. And you do have both sides, right? So if you deny chips completely to China, then it might slow them. And then the question, but as you say, right, it also invites them to develop domestic industry. So this is exactly what happened with Cisco and Huawei. And, you know, Huawei was a joke company compared to Cisco. Cisco had the top technology and networking. And Huawei at this point is completely competitive with Cisco. And you can no longer control. You can no longer control them. You know, given how much easier it is to manufacture and to innovate in China, they don't have the climate gang. They don't have a whole bunch of stuff, political constraints. Once they get to the leading edge of something, it's very, very hard to recapture that. So there's that argument. And, you know, you could say that the next couple of years are pivotal. But, you know, AI is probably about a 20 year race here. And so any particular couple of years could be pivotal. Now, I mean, if you zoom out, you know, broadly, ideologically, I don't like the idea of sanctions. You know, for the same reason, I don't like all the wars that we start. But having said, you could. It's one area where I'm kind of open to the military industrial complex's lies, where in this case, yes, I think it is actually a big problem if China gets ahead of us on AI. But, yes, there is a tradeoff involved.
Speaker 2Because we're dealing with a matter of months. Here in weeks, right. The lead that we have over open source in China.
Speaker 1Yeah, I think the last consensus I've seen is that on LLMs, we're looking at something like six month delay. There's caveats to that, like the latest Z.AI, which is one of the leading ones over there. They just apparently run everything through Claude. So they sort of cheat on the benchmarks, which is cute. So, like, I don't know what the real number is, but anyway, yeah, it's something like six months. I don't know, two years. Like, in other words, it's not it's not like a Europe type decades behind Stone Age competitors. So, yeah, right. They're not that far behind. The other question is, you know, whatever that gap is, if it's six months or if it's two years, keep in mind, they've been doing that on seven and a half billion of smuggled NVIDIA chips. Plus whatever else they smuggled through Hong Kong and Malaysia and, you know, the Congo. Who knows? Where? So, you know, it's a separate question if if you could actually go through and stop all of the illicit chip sales into China, which, by the way, you can't do for the same reason that, you know, there's still cocaine on American streets, which is not produced domestically. But, you know, even in theory, just to sort of get an idea of the scale here, if you could stop all American chips or European, for that matter, going into China, what would that gap look like? I suspect it would be larger. It could be three to five years. I don't know. Again, somebody who's less technologically retarded would know better. But like, that's probably a bigger gap.
Speaker 2I mean, there's so many levels of intrigue to the story. And I mean, it's come out that the government might have known about this woman, Alice Wong, his counterpart over there in Singapore for over a year now. Right. And then it came out just last night that there was another affair with a member of the NVIDIA company.
Speaker 1Not for the course. I mean, that's.
Speaker 2But that was also known, too, for months. Right. So the question is, the classic question in Chinese, I'll say it to keep it as part of the theme, weishima shansai. Why now? Why now, a couple of days after midterms, did this all come out?
Speaker 1So that's a good point. And yeah, that gets over to the speculation about motives. You know, I mean, normally sleeping with an executive wouldn't rise to a DOJ question. But covering up or doing favors for the spy, that gets us back to the Lord of War scene, which then by extension, so, you know, why now, right? So weishima shansai. There you go. You're Chinese. I actually lived in Taiwan for five years. My Chinese is lousy. Yours is much better. But yes, that's the question, right? So what's the play here? You know, officials in DOJ threatened to leak because they see the tide shifting and, you know, they were always anti-Trump. And so, you know, now they see light at the end of the tunnel on their exile. It's an interesting question. I have no expertise.
Speaker 2And then another level is the corporate governance structure, right? Jensen started the company. He has, I think, about four to six percent of the shares, but he doesn't have super voting rights. And that is sort of like a woke thing. If you're a founder to give up those super voting rights, this could never happen to Zuckerberg. This could never happen to Elon. Why would any CEO allow themselves to be stripped of those rights?
Speaker 1Yeah, I mean, he had a very, very long runway on this. I think he started it in 1987 or something. I mean, bro's been at it for a while. You know, Zuck, he, you know, everybody knew that, you know, social media was going to be a big thing. And there were, you know, a dozen competitors. All the products were identical. Zuck nailed the financing is really his win. And so, you know, he could get big fast, but. I mean, in Jensen's case, he's been, you know, pounding away in the driveway or in the garage for decades. I think his wife pawned the wedding ring at one point because they couldn't make rent, which kind of makes all this a little bit more salient. But at any rate, yeah, I mean, I think it's surprising for a modern tech lord because the financing makes it go so fast. You know, you look at somebody like Uber or Airbnb and something. I mean, the money just, you know, just poof. You're seeing it in certain companies now, like CalSheet, for example, the prediction market company that's got value to 30 billion. Right. And probably nobody heard about them like two years ago. Right. So it's so much faster now that a tech company can get big that I think it feels strange. But I mean, NVIDIA is sort of a 1980s company, right, which is not long after Apple Computer was founded.
Speaker 2Yeah. So now he's gone and he served as a central figurehead. Talking, of course, Jensen in this whole entire AI trade when it comes to the circular financing, all the deals that were being made. Talk to us about that. Does this have the whiff of a bubble? I know that you cut your teeth on the dotcom bubble. What similarities do you see right now with the market and back in the late 90s? How many apps does it take to run your life on Bitcoin? One to buy, one to send, one for lightning and another for loans. That's a lot of friction. Rhino replaces all of them. Stay to the end and I'll tell you how Rhino can set you up with $21 in free Bitcoin. Rhino is a next generation Bitcoin app built for actually living on Bitcoin, not just holding it. Buy Bitcoin with zero fees. Borrow against your stack without selling it. Move money globally on Bitcoin rails in minutes rather than a wire that takes days. Set a DCA plan so your stack keeps growing, whether the market is up, down or you're out touching grass. touching grass. Bill pay, cards, and Bitcoin retirement accounts are all on the way. And Bill pay, cards and Bitcoin retirement accounts are all on the way. your Bitcoin is never rehypothecated, never lent out, and never touched by anyone other than you. Biometric protection, 24-7 fraud monitoring, and you can move to full self-custody whenever you're ready. It's your Bitcoin always. Download Rhino for iOS or Android at the link below and use code BITCOINNEWS. Deposit your first $100 USD and Rhino drops you $21 in free Bitcoin straight into your wallet. At Rhino, you get the most Bitcoin for your dollar.
Speaker 1Yeah. So I did cut my teeth on dot com. I saw an article with Paul Krugman where he said that the Internet was going to turn out to have less impact than the fax machine. I thought that's so stupid that there's probably a lot of people on Wall Street who are running a lot of money who are that stupid and they're going to figure out the AI is real or the Internet is real. I feel like we're very much in that stage. So what you're talking about breaks down into two questions, right? One of them is, is AI in a bubble? And then the separate question. Well, here, let me just hit that first. So depending on the metric you use, the AI infrastructure is not in a bubble at all. It's come up a lot recently. I think, you know, somebody like Micron, for example. So their stock is up, I don't know, 4F, 6X in the past couple of years. It's up a bunch. However, the reason it's up is because the profits are absolutely outrageous. I think their gross margins is 80 percenters. It's just absurd. The price earnings ratio on Micron, which is, you know, with NVIDIA, it's one of the poster children. The price earnings ratio is like six. It's that's distressed furniture store and, you know, Tulsa. I mean, it's it's nothing. And of course, the reason is that semiconductor industry has had so many boom busts. It's second only to biotech that investors, I think, correctly discount earnings. They say, OK, sure. It's a, you know, for reference, the standard poor in general is something like 20 on PE, right? So the company costs 20 times what its current earnings are. And in the, you know, sort of one of the poster children of AI, it's a third that. So, you know, but that's I don't think it's undervalued necessarily because that is how semiconductors work. You have this giant boom. Everybody has too many chips. And then you get this bust. So that takes us to the two. And in dot com, you had a similar pattern across the board. So you had all this money going into building out fiber. You had vendor financing, which is standard in every single boom like this. The company that I mentioned earlier, I didn't see anything compromising about it. So Harris Corporation, big telecom supplier. We built infrastructure for the Internet back then. And they were, you know, tons of vendor financing. Going on by all the manufacturers, Lucent, Nokia, Ericsson. They were all dumping out billions of billions. A lot of people were raising the alarm. Now, what happened in dot com is that there's a certain built in limit where. I think if you put it simply, fiber optic is just too good. All right. So when you're running copper, you can get 32 conversations, a single line when you're running fiber optic. I don't know. Tens of thousands, millions. And so that meant that once you go through and, you know, dig up America, you only need to do it once. Right. Like because they were putting in tons of fiber. I don't remember the number, but it was like hundreds of fiber and each individual fiber can carry thousands of time. And so the infrastructure boom very, very quickly ended. It overshot current Internet traffic. And then even when YouTube and all these other bandwidth intensive uses came in, it took decades. I think it took like 15 years to start bumping up against constraints with the Internet. The difference in AI, I think, is that anybody who power uses AI, like if you use Grokbot, OK, or if you're actually using AI in your company to start replacing, you know, serious business tasks, not just goofing around, you will very quickly bump up against the token limit and the numbers start going up fast. Right. So most people, when they're using. Using AI, they're going and, you know, doing chat GPT to come up with birthday presents or whatever. You're not going to bump up against the token limit. So you're under the impression that it's free, like the Internet was free. But once you start intensely using it, whoa, there's I mean, like personally, I got to keep myself below five hundred dollars a month. Right. I got to, like, figure out how to token max and optimize, I think that and also the Internet, as amazing as it seemed at the time. If you really take the Internet from sort of call first light nineteen ninety five, which is when Mosaic came out and, you know, AOL opened up or came onto the Internet, which scandalized everybody in the chat room. At any rate, when you take nineteen ninety five as the moment of birth, really like dramatic improvements in the Internet experience, they didn't come along like maybe five, ten years. Right. You had OK, you had social media, you had, you know, mobile because the Jesus phone. Yeah. The iPhone, you had video, obviously. Beyond that, in terms of dramatic capabilities of the Internet, you know, you had, I guess, Internet of Things. You know, you can spot where your stuff is, whatever GPS. Right. I mean, you had pretty incremental improvements, to be honest, for 30 years. I think speaking from somebody who was back in the halcyon days of the 90s when it was going to be the sci fi reality, it's been surprising how little the Internet has developed since then. I think AI is a completely. Different beast like every, you know, currently what they're working on is that you've got the agents who can do tasks. The next round after that is going to be, you know, not monitor my, you know, Twitter feed for news about the Fed. OK, the next is going to be take over my Twitter feed and put out interesting stuff in my voice and get a bunch of followers and then you go back to sleep. All right. Think of the tokens involved in that. Right. And think of the willingness to pay. Pay for that. Right. So if if I had somebody who could reliably do that, I would pay thousands of dollars a month for that kind of service. If you've got a job, right, you could hire them to do your job now, which, you know, they would still have to employ you because there's somebody's got to close the loop. The point is that they're the the functions of AI are moving so much faster than the Internet did. I mean, like we're seeing like a decade in like nine months. And so. So because that all translates into not filling up a thousand fiber optic cables that are already paid for. OK, they're free. Not that. But token maxing where I mean, in three, five years, anybody who's educated, who's curious about the world, who has ambitions in life, they're going to be constantly talking about how do I get this done with less tokens? It's it's going to compete with mortgage payments for like how much money you're spending on this. So I think that, you know, in investing the you always kind of fight the last battle and use the same taxes the last battle. Right. So, you know, the Polish are charging the Wehrmacht with cavalry. I think AI is a completely different beast than the Internet for that reason. It is similarly transformational. I think it's actually more transformational than the than than the Internet was. But I think more important, I think the appetite for intelligence. The appetite to rent intelligence, to have an opportunity Optimus robot, which has the AI brain, it cleans your house and it does brain surgery because why not? What's the willingness to pay for that? My God, how many dollars do you have? So I think that as an investor, I think there's no doubt going to be a dip at some point because there always is. People get excited. At some point, everybody looks around and they're like, are we sure this is going to continue? So yes, I think there's going to be a dip at some point. I am more concerned about the LLMs, the final models. And the reason is because I think any listener will agree. If you have multiple, like if you sometimes use ChatGPT, sometimes Clause, sometimes Grok, and you looked something up three weeks ago, you can't remember which model because they're all the same. They're darn close. There's a couple of things one model does better than the other, but they're also catching up and they're leapfrogging. So I think that unlike. See, what happened in the internet is that the builders did not capture the value. The end user, like Google, the customer-facing companies is who captured all the value because from the customer's perspective, they didn't care about what was in the guts. It's a little bit like we often have this discussion in Bitcoin. People will say, well, normies will never use Bitcoin. You say, well, from their perspective, it's not going to be Bitcoin, right? You're going to have Bitcoin rails and then you're going to have interfaces. So that's what happened with the internet and the value was all captured by the end users. So I think it's a little bit different because the competition is so intense that the models are almost commodity. And I think it's fun to sort of step back for a minute because I think, I mean, I was there, okay? I was writing articles warning about bias in AI back in like 21, 20. And we all were, we were concerned about that. There was also an assumption in the community that whoever cracks AI. Whoever gets the super genius God is going to like own everything. And there's been the assumption going back forever. And in fact, when it arrives, what we discover is that some like seventh grader has the same cognitive power as, you know, billionaires. I mean, I think it's worth stepping back and just considering the fact that that's shocking, okay? Nobody expected that. Nobody in the community, the experts, nobody thought that that's how it was going to pan out. That the supervillain who invents AI is going to have no power. In fact, it's going to be massively leveling. That regular people are now going to have geniuses who they can literally just go online and, you know, you can ask them anything. There's a saying in, I can't remember the term, but when an expert reads a newspaper article about their own subject and the article is completely stupid and they're like, this guy's an idiot, newspapers are worthless. And then they'll read an article about another topic that they don't know about and they'll just kind of believe it.
Speaker 2Yeah, Gilman amnesia.
Speaker 1There you go. Thank you. And, you know, for me, that was kind of the test on AI, right? Which is that I ran through like a bunch of sort of sci-fi, okay? What if we have immortality tomorrow and so people stop dying? Like, you know, what happens to, you know, population, to resources, to productivity, blah, blah, blah. And I was like, okay, let me, let me just run this through. Cause this is the kind of thing that I enjoy playing with. And, you know, I've got my thoughts, but let me see what AI thinks. And I was like, that's, that's better than I could have done. I mean, forget it. And I'm, people see me as an expert on that kind of sci-fi stuff. Okay. And AI is absolutely better than me. It's free. The seventh grader can do it. Okay. So. Well there,
Speaker 2I want to push back on you a little bit because I think you might be contradicting yourself, right? First, you're saying that people are going to take a certain amount of their income and have enough money to pay for their tokens. Right. But then you're also saying it's going to be commoditized and be completely free. So where does it land? Most Bitcoiners start the same way. You buy some Bitcoin on an exchange. Then you get a bit more serious, decide it's time for a hardware wallet and take self custody of your Bitcoin. Eventually you start thinking about running your own node or maybe even getting into home mining. But what if you could build out your entire Bitcoin stack in one place? That's where our sponsor Solo Satoshi comes in. Pick your hardware a la carte or one of their new bundles and get everything you need in one go. They've got hardware wallets, Bitcoin nodes, home servers, solo miners, and even dice and entropy tablets. If you want to go all the way down the rabbit hole and generate your own entropy. Having sold over 50,000 devices, Solo Satoshi carries the brands you know and love, including BidX, Blockstream, Start9, SeedSigner, Foundation, and Kanan. Solo Satoshi is building the one-stop shop for the sovereign Bitcoiner. Whether you're looking for the latest cutting edge home miner, upgrading your self-custody setup, or just putting together a starter bundle for your mom, they've got you covered. Find the link to solosatoshi.com in the description below. Solo Satoshi, your trusted source for Bitcoin hardware.
Speaker 1That's exactly the thing, right? So I think there's nearly unlimited willingness to pay. And then the next question is because of the commoditization, how quickly do the prices compress, right? So that's why I'm. The willingness to pay ultimately accrues to whichever manufacturers have a moat, right? And companies like ASML, the one impressive European company in AI. God, what is it? LAN Research, Micron, NVIDIA. There's a bunch of manufacturers who currently have very strong moats. They have very strong competitive moats. So when you have an enormous willingness to pay, you have a commodity middleman, and then you have genuine moats at the end that are, yes, China can compete, but nonetheless, because of IP rules, even if China has better chips than NVIDIA, they can't necessarily sell them anywhere outside China because NVIDIA will sue them into oblivion. And so you get this kind of interesting profit situation where in internet, the manufacturers didn't capture any of the value. They built the fiber once, it was Christmas for them, and then that was it. And then the middleman captured all the value. In AI, what it seems to me, I mean, this is early, but what it seems to me is that there is far, far higher willingness to pay. Normally, that would cascade trillions upon trillions onto the manufacturer, the middleman. But I think in this case, there is abnormally low lock-in for the middleman. It's not Google. I thought for a minute that agents were going to we're going to do it. And then I literally just the other day, I saw that all of the AIs now allow you to export their entire memory. So every conversation you've ever had with ChatGPT, if you'd like to switch to Cloud AI, you just say, export my memory. So there's like no lock-in. There's nothing to catch all the money. They can do it for now because they're capacity constrained, because there are some, you know, there are still some real differences. Like Grok sucks at writing. I love Grok. I love Elon. Don't ask Grok to compose, you know, an essay for your kids to get into Harvard. They'll get into community college. There are some differences between the AIs, but I've been struck that they're small and they appear to be compressing.
Speaker 2So I guess what you're saying is this time is different.
Speaker 1I think this time is fundamentally different. But having said, you're almost guaranteed that it's going to overshoot. Okay. That point is not different. The question is, you know, if SMH, which is the semiconductor that kind of gathers together all the producers, if SMH is, I don't know, something like 500 today. Okay. If it's 500 today and it's eventually going to overshoot and drop by half, then okay, is it at a peak and it's going to go to 250 or is it going to go to 1500 and drop to 750, in which case you should still buy it? Because that's the thing, right? Is that in the dot-com era, call it 1995 through 2000, everybody from day one said it was a bubble. So I first bought into Yahoo was the first internet stock that I bought and that was in 96. And everybody told me I was an idiot, right? Everybody was like, this thing's a bubble. Any Bitcoiner knows. Okay. When was Bitcoin first called a bubble? Right. So if you would listen to them back in, I don't know, three weeks after Genesis block. Okay. So this is always the trick. So, you know, is, is AI ultimately going to be a bubble on both the producers and the LLMs? Guarantee it. I will, I will bet my eye tooth. However, where are we in that bubble?
Speaker 2That's the trillion dollar question, right?
Speaker 1Always. Yep. If you can figure that out, you will, you will be very rich.
Speaker 2Well, then I want to touch on the fact that this thing is so amazing. We're in the middle of this technological revolution, but the people seem to hate it. Uh, the, uh, the election was a clear sign of that. Everybody voted against these data centers on, I think, 46 or 48 resolutions. They all voted against people would rather live close to a nuclear power plant than a data center. Uh, so what is going on here? Why do people hate it so much? And will the Democrats slow it down?
Speaker 1Yeah. The number you cited is that's the money number here because that is completely a rash to fear a data center more than a nuclear power plant. Now there's safe nukes, blah, blah, blah. But you know, sort of understanding what most people are thinking when they're afraid of nuclear power plants. And to say that data centers are worse, that's irrational at that point. And so what's driving that? And, you know, people have been anxious about AI since the beginning. For whatever reason, the AI community has absolutely loved the doomsday scenarios. I think they have like Manhattan Project envy. And, you know, they talk about wiping out all the jobs. And this has been going on for a long time. Like the first study saying that AI was going to wipe out the jobs, those are like 15 years old. Yeah, this did not start last Tuesday. The entire industry, and, you know, maybe they did it to raise money to, I don't know why, but it's very rare. Like biotech does not have conferences where they're like, you know what, though, we could just kill everybody too. It's weird. Like it's almost like suicidal, I don't know, apathy towards their own existence. But anyway, AI has done that for a while. So people have opposed it. Because they buy it, right? The experts say that all our jobs are going to be wiped out. We'll buy gum. That's probably going to happen. So that, you know, predated the sort of latest talking points. Like the strongest one at the moment appears to be electricity bills. That's what really gets people excited. And I mean, fair enough. We live in a democracy. You know, you have to take voters as they are, not as you wish they were. I mean, if you look at the actual data, the impact is very, very small. It was actually a study that found that AI, having AI data centers lowers electric bills. It's counterintuitive, but consider that a lot of electric bills is fixed generating costs. And so if you amortize that over a larger base, you know, so if you limit this up to Rust Belt Town, where half the, you know, power plants are shut down because everybody left, well, then having an AI data center in there actually amortizes the cost of running the plant and actually lowers your bills. So that's the data. But of course, nobody buys that. This is egghead stuff. And, you know, the Democrats have been very, I mean, it was kind of a freebie. Like, you know, many polls had found that people in general are terrified of AI data centers and AI data center. It looks like an Amazon fulfillment complex, but it doesn't even have the traffic. You don't even have the trucks coming in. I mean, it's nothing like I lived in North Virginia. These data centers are everywhere. You don't even notice them. Sometimes I paint them pretty. They are literally nothing.
Speaker 2They do hum though. They make noise.
Speaker 1They hum if you're standing right next to them.
Speaker 2If your house is like next to it, like you're going to be, you're going to hear it. That is, that's. The one objection. I think that the electricity you're right. Complete with that. And the water usage, right? Like the water usage is a joke.
Speaker 1There was Jason, who's not very popular in our community, but you know, he's part of the all in podcast. Right. And you know, he was pointing out that like golf courses use like 28 times more and like, they're not curing cancer. They're not, you know, like AI already won. I think it was the chemistry Nobel last year. Cause it figured out a way to fold proteins was really important for drug development. Okay. Golf course, isn't doing it. And yeah, Jason had a tweet. He was like, golf courses. And yeah, I mean, I'm, you know, I'm kind of with him. Alfalfa production in California uses someone got like a hundred times. It's stupid. And I mean, obviously they only do that because it's subsidized because government gives away for free. We should just buy our almonds from Mexico or whatever at any rate. Yeah. Most of them are BS, but the point is that they're landing with voters. Right. And so then the question is, is this actually going to impact? I mean, what I've been pushing for over a year, is that like the data centers need to pay the damn people, right? Like if you want to build a data center in some town and the locals are like, no, you know, it's going to, it's going to do the thing with the contrails. Okay. Cut a fricking check, dude. Like come in and say, listen, you guys are never going to have to pay property taxes again. And, and we're going to double the school, you know, budget, whatever. I mean, just like y'all are dealing with trillions of dollars, man. Cut a fricking check for like $8 million and call it a day. Again, just, I don't know. I don't know why the industry has not done that more.
Speaker 2They're just autistic and don't know how to deal with people. That's what it really is.
Speaker 1This is tech people. Yeah. I mean, this is partly why like, um, you know, they need to hire people. The Silicon Valley got smarter about it after the Microsoft trial and they started hiring lobbyists in Washington. Unfortunately, that was probably, you know, taking a bite from the poison apple, uh, corrupted them. So that that's a whole different discussion. But at any rate, um, if they had, you know, started paying people early, so that communities would see it as like a jackpot, which I mean, it should be, there's, there's so much profit involved that these guys could slice off a chunk for the locals. Um, so, you know, those will still occur. Uh, states are banning them, not all the states, as long as you have some states, then, you know, maybe the industry gets its act together, get some sort of association, essentially a cartel that can actually, um, bribe locals to put data centers in. So there are some strategies that they can do. There are some states that are just going to be kind of, you know, they're going to be, you know, they're going to be, you know, kind of knee jerk, um, hostile to regulation. Um, but broadly speaking, it is going to be another roadblock in the way to building out data centers. You know, the industry can partly, um, they can compensate, you know, they can focus resources on the jurisdictions are still open. Um, but to the extent they can't do that, then they have to put resources into trying to squeeze more out of less, which are resources that you are not putting into maximizing your output or your, you know, um, performance. Uh, so that then goes back to the China issue, whether China's, you know, still going to be whatever it is, six months, two years behind, uh, if these activists manage to slow down data centers.
Speaker 2And another part of the election that just happened, it's sort of the rise of democratic socialism and you as a libertarian Austrian economist, uh, I'm sure you don't like this new trend happening here in America. What do you take of it?
Speaker 1Where do you think it's going? Yeah. I mean, it's understandable. Um, people are pissed off about the unipartisan. Um, people are pissed off about and they're pissed off on the right and the left. And this is actually kind of a trend that we've seen for a while. So like in the 2008 crisis, everybody in America was pissed off at the bankers and one of them in jail at a minimum and nobody, nobody did time. I think there was one kid at like, I don't know, Bear Stearns or something who did like, I don't know, a year or two, nobody, nobody got punished for that. And in the wake of that, you had the tea party and then you had occupy wall street and they were the same fricking movement. And they should have, they should have hooked up and had a baby. But of course, at that point, the media came out with, uh, you know, we found the racist in the tea party. Now these guys actually hate you and they murder you. So this is kind of the play, you know, you, you, you've got the 1% setting the other 99% against each other. Uh, but of course occupy wall street was, was correct. Right. We, we live in a financial oligarchy and, you know, we get strip mined. Um, you know, capital strip mine. Uh, and so, you know, in a sense, uh, I do sympathize with the people who are voting DSA. Uh, and, you know, I think there's a kind of ongoing battle from the right populists and the DSA to, um, sort of capture that energy. Uh, I think we're living in a fascinating moment right now where the mainstream is collapsing across the board is collapsing in popularity. It's not collapsing in power because it's not collapsing in popularity. It's not collapsing in donors, but it is collapsing in popularity. And you're seeing that across the West. At least you're kind of seeing it worldwide. You're seeing it in Japan. You're seeing a lot of places where the mainstream simply has no credibility that I put down to social media, uh, you know, in a word Gutenberg 2.0, right? So the Gutenberg revolution was that he made printing cheap. It was actually Chinese tech, ironically, but anyway, he imported Chinese tech and he made it very cheap to print, which made it very cheap to organize the peasants the peasants will always outnumber the insiders. And the insiders had been in control previously because the peasants couldn't organize. Uh, you couldn't like hold a rally next week at the Abbey's market and like hire a bunch of monks to hand letter your, your, uh, uh, you know, pamphlets. Uh, you could do that with, with Gutenberg press. That's part of the reason why government censored presses early on. You had to be licensed. Uh, that's why governments created post office is not junk mail. Surprisingly, it is censorship. Every state started it because you had private sector. Okay. Like, like you would take your mail to the market and then you had regular runs going to the next market. You didn't need a government post office. The reason for it is that they banned private carriers so that the government censor could see when you're organizing peasant revolt. Social media is the 2.0 version of that, right? It's the same thing. So, you know, whether it's the DSA or whether it's the populists, you are seeing these explosive movements because of social media, which is of course, like in a sense, respect to the censors during COVID. Yes, they know exactly where the pressure point is. Okay. Social media is how you organize revolutions, Iran, China, all of them know exactly what the threat is. Uh, so, you know, I, I think it's going to be fascinating, uh, what's going to happen next in the U S because both wings, especially for the younger generation, if you're Gen Z and you know, you're going to be pissed off at the world, there's debates, whether you have a right to be, you know, things are much better than they were in 1978, whatever. Setting that aside, the point is that a lot of people are angry. They had expectations are not being met, right? Um, they can't get married. They can't buy a house. You can't find a job that can support a mortgage or that can support a family. They're pissed off and they are not right. They're not like, they don't have posters of like Mitch McConnell on their wall. or Hakeem Jeffries, okay? Those people are up for grabs. And I think we're at a fascinating societal transformation right now where like kind of all the power is up for grabs. It'll be delayed because, you know, the donors are going to fund, you know, whatever. There will be rear guard reactionaries. But I think we're in an absolutely fascinating moment. That's partly what gets me up in the morning. I know that we're in a battle of hearts and minds and the Uniparty doesn't have the goods anymore. There's, you know, people have been voting that way, like voting Tory or whatever. There are people who sort of vote that way out of habit, but they're being replaced by a new generation. And that generation is they're much more creative in both a good way and a bad way than the people who are leaving.
Speaker 2And so what does motivate you in terms of your message? What do you want these Gen Z's to hear before they get sucked into the Hassan Pikers and Nick Fuentes?
Speaker 1Yeah, I think, you know, number one is so I'm an anarcho-capitalist. I'm a Rothbardian. If you're familiar with Rothbard, I'm in his Pat Buchanan stage, which explains my relative support for Trump. I'll leave that to listeners and AI can help if you're not familiar with Rothbard. So anyway, I view the government as a gang of criminals. Every pretty much every single thing the government does is, you know, either it destroys lives, it makes us poor, it makes us into slaves. And so, you know, I think that message is very important. I think Bitcoiners are kind of have been ever since Genesis block. They're kind of the target market. These are people who are very creative. They're very interested in like alt histories there. They have an incredible imagination and they have the sort of free thought and the what would you put it? Grid. OK, to persevere and to build things. I think that kind of people are those are our soldiers. I think those are the people who are really a question. You know, if you look at revolutions through history, it's it's one percent, five percent who actually launched the revolution. Right. Everybody else just kind of they just kind of stand back. You know, like we go to a disco and there's people with their their hands in their pockets. OK, most people are not on the dance floor. Right. They're standing there and they're looking for guidance. And they're looking somebody to follow. So I think it is exponentially important right now that we emphasize the value of liberty. The fact that all of these things that are pissing off Gen Z, whether it's, you know, houses being expensive, jobs not paying, whatever, that all of I mean, that's that's kind of an economics education question. But, you know, so underlining that all of those things are downstream of liberty. If you have liberty, those things all fix themselves. If you do not have liberty, then you are on the road to serve them. I think that's that's the critical point. That's partly also why I do what I do. So I'm a former professor. I used to write papers, academic papers, and I concluded that, you know, economics fixed the R&D part a long time ago. You could argue to go 200 years ago. We we we understand in economics what you do. OK, like how do you make a country rich? You need property rights. You need people to have the freedom to innovate and build and also the incentives that they can keep what they earned, et cetera. All these things are solved. We know all of the big questions. In economics, what economics is singularly fits pathetically failed out is marketing. Right, so, you know, the vast majority of voters, even in educated countries like the US, we like to shit on ourselves, but it is a very educated country, even, you know, people with, like, PhDs in sociology. They have no idea what, you know, rent control does to housing availability. OK, they have zero idea of this. So I think economics is a field has failed spectacularly. OK. That's partly why I do, you know, I make three and a half videos, minute videos, trying to explain stuff for normal people in normal language, no jargon, because I feel like that is where economics has failed humanity.
Speaker 2And let's talk about that 25 year old who's starting off looking for a job, maybe has a job, maybe doesn't. It's pretty hard for kids to get on the ladder right now. In terms of economic advice, how should a kid start a young adult start off their their life? And gain that economic freedom? Yeah, so you've
Speaker 1got a couple of different horizons. I think it's good to start with. Where's the end point, right? So, like, during the course of his life, what is the labor market going to eventually look like? It's not going to look like that tomorrow, but what's the trend? Where are we headed asymptotically? And in there, I think that the best pattern is the Industrial Revolution. So the Industrial Revolution commoditized labor to a large extent. And what the. What the AI revolution is going to do is commoditize experience or experience as an expertise, not as in doing fun things. So it's going to commoditize expertise, knowledge, education, all of those things are about to get dramatically cheaper. And so how do you position your career? So I think when you take the let's first talk about the period before the robots, because that's very important. And there we're probably looking at. Wild asked gas 10 years that the market, you know, before robots start to make a huge impact on the market during that period, I think that the kinds of jobs you can be interested in, again, looking back to the Industrial Revolution, trying to trace out what's similar, what's different. So you want to identify the things that remain if intelligence and expertise become commoditized. So the starting point is human to human contact services, broadly speaking. So, you know, vending machines have existed for a long time. Shockingly, there are still bartenders, even in the era before age checked. Nobody viewed a bank of vending machines as equivalent to a bartender you've had back in nineteen hundred. There was something called an automatic where you would walk in and you had food behind little doors and you would put a dime in the door and take out your pumpkin pie. Shockingly, those did not replace restaurants. People people don't like to put the number in the tape and the you know, and then. And then the food comes on a conveyor belt. Those exist, but they're fringe. Generally speaking, people like to interact with other humans in the services. You also have before the robots, you have maintenance, you know. So if you think about the people that you hire in your personal life, you got an electrician, you got people who deliver your groceries. Right. So, you know, again, services, but relatively skilled services, you also have makers, people who create things. That could be YouTubers, the much maligned YouTubers, of which I'm apparently one. My kids are embarrassed about that. It could be, you know, hand making stuff. Remember years ago there was some kid, he made a steampunk keyboard and he showed it online. Everybody's like, wow, that's so cool. Can you make one for me? It was like, sure. And people wanted so many that he was charging fifteen hundred dollars. This is like in the year 2000. So it's like three grand today. He was charging fifteen hundred make a keyboard. So he quit his day job. OK, what happens with A.I.? So the the the common. Threat in all of those is that the A.I. Sorry, the human adds value. So today you can hire somebody, you can commission somebody to make you a chair. You can also buy one at the factory. The one at the factory has fewer flaws because somebody with a big brain designed the stamps and any flaw they saw on the stamp, they took extra care to take that flaw out. So you literally have you probably have, I don't know, five, 20 times more flaws in the handmade one. And guess what? People will pay five, 20 times more for it because it was made by a human. Right. So this is what people forget is that huge chunks of our existing economy, like more than half of jobs, nurses, for example, you know, dog walkers, coaches, you know, like personal assistants at the gym. OK, there's a ton of jobs where the value is that it is a human. You can already replace a lot of those in theory. You know, you could I guess you could hook up your dog to like a robot, you know, one of these little delivery carts they have or so you could do it in theory. Nobody wants to do it. Right. And so those jobs are where the growth is. OK, so if you're young, that means get into HVAC, anything to do with home repair, anything to do with the skilled trades, electrician, the data centers themselves. So the amount of investment going into data centers currently higher than it is for offices. There's poetic. It's higher than it is for for residential construction. Right. It's absolute mass of the mouth. So get into the trades. The trades are going to explode. The only so I could, you know, turn out to be Saab and it did absolutely nothing. And it was Paul Krugman's fax machine possible, in which case false alarm continue doing what you do. But if I does what it's supposed to do, if it replaces the jobs because it's super brilliant, then that also. It also implies you have a dramatic increase in productivity. The Industrial Revolution, which didn't do anything for intelligence, it only did manual labor. And if you look at societies before and after the Industrial Revolution, depending how you measure, it's like ten to one hundred times richer. Intelligence is more important than manual labor. If your GROK bot or your, you know, Tesla robot is doing brain surgery, OK, you sum up the improvements, the quality of life of all of that, it is far bigger than the Industrial Revolution. So if you. have a society that is conservatively 10 times richer you got stuff all over the place everybody's got a second home nobody who lives in florida is lacking a swimming pool swimming pool in every pop right you have massive amounts of stuff that needs taken care of so so that's the pre-robot okay pre-robot you want to focus on services um you know uh being a barista which now pays double because the ai made you know everything more productive all right so uh services making stuff entrepreneurship skilled trades now once robots come in the skilled trades the bottom falls out right yes there's still some people who don't trust the robots and so no no no i want a guy coming in you know i want to be able to look him in the eye and make sure he's giving me a fair price okay this will always exist but it'll be increasingly fringe uh as those kinds of people age out of the market um but even there right okay so the skilled trades dropped off however at that point you get another 10x bump in productivity which is the productivity has nowhere to go it's going to go to people like stuff doesn't actually take money and like hide it and go you know ai like you know sips on juice boxes when nobody's looking like the money ultimately accrues to human beings and then those human beings go out and buy services uh so the the skilled trades and then i can disappear overnight right because the thing is remember if you're trying to replace all the accountants with an ai you need one ai to do it really really well and then all the accountants are toast it'll take time but anyway on the other hand if you're trying to get a robot to replace mcdonald's workers we need like six robots per mcdonald's so that's going to take a much much longer time uh for perspective the first electrified factory in the u.s okay which was 1870 or something somebody's gonna do the ai thing and i'm wrong anyway something like that for illustrative purposes so if you count from the first electrified factory which is obviously better than shoveling coal in there to half of american factories being electrified something like 50 years so the technology was better and then it took 50 years why because you run out old infrastructure you know you already got the mcdonald's if you do the robot thing you gotta gut it you gotta completely rework uh so the robot's gonna take a while so you know when i say 10 years until the robots are here okay but the robots replacing your hvac job you know skill jobs in general i think you're probably looking at closer to 20 30 years so for a current gen z i think that's the recipe uh services makers entrepreneurs and skill trades now and then the next question is what comes after the robots then you get another bump in it the skill trades thing falls out but that's not a problem for gen z that's a problem for gen beta which we're gonna have to rename because it's insulting
Speaker 2yes i mean everybody's gonna have to become an artisan in some kind of way which
Speaker 1is beautiful like if you've seen the movie office space so the whole premise is that he's got this shit cubicle job it's golden handcuffs he hates the job and then he's like you know what eff it i'm out and he goes out and the end of the movie he's doing construction he takes his helmet off blue collar the kind of job that babes look down on he's doing the blue collar job life is good this this is the ai economy people will look back when it's done and they'll say wait a minute you guys you all gen zers you all you wanted the cubicle jobs like half of people sat in cubicles all day and they tried to like buy stuff on temu so they could brighten it up so it wasn't quite so soul crushing that's that's the job so which is the equivalent of today when you hear about people you know smashing tractors and you're like you wanted to to pitch hey that's the job you want and thank god you failed because now you sit in a cushy air-conditioned uh cubicle watching youtube videos and i i know this because i can see when you watch the videos america instead of working okay so that's that's what we're talking here right you go from pitching hay to a cushy air-conditioned job sharing memes at the office now you go now take that exact measure that and now do the whole thing so yes jobs in the future are going to be incredible uh whatever you like doing is what you're going to be doing okay i'm a tour guide in the amazon just whatever whatever your thing is and it's going to make poultry poultry poultry wages in today's terms like three bucks an hour that's why you don't do it now okay however magnified through the ai and the robots that's gonna be a buck 50 an hour so when the smoke settles you will do the exact job that is your calling and they don't pay 100 grand could you make 400 grand working in a cubicle yeah but who wants to do that and we know that that's exactly what the math will look like because that's how people believe when they have fu money you can they no longer do the thing they hate now they do something they love if it makes money grand if it doesn't who cares i like raising sheep and then how
Speaker 2does bitcoin fit into all this in an ai future where is bitcoin
Speaker 1so i think the trick with bitcoin is that through all of this glorious productive economy you've got overlaid on top of it this parasitic monetary economy where you know if you zoom out and ask what is money fundamentally money a dollar bill is a bidding ticket on real resources right this is why printing more money does not make you rich javier millan argentina's got a great line right he says uh what is it um printing degrees that doesn't abolish stupidity okay and printing money doesn't abolish poverty they are merely bidding tickets right so if you print some extra bidding tickets all you're doing is competing with the existing bidding tickets you're effectively stealing whoever you give the counterfeits to uh or the injected money if you're a central bank uh that has been stolen from everybody else right and so it is you know partly it's dangerous because it it sort of gives the patient morphine uh while they you know continue i don't know sawing off their arm uh so they're not really aware they're doing it it's like a slow motion saw uh so it's very dangerous for that and it's it's sort of anesthetize us now into for example 40 trillion dollar debt in the u.s uh you know i don't know i don't know i don't know many many trillions in other countries i think the total government debt at this point is greater than all the money in the world uh which again is poetic uh so it's very dangerous for that and then the other danger of it is that it it um you know like at some point that's not sustainable and you could get some sort of collapse and then the question is what's the collateral damage in that collapse if you look at every crisis every economic crisis in american history i think one of them turned out well which was getting rid of the fed the second bank of the united states that's about it okay every other crisis was bad all right whatever caused the crisis did not get fixed in fact um like in the case of 2008 crisis actually got institutionalized we saw similar bailouts in 2023 and like nothing happened the reason like nothing happened is because they're automatic now we should make you sick um it should get you back to occupy wall street um but anyway so i think that's the concern is that the the system we're building is it's sort of building up to this ragnarok where you know you have this massive explosion and the question is what gets consumed in that supernova if you go back through history it is very very likely that in that supernova we come out the other end with a european style economy which is to say strangled and sort of dribbling off new living the only reason why europe is richer than it was in 1980 is because they copy american tech our internet our iphone our everything our you know science diseases cured just across the board they're parasitic just like the soviet union was parasitic that's that's my concern is those two things
Speaker 2and so bitcoin so should people own bitcoin for this future is that the
Speaker 1yeah so bitcoin or gold of course are both life rafts right so if you are in a system that is uh inflating and is starting to flirt with hyperinflation then you know the obvious way is you buy something uh hard uh gold historically served that role but of course the problem with gold the sort of irony of gold is that the reason why gold bugs love it so much is because it's real you can hold it it's made of atoms you can drop it on your foot and this is of course also the fundamental flaw of gold is that you can't tell people like i have gold but it's secret i can't show you where it is right nobody's going to trust your currency based on hidden gold you have to actually show where it is you have to set up a tv camera you have to put walls you have to put the crocodiles to dissuade thieves uh at which point the government's going to come visit you and you know they're going to check to make sure you know whatever that your w2s have been calculated properly they're going to have opinions about your gold um so you know near term in other words if if in that ragnarok right if you think that it's going to be limited to you know just massive bailouts and then regulations to make sure it never happens again which of course make it much much bigger uh the next time so if you think that that collateral is going to be limited uh then you're going to have to make sure that it's going to be limited to you know sure park it in gold um you've got a separate question of gold of course which is that if you're not paying storage costs your gold is not there it is re-hypothecated it's re-lent out uh so that's a you know completely separate that's the paper gold debate but that that's absolutely true um the only way that you can be sure your gold is actually where it claims to be is if you're paying them to store it because they're not doing it they're not charities okay um so that's a separate question to gold but i think gold is kind of a bet that when the explosion comes it's not going to be that big if on the other hand you think that it's going to be you know catastrophic that government's going to reach for every last thing they can they're going to seize your your stocks they're going to do you know europe now has this uh i guess they want you to list everything you own like all your jewelry, which is a little bit ominous. I mean, that was done in the 30s. So if you think that it's going to be a massive explosion, then you want to hold something that governments absolutely cannot control of, which is the premise on Bitcoin. I mean, that was from day one, that was the premise.
Speaker 2And as you said, born out of Occupy Wall Street, the Tea Party movement, and now today it's all coming full circle. Well, Peter, professor, thank you so much. That was such an awesome conversation.
Speaker 1Yeah, I enjoyed it.
Speaker 2Great questions. If people want to find out more, where should they go?
Speaker 1I am on X, the artist formerly known as Twitter, at Prof St. Ange. And I also do a, let's see, I do weekly deep dive articles on what's going on in the economy. I do a monthly investment letter.
Speaker 2There you go. Watch Peter's videos. They're awesome. They give you everything you need to know. Oh, that's right.
Speaker 1I do daily videos talking about the economy. I've recently gone to weekly, unfortunately. I got to spend more time with my kids because they're almost 18. However, you can check those out. There's, there's archives. The archives are over at YouTube. Same username, Prof St. Ange. All right.
Speaker 2We've got to have you back on. That was a real pleasure.
Speaker 1Absolutely. Enjoyed it.
Speaker 2Cheers. All right. Thank you so much for making it to the end of this interview. If you enjoyed it, hit that like button, subscribe to the channel, turn on notifications. So, you know, every single time we drop a new one, if you want to dive deeper and learn how the blue tsunami affects the AI trade, be sure to check out the latest interview we just did with Matt Dines and Marty Ben. Two geopolitical monetary minds who go, deep, deep, deep into all the things happening behind the scenes, putting together the headlines and making sense of this rapidly changing world. I think you'll enjoy it. I'll see you over there.

Podcast Summary

Key Points:

  1. Jensen Huang's resignation as NVIDIA CEO raises concerns about the future of open-source AI and American dominance in the technology.
  2. The debate between open-source and closed-source AI centers on whether AI should be free from government and corporate control or locked into an oligarchy of major labs.
  3. The controversy involves alleged chip smuggling to China through a Singaporean company and questions about what NVIDIA's leadership knew.
  4. Restricting chip sales to China presents a tradeoff between slowing China's AI development and incentivizing their domestic industry, with China currently estimated to be six months to two years behind.
  5. The guest views government as a gang of criminals and argues that liberty is essential for solving economic problems facing younger generations.
  6. AI's economic impact may differ from the dot-com bubble because willingness to pay for intelligence is potentially unlimited, though commoditization of AI models limits middleman profits.
  7. Local opposition to data centers is driven by misinformation about electricity and water usage, and the industry should compensate communities directly.
  8. Bitcoin serves as a hedge against inflation and government control in an uncertain economic future marked by rising debt and potential crisis.

Summary:

Peter St. Onge joins the podcast to discuss the resignation of NVIDIA CEO Jensen Huang and its implications for the AI industry. St. Onge expresses concern that Huang's departure threatens the push for open-source AI, which he views as essential for preserving liberty against both government censorship and corporate oligarchy. He warns that without competitive, government-independent AI, voters could be brainwashed and freedom could be lost.

The conversation explores two central debates: open-source versus closed-source AI, and whether to sell chips to China. St. Onge notes that Huang advocated for open models as an alternative to the dominance of Anthropic, ChatGPT, and Google. On China, he acknowledges the tradeoff between denying chips to slow Chinese development versus incentivizing their domestic industry, comparing it to the Cisco-Huawei dynamic.

St. Onge, a libertarian economist, argues that government is a gang of criminals that destroys lives and wealth. He advises young people to pursue skilled trades, human-to-human services, and entrepreneurship in the near term, as AI commoditizes expertise but not human connection.

Regarding AI's economic impact, he believes this time differs from the dot-com bubble because the appetite for intelligence is nearly unlimited, though model commoditization limits lock-in. He recommends Bitcoin as a hedge against inflation and government overreach in an era of massive debt and potential monetary crisis.

FAQs

He argues open-weight models let companies run AI in-house and set their own parameters, preventing control by a small oligarchy like Anthropic, ChatGPT, or Google. He sees this as essential for competition and liberty.

Huang resigned from NVIDIA, the company he founded, amid reports involving a Singapore-based company allegedly reselling billions in chips to China. The guest considers this significant because Huang was the biggest advocate for open-source AI.

The guest sees a tradeoff: denying chips may slow China but also pushes them to build domestic industry, as happened with Huawei versus Cisco. He is open to restricting sales but notes illicit channels cannot be fully stopped.

The guest says AI infrastructure is not obviously in a bubble by metrics like price-to-earnings, though a dip is likely eventually. He thinks AI differs from the internet because demand for intelligence and willingness to pay are far higher.

Concerns include electricity bills, water usage, and noise, though the guest argues these fears are largely exaggerated. He suggests data centers should pay local communities to win support.

He recommends human-to-human services, skilled trades like HVAC and electrical work, making things, and entrepreneurship. These are harder to replace with AI than knowledge-based jobs.

Chat with AI

Loading...

Pro features

Go deeper with this episode

Unlock creator-grade tools that turn any transcript into show notes and subtitle files.