Lemonade Stand - Day 228 | Ep. 084 Lemonade Stand 🍋

Episode Date: October 14, 2026

On this week's show... Aiden follows the oil, Atrioc looks inside a data center, and DougDoug messages the group chat. We launched a Patreon! - https://www.patreon.com/lemonadestand for bonus episodes, discord access, a book club, and many more ways to interact with the show! Episode: 084 Recorded on: October 13th, 2026 Clips Channel: https://www.youtube.com/channel/UCurXaZAZPKtl8EgH1ymuZgg Follow us TikTok - https://www.tiktok.com/@thelemonadecast Instagram - https://www.instagram.com/thelemonadecast/ Twitter - https://x.com/LemonadeCast The C-suite Aiden - https://x.com/aidencalvin Atrioc - https://x.com/Atrioc DougDoug - https://x.com/DougDougFood Edited by Aedish - https://x.com/aedishedits Thumbnail by Cheyenne DeWolf - https://x.com/cheyedewolf Produced by Perry - https://x.com/perry_jh New takes on Business, Tech, and Politics. Squeezed fresh every Wednesday. #lemonadestand #dougdoug #atrioc #aiden Learn more about your ad choices. Visit podcastchoices.com/adchoices

What if Bruce Springsteen had never become the superstar he is today? I'm Will Anderson, and this week on my music history podcast, The Monday Music Club, we're talking about the making of his breakthrough album Born to Run with the guy who wrote the literal book on it, Peter Ames Carlin. He had put so much of himself into his music, like all of himself. I mean, it was hard for him to have romantic relationships that could last. You know, he didn't really do anything, but put all the energy into his guitar. It's a fun ride, so listen to the episode right now.

by searching for Monday Music Club anywhere you get your podcast. What if Bruce Springsteen had never become the superstar he is today? I'm Will Anderson, and this week on my music history podcast, The Monday Music Club, we're talking about the making of his breakthrough album Born to Run with the guy who wrote the literal book on it, Peter Ames Carlin. He had put so much of himself into his music, like all of himself. I mean, it was hard for him to have romantic relationships that could last. You know, he didn't really do anything, but put all the energy into his music.

It's a fun ride. So listen to the episode right now by searching for Monday Music Club anywhere you get your podcast. Three, two, one. Ladies and gentlemen, we are. Great show. Why are you laughing? We've a show this week. I've never seen you do that for the control ever. What? You went like this head down with three to one of the last place. Dude, I don't know if I tell you guys. I meditate for five and a half hours for every episode to get into the mood. Yeah. That's why I show up. And this week you didn't.

into the mood. I show up and I maintain an erection for the entirety of this recording. No, we know. That's why you wanted to get started. You're worried about running out of that time. No, I'm not. It goes four plus hours. Although it was starting to get a little flaccid when I saw your face. What happened to your beard? This is everybody's talking about it. You know what sucks is I was playing basketball with Aiden and I got elbow in the face and I got hit right here. Oh, wow. And then I shaved it the same time. So it just stands. I couldn't shave that part. So I have like a false goatee going on. Yeah.

So, thanks for you. We got a great show for you today. We're going to talk about the hilarious Russia deal. We're at Diesel and a Saudi airport attack and bond yields and some stuff about neuralink and mathematics in the UK and brands and creators. It's a wild and wacky week. I think I speak for everyone when I say, what the heck? Question mark or however we chose to frame this particular episode title. This week was crazy.

This week was crazy. This week was. Every week we sit here for 10 minutes generating new synonyms for the word crazy to describe this week. But let's start off. Let's move over to Trump and Russia because my gas is getting pretty expensive, Aden. And it's mostly because of this one little thing that happened recently. One tiny little. It's one little recent thing.

There was big news this week because on Friday last week, Trump announced that he had made a new deal with Russia to import a bunch of diesel fuel into the United States. And this is breaking precedent because we've had sanctions on Russia since the beginning of the Ukraine war. And we have not been accepting any of their diesel fuel in that time. And it comes at a time when diesel prices are at an all-time high, something that is boating very poorly for the Republican Party's chances of winning in the midterms. It's like one of the number one issues affecting people, you know, whether or not they need to fill their own car. with diesel. It affects the prices of everything that is transported or brought around with diesel. So this is made in an effort to have access to more diesel to relieve this price, right? There's a few big reactions to this, obviously. As you can imagine, Ukraine, not super happy with this,

because what does it mean? It means an upwards of $4 billion going to their war adversary right now. also the rest of Europe this is completely out of step with the efforts they're making on their end they want to maintain the sanctions they're holding or the support for Ukraine that they have in the war in order to end it sooner

and this is like a big signal from the United States that we're not as in support of Ukraine or we're allowing Russia to cross certain lines in this in this conflict yeah like we have a tier list of our values and above war

on other nations is oil prices. Yeah, right. Large quantities of diesel. We're willing to make exceptions. Yeah, right. It's sort of a porous moral framework. A porous moral framework is a great way to put it. But Trump's response to this is like,

oh, the fake news never pays attention to the amount of diesel and fuel that Europe is consuming from Russia over the same time period. And I was like, you know what, that's maybe a good point. Like, let me look into that. and because there is still a massive amount of oil and natural gas that is imported from Russia into the European Union in order to handle their own energy, right? What I did not realize is the amount that the EU has reduced their dependence on Russian energy by in this time period.

So at their peak in 2022, the highest per day that they were importing at the time was 800 million euros worth per day that they're importing. that they're importing from Russia. From Russia at that time. That's gone down to 47 million euros now per day. From what number? From 800 million euros worth at its peak in 2020. Which is an incredible falloff. The additional argument to this is you haven't seen the same decline in natural gas imports over that same period of time.

But they're still falling through with plans to phase out Russian natural gas by like 2030. So the EU is putting like a goodwill effort into like down ticking their reliance on Russia, Russian energy over this time period and is not really like breaking from that at all. And the thing is when you contextualize this with the grand demand for diesel fuel, this is actually not that much that the U.S. is important. This deal is actually like pretty small. It's like half the amount they're using a day, right? Yeah, it's not even, it's not even close.

So the important thing is this is more of a signal to these parties in Europe at a bad time to be like, fuck you guys, we're making deals with Russia. But also it's not an amount that it will have a material impact on the price. Like this is very unlikely to work out in Trump's favor in his goal is to bring the price of diesel down by increasing access to supply. But it's not enough on the day. relative to the demand to heal, to make this meaningful difference. So it's just, all it does is piss off Ukraine, piss off the European Union, doesn't relieve prices. And so it's really failing on all fronts, or rather it's expected to fail on all fronts. And even if you think the EU is being hypocritical,

the like practical function of this decision is just not very good. Yeah, yeah, the UK's vote out against it. And what's interesting, this is unconfirmed, but Russia, before this had banned diesel exports in general because they have a diesel problem in their own country. And a lot of the refineries have been getting blown up by Ukraine. And so for them to get even 300,000 tons out is something.

And there's unconfirmed reports that some of it is imported from Iran. So they're in, Iranian takers were spotted dropping off diesel in Russia. And so the idea is that they're just buying Iran. Iranian diesel and then reselling it to America. So in which case, we are just funding Orion, you know, it's like, and Russia gets a skim off the top. So we're just funding Iran and Russia to get diesel for a, you know, it doesn't make a lot of sense.

So in the buildup to this decision, the U.S. and Trump had specifically requested that, you know, Ukraine, please stop hitting refineries in Russia. We really don't want you to do that right now because of the strain that is adding to the global oil market, the global diesel market, right? And they don't have a lot of places in Russia right now to refine the diesel in order to supply this amount. So there's this suspicion of, oh, could this be like turning around Iranian product? And that's what's getting back to the US. But even if it's not that, if you look about a couple months ago, because refinery capacity in Russia is so limited right now, Russia,

has been importing diesel from South Korea, like over the last couple months. So if your idea... Let's just pass the diesel around to everybody. Everybody gets a taste. And you sell it to that. We're going to sip and you pass it along. And we're going to resell it to Brazil.

We're just going to pass it along every single week. So if we were to give, let's give Trump the most credibility possible here. And let's say this really will be an increase. Like this was supposed to be a genuinely, a genuine increase in the supply of diesel available to help, like, reduce the price. Let's pretend it would be enough of a bump to make that difference. In reality, it's the same diesel that already exists in the global market just being shuffled around between those parties. So it

really makes no difference. I mean, that's true of almost all these sanctions, though. Like, even though I, you know, in theory agree with limiting Russia's ability to sell diesel as part of a thing, Like a lot of the oil and natural gas that Europe is buying in that 90% drop is like it goes from Russia to India and then India to, you know what I'm saying? Like everything gets tranchet because it's a it's not a non-fungible. It's a fungible token. Yeah. It's an FT. Sorry, I'm not really following this story.

How do you explain this but use NFTs? I'm trying to pitch my NFT here. Yeah, I mean, and presumably Russia is still just making a ton of money from selling gas, right? It's not like they're short on money to fund the war. That is not like the key problem. It's more it's just ruining all these international relations, right? And I mean, maybe you have a better sense of like the impact. I mean, I wouldn't say it's like so easy to fund the war.

You know, it's a long layering war and it's expensive and you lose lives and people. But yeah, they have an income and it's largely from selling energy. Yeah. So the best thing to have done would have not spike energy prices with a war in Iran, which by the way, I think we're on day, what do we say, 212 to 228 of the four-week war in Iran. They say 230 is when it usually wraps. Yeah, I think we're out of 230. We're going to wrap this baby up.

But there's been some additional chaos in the Middle East. It's like... Yeah, so the other big part of the story that is kind of unfolding in this region is the Houthis have been attacking a lot of refineries and oil depots in Saudi Arabia because they have had a historical conflict between them in Saudi Arabia for like the last. 10 years, I think more than 10 years now. And more recently, there has been attacks on Riyadh's International Airport. So there was one just yesterday that was a little smaller, and then one before that that killed 13 people and injured over 300. It's a huge attack of

this airport. There's like security cam videos of it. And this is, so there's a massive industry in Saudi Arabia, obviously, oil, right? But as time has passed, they are making efforts to diversify away from that. So there's huge, like, visit Saudi Arabia tourism campaigns. There's F1 races.

There's comedy festivals. All these things we hear about in the news, right? They sort of organically sprung up inside of Saudi Arabia. But really grassroots efforts within Saudi Arabia. A real desire in his heart to perform in Riyadh, and he just made it happen. It's very smash bros-esque. you know, like ultimately arose from the community.

And then the Saudi prince said, you know what, we'll kind of help out. We'll just give a little bit of extra funding. So Saudi Arabia has made this like multi-tiered effort to, uh, let's become a tourist destination, not just one for religious purposes. Let's become a hub for business,

like our counterparts in like the UAE or maybe like Doha. And let, let's build out our economy. So it has more presence, uh, globally and is, is just more relevant than, just being a hub for oil, right?

So the big thing that the Houthis want to do is attack places like the international airport to destabilize this vision of a future Saudi Arabia. Like if your airports are under strikes and people, like civilians are dying there, your tourism, your business conferences, all these events you want to host, people are much more tepid to attend and go to your country.

Was this like it was an airplane? This happened in an airplane? in the airport. Happen in the terminal. And so if you pull up, like I had read the FT article initially about it, but the whole side of this terminal like explodes as people are like waiting to board planes.

Oof. I guess, you know, the silver lining with all this, we've talked about it a little bit, and maybe it should highlight it a bit more of how many countries are dramatically increasing their green energy. And then something happened,

I think over this last week, week or two weeks. Germany like actually published their plan to fully phase out fossil fuels by 2045. So they've done that. and then France also, France and Netherlands have done this. So there, I mean, there's, there's, there's, you know, you talk about the 90% reduction. Like, there's, people are really ramping up renewable energies as a result of all of this chaos globally.

Yeah. So that's, that's, that's sick. Like, there is good things happening from this, which is that is dramatically accelerating us finally getting off of fossil fuels long term. That's why I've called Donald Trump the world's greatest eco warrior of all time. He does what needs to be done. Have you thought about like the, the version of this where, Trump, Trump like dies. Like he gets through the rest of the term.

He passes away. He leaves like a manifesto. And he's like, he's like, I only thought I could change the world in one way. And I wanted to foster a world of energy, like renewable energy and energy independence. I love clean energy. It has been my dream.

It has been my greatest dream. I realize the ultimate deal. De-stabilized the Middle East. Destabilize Russia. He signs it off. He's like, I had to be. I had to be the villain no one wanted to be. Like, this is like the fifth time we've done the Lushtrae.

I just don't think it's true. It could be true. The only piece that doesn't make sense is the Venezuela, because he does seem to want their oil. But... You have to throw them off the center. Oh, that's true. Obviously.

If it's too obvious, that Trump is clues, that he's a great person. You have to occasionally. Of course. But yeah, I think that the main thing here is like, as they are pushed to attack more and more like facilities in Saudi Arabia.

It furthers the like global oil crisis that's going on. And I don't, you know, I don't know how those things are going to get better, but I do know what's kind of getting better is people are finally buying bonds again. Well, that's not. Two closely related things, Brandon. I love that transition. This is not like a feel good. This is all tied in, which is that, you know, as diesel prices rise worldwide,

we've talked about this story. Every major economy is starting to see their borrowing costs go up and up and up and up and up. But if you think about what that means, that means people aren't buying their debt. They aren't not loaning the money because if they don't loan the money, you have to offer a higher amount. Then I'm like, maybe. And then I say no, then it goes higher and it goes higher. So finally, it seems like they're kind of like topping out a little bit, possibly in that there are a bunch of inflows into government debt. Basically, you might be someone who's got all your savings in stocks.

You know, you got a 401K, you put it all in stocks. But if you start to see, hey, I can lock in a 6% return for 30 years or something like that. It starts to go like, that's guaranteed pretty much. Unless this country collapses. I can, you know what I'm saying? And so for a lot of times, people will be like, well, no, that's still too risky or still too low. I want higher. But now it's getting so high that we're seeing a good chunk of like real wealth start to transition over into those government bonds.

And people are starting to really question themselves, like, is that a number I could live with? Because if it hits that high, if it's like, you know, six and a half percent or something, I'm like, I can lock that in forever and not have to worry about the ups or downs of the market, maybe I do it. And so that, that's really the story. It's like people are starting to bite. For a while they weren't biting. It was like a no bid. And now it's like people are starting to bite.

So maybe they won't go to catastrophically higher levels or maybe they will. So was there a concern that it's just going to kind of keep snowballing up forever? governments would just be like we literally cannot get anybody to loan us money. I mean, not anybody, but you can't get enough for the amount you need, right? But it's seeming like at least right now there's enough demand to stabilize these levels.

But again, it could, another, you know, things you get crazier. I'm just saying the cost of borrowing has risen in so many countries so high, so quickly that it's kind of rare in our lifetimes for it to all go up this fast as quickly and people are wondering, well, what is it going to stop? And now maybe it's, you know, this is an early session,

but there's real money flowing into actually buying it. If it's like six, six and a half right now, and that's where it's topping out, what was the norm historically? It's not six and a half right now. It's like five right now. Okay, but a little over five. But the norm historically really depends on

whether you're not in a period of crisis or, you know, you could say four percent or something like it's like a stabilized. But yeah, I don't know. I think the real question for people think, I'm kind of thinking like this. I'm like, hey, do I want to grab some of my savings and put it into this? super rock solid guaranteed, you know, 30 year thing.

And the truth is, how bad do you think inflation is going to be in the future? If they get it down to 2%, then you're getting a 4%, you know, if you're locked in at 6, a 4% real yield, real gain. And you get that for all that. That's really good. Historically, that's a really good thing. But if you think inflation is going to spike back up and you're locking in a 6% and it goes to 10, then you're fucked.

Right. So that's the big question. And I don't want to get to financial jargonie. I just think this is an interesting development, which is that at least there's something had, like people were just not, we're just not blinking. They're just like, it kept going up and no one was buying and now people are, some people were buying. So can I be perhaps a misinformed villain here? Sure. And, and say that is this, this isn't great news in a sense. Like, like, like, like, this isn't, this isn't, okay, this isn't, this isn't like, oh yeah,

people are buying the bonds at this high percent. It's, no, they're locking, like bonds have, uh, bond yields have been going up and people want to lock them at these high percentages, which are bad for the governments that owe that debt in the future. Like this is still really, like the market, the demand has met the supply, so to speak. And we finally come to terms on a new higher percent that people are willing to buy at in large numbers. But it's a number that we're going to pay a hefty price for in the long term. Not, and when I say we, I kind of mean the world, not just the U.S. Well, we can borrow more money when it comes to it. Yeah.

Yeah. We'll just keep doing that. And I'll hire, right. And what you're saying makes no sense, because every single time we've had to borrow more money, we just borrow more money to pay for that.

And it's interesting you say that. I don't think it'll stop ever. And it's, it has, and it's worked up until right now. So it'll probably work forever. So extrapolating it will work forever. Yeah.

Again, this is not like the, we've solved it. It's not like it's all bad. It's just a weird thing to notice. And I think people, anyone with money right now is starting to notice like, man, the stock market is so crazy. It's up and down. It's wild left and right.

Do I want to just, you know, when the stock market is average 10% a year and like the bonds are giving you 2%? It's like, I'll take the craziness for that 8% gap. But as it gets closer, people start to go. And so the real threat is if the bond yields start to rise again, that it can crack the stock market. The idea is that people start flowing their money out of the stock market. and then we start to see.

This has happened like, you know, bond yields were at their highest in 07. And like it, prior to big crashes before, it's been like people jump. And it starts the beginning of the dam breaking. So anyway. Well, when they start taking their money out of the stock market, it starts performing more poorly, will simply lower rates with no consequence.

And the money will just flow back in. I do not see what the problem is. Yeah. I think, actually, there was a quote from Trump recently where he was talking about the debt. And he said, I mean, you just said openly, we're going to inflate it away. But if you think about that, that is such a scary thing to say because that is to inflate away the amount of debt we have. It's like all your savings go to zero. All your, that's it.

It's a wild thing for a president to say. But he says a lot of things. I can't really, I can't see how much he believes in it. I really dish that up because I kind of blindly dish that up as a win. Yeah. Just hoping. Like I'm hoping it would be a pivot from like... We've got some wins in here, I think.

We've got some interesting things, but I think the opener was about how, you know, the Iran war is... Three wins in a row, ladies and gentlemen. Russia killing it. I don't want to use that same joke for the next story. And Bond's killing it. They're killing it. They're killing it.

Yeah. And this is all the consequences of a really, really necessary war in Iran, which I'm glad is... I mean, do you think that... Oh, okay. I'm going to set you up for... What I got, I hope, is a win, duck. I think it's a win.

All right. So somehow you need to, good luck. Transition from where we are to where I'm going. Because all this, all this crazy world that's changing technology, one thing is stood out to me a little bit, which is that video we watched recently about Neurrelink that allowed that guy to speak again. Yeah. And I thought that was very moving.

And you have more Neurrelink news that I'm hoping is equally positive. It has no consequences. No, I know. But how is that related to bond yields? Ooh, it's not. All right. This is going to be, I'm just going to talk a little about what Neurrelink is doing. So Neurrelink is a company.

It was founded by Elon Musk. And what they have been doing, I think, is one of the cases where you can look at a new technology and be like, holy shit, this is absolutely incredible that this exists in the world. There are, of course, caveats to this. And I will give some of the criticisms of this company. And with one of the most notable being, it's an Elon Musk company, which comes with all sorts of baggage. But in general, one, he's not like that. involved with this company day to day. He has largely...

He's got tweets to make, bro. He's the bank. He funds it largely. But I thought, let's just walk through a little bit because recently they started showcasing some cool things. So the origin of this is this 2016. Elon Musk founds NERLINK with eight other scientists. And the idea is, can we put a chip in the human brain that connects into the brain

via threads, like literal, like a sewing machine type of like needles that go into the brain where you can track the neural signals that are happening in the brain and translate that into computer motion. And the idea being, obviously, somebody who is extremely ill, paralyzed, has some sort of neurological disorder. This could help them regain some sort of major functionality. So starts at 2016. By 2019, it had $158 million in funding, but $100 million of that is from Elon. So, again, he's largely the funder of this. But he is admitted he's very, very little involvement with the day-to-day.

but they invent this like sewing machine type of a device. So I want to walk you through a couple notable milestones. The first is this one if you pulled up Perry. So in 2021, they start this with chimpanzees. And this is, I believe the name of the chimpanzee is Hager. So they install this chip into the chimpanzee's brain. And they have the chimpanzees start doing actions with a computer. They're moving this joystick.

And they over time are able to analyze the actual brain signals that are happening while the chimpanzee is doing this and learn what the brain is doing when it's trying to move in a certain direction. Oh, interesting. And as you can see, this was released in 2021. The monkey in this case, is monkey the right term? No, chimp, these are different, right? I'm way, let me cut that. You're canceled on monkey blues. I don't know that there's no monkey to chip. I'm sorry. Somebody's going to write a furious comment and you're, and you're validated. But one of the things here, if you notice, the controller is not plugged in. So at this point in their study. They already are at a point where what the monkey initially was using a joystick

to control a cursor, but they tracked the brain signals and over time realized they can actually just see what the monkey is trying to do when they think about it. And actually, so this right now, this is not plugged in. This is, it is controlling a computer with its brain. And the monkey's hitting that shit. Audio is pretty good at this game, honestly. He's doing amazing. Audio listeners, we're watching a monkey move a, basically a mouse across the screen to like track a little square. And now he's playing pong. And now it's playing pong. So Pager's favorite game was Pong.

And this is, again, it is controlling this through its brain because over time they picked up what kind of signals correspond to what the chimpanzee wants to do. And then if you're wondering about the like straw that it gets, every time it does well in one of these games, it gets a little banana smoothie, like sent into its mouth. So that's how they encourage the chimpanzee to do this. Which is something I've been thinking about we should probably do on this show. Yeah, yeah. Just to pick up performance.

That would be good to help improve our. He's so much better at Pong than me. He's really good at Pong. So you fast forward to 2024. 2024, Noland Arbaugh is the first human patient. Eventually, NERLINK is approved for human patients. So pull us up. This is Nolan.

There's actually, dude, go watch interviews for this guy. He's just a phenomenal human being. So he is paralyzed from the shoulders down, basically, fully paralyzed. And he is the first person that had a neuralink chip installed. And he, this is here a video of him, playing chess with his brain. The chip that they installed has threads into his brain, which over time they did a whole bunch of training

where he uses a computer and they learn what things he's trying to do when he thinks them, what he wants to happen on the computer. And over time, Neurrelink can actually change, you know, convert those brain signals into a computer actually doing the actions. So this came out in 2024 where they showed, hey, we have a real human who has this in their brain.

It is wireless, which is unique, because there are other companies doing things like this. And if you listen to some of the interviews of Nolan, it is so inspiring. He talks, one of the most notable things is he said he got really addicted to civilization of like playing that game. Yeah, I've seen this guy. Yeah.

Right now he's saying it's like using the force on the cursor. Because over time, he just is able to start thinking about what he wants to do and they're able to translate that into actually using a computer. So I just want to like, you know, step back and say, this is a person who's paraplegic, who's unable to do almost anything. And this technology is allowing him to control a computer. And these are just the early stages.

Another just quick one I thought was fun. This is from a talk last year where they now have like at this point like 20 something patients. And they show an example. This is a neuralink again of different uses that people are like people who have neuralink installed, usually paraplegics are doing. And this includes Mario Kart, Halo, Call of Duty. There's people playing Counterstrike 2 by the way, Aiden,

which you maybe play. against them. I don't know what rank you're in. People using CAD software to like design things. People making art again and using drawing programs. I mean, it is, it is incredible. And then this is, they're starting at the bottom right. You can see to obviously the hope is eventually you hook this up into robotics, right? You allow somebody to control physical arms in their environment and whatnot. So as, as, you know, if you follow this stuff, it's just, it's incredibly cool to see the advances they're making. So finally, I'm going to show one that they, uh, this is a video that they, they published a couple weeks ago. So as of this year, they have been working with ALS patients.

So ALS is less of, you know, there's this, let's say, physical injury in your spine, but instead your brains, I guess, just function is deteriorating, right? So ALS, this is a, this patient, this is their third ALS patient, but they posted a video about it where here's somebody who's this extremely extroverted person who used to do like public speaking, basically, and has lost his voice. And you'll hear in a sec, you can barely understand him. And the idea with this is we're going to try to basically, we put the chip into your brain and we'll figure out what you're trying to say over time and then convert that into text. And this is him basically using it for the first time. You see a sentence on the screen and you'll attempt the best you can to speak that sentence.

Our understanding doesn't matter. All that matters are your brain signals and your intention. Make sense? Okay, let's do this. So this is him basically training, right? Even though you can't understand what he's saying, he's reading certain snippets of text so that they can start to understand what is his brain do when he's trying to say something. And over time, they can then give him this button

where he can press a button and start trying to talk and they'll monitor the signals and convert it into text. Sound good? And press that sound blue. Hello, how are you doing today? Wow. He's so blown away too. Yeah.

And his reaction is so cool. And his wife is like, dude, it... You just got your wife's back. I think the beginning of this piece does, like, a really good job of setting the stage of, like, how much, you know, being able... Like, meeting people and speaking to him, like, really meant to him in his career and, like, how he had to... He sidelined the whole thing, like, when he got his diagnosis, right? like he put everything to his job and it was about like presenting to

and like meeting people and or rating all the time and he lost that. So it's like this, this massive piece. I mean for anybody like speaking is important right but for him even more so I think. Yeah. And he,

it's like a piece of his identity being given back to him. And he's so, it's so cool at the end you see him. Pull it up. Yeah. Um, oh, oh, oh, what is?

he's demonstrating this with his wife what you're thinking the words and they're popping up I love you so cool it's just wait that was him just thinking it was him thinking it

he's able to say I love you to his wife after not being able to speak it's like dude I'm like tearing up when I watch these videos it is it is amazing so yeah I mean getting

I think this is an example of like telepathy almost Like the version of this in the future, I'm not saying everybody's fucking lining up to do this. But if you wanted, you could have this chip implanted in your brain. You have this chip implanted in your brain.

And I could speak to you without us actually talking. Like we'd be able to communicate at a distance. Right. Or like I think all of these examples, I think it's really easy to see stuff like this, especially the video games in the same way that something like AI was being tested on like Dota and StarCraft. and go early on. People are kind of wondering like, you know, what does this actually mean?

And the practical applications of what these mean like way down the line as this technology gets better and better and more available is simultaneously amazing and scary. Like the amount of people that this could could help is unreal. Yeah. The wide range of medical conditions and problems and man, just getting, again, I encourage, if you're interested in this stuff, watching interviews with Nolan Arbaugh, who was the first neurologic patient, and just his experience talking about it and what is like, at first time, he's just chill-ass dude, but then his experience of how his life opens up when, after years and years of being paralyzed, he's given this opportunity and presumably it's just going to get stronger and stronger is,

is really cool. So, man, you know, in the medical, I think broadly in the medical space, there's still like really cool stuff happening right now, even though technology broadly, I think, feels a little annoying and scary at times right now. I mean, you have some caveats as well to this, so it's not all. So, you know, so just some quick ones just so we acknowledge. First off, Neurlink has one competitors. They're not the only person to do this or only group, but they're the most kind of notable right now. But even if you just consider that they're notable, like it's pushing other companies to try similar things. So for example, Black Rock Neuoreh is a Chinese state one that's doing similar things. Syncron is a New York City-based American company.

These are all doing it in differing ways. Another big criticism is basically, as you can imagine, Elon has run the company in a very Elon way. So there's a lot of criticism about rush timelines, rush timelines that result in things like animals not being treated as well as they should have been. Within two or three years, like most of the co-founders had left because of how he was running it. and a quote, toxic environment. So it's a bummer because this is like such an incredible thing that is happening. And it's like, ah, Elon's still doing Elon things behind it. But this is just, this is fucking,

he's not touching a day to day and just supplying money. He's largely just funny. That's kind of cool. Yeah. I don't know. This fucking rocks, man. It's a real feel good thing to see people's lives being fundamentally changed. The super cynical villain chair argument I can bring to the table is,

uh, this is part of the, the Peter Thiel, Larry Ellison live forever in my, my digital world never die. I don't need my blood boy even anymore world, you know? It's funny. You say that. Well, somebody think of the blood boys. Do you know what I'm saying? The blood boys will be out of the job. It was a growing industry and now they're all getting fired. And this doesn't seem like it's going to be at that level. Bonnie, you say that. Okay. Because Elon's stated goal is to turn this into like a neural lace device where he says it's something analogous to a video game, like a saved game situation where you are able to resume and upload your last state.

Jesus. So he has been explicit that he made the company and the starting point of the company purpose is to address, quote, address brain injuries or spinal injuries and make up for whatever loss capacity that somebody has with a chip, which is where it's at right now. And I think it is a phenomenal use of human technology and ingenuity. It is phenomenal. And then long term, his goal is to achieve symbiosis with artificial intelligence. So that is the downside. See, that's the problem. I think. I think that's, this applies to so much tech on the show is, is like, oh, these, these human applications that, like, emphasize or augment human life in, in a way that, like, helps people are so good. But then there's the other version of events where I'm strapped up to the machine in the matrix, and I, and I don't know I'm in the tube.

You get unskippable ads. It's this weird? And so I can power Larry Ellison's fantasy. So I'm fearful of that world. But when I watched that fucking video that you just showed us now, like when I watched it the first time, I fucking cried too. Like look at how much it means to that guy.

And is Larry Ellison weaponizing my emotions so that I give him eternal life so he can fire his blood boys? I don't think he's on this one. Yeah, you keep venting to Larry Ellison. I'm being hyperbolic. Because Larry Ellison, like, wants to live around. He is aspiring to put all of the Paramount media into our brain in an unskippable video.

Yeah. To be clear, I don't think Larry Ellison's going to make it. He's too old. But Peter Thiel might. Peter Thiel might be at the cuss. He's got so many blood boys you can get to the escape velocity of. And outside of the fear of all the blood boys that will lose their job because that is a concern. That is a big concern.

And I mean, he's even, he's pitching the idea. So Neurlase is from a series called the Culture Series by EMM. Banks, who is a science fiction writer who create books. And the idea is basically humans have a, in, you know, in this far future, all humans have this chip in their brain that allows them to communicate with the broader AIs that run all of society. Yeah. And that you can't- Science fiction written as a positive version of the future or a parable where it's like the torment nexus meaning where they see this and go, it's great. It's, isn't this like Warhammer 40K lore? Isn't it? Am I? Never mind. I mean, it's open-ended in the books. But that's why I like the books.

It's a really interesting thing of like it's basically what if we achieve utopia type of vibes. And then it's very like brave new world ask of what is that even good? It's that type of vibe. It's like you have soma in your brain where you can. Yeah, that doesn't sound. Yeah, we all get, we all get early. You just talk about that earlier. And that alone's scared.

I'd send you crazy shit too. I'd send him crazy shit. Like I'm just trying to like read something and then Aiden can pop up in my right or an ad or an ad or something. I would send you a lemon party in the fucking telepathy group chat. It doesn't sound good. I think this is really cool. Like if you blinked, if you close your eyes, it would still show up.

You could see Lemon Party. Because you paid for VIP. I'm at like Trader Joe's like in the eye. I get flash banged by Doug's message. Because I will pay the $8 for the blue check that allows it to guarantee. Priority eyeball. Yeah. There's a premium tier.

Yeah. I just think Nintendo DS good phone bad is my solution. We always have this great tech and we take it a lot. little too far. Do you know what I'm saying? Nintendo DS good. It's good. Phones bad. We should have stopped at one. We should stop with this. We should solve these guys problems and stop there. And then no ads in the brain. I just, we always take it one step too far, bro. I think it's cool. But I'm, I'm worried about the implications. But it should happen. It should still happen.

More for this show comes from Shopify. Is it Black Friday coming out? Friday is coming up. I was just thinking the same thing. Wow, you and I are mine. I bet you, I know what you guys would like. you guys love little peepo emojis, don't you? On Twitch. Aidan, I do love that. Yeah. I want you to imagine I've taken your favorite peepo emoji.

Okay, my peepo emoji. I've taken it and I've embroidered it on a sweater. And now you're like, where can I buy that from you? How would that have been possible? Well, a few years ago, when I made this sweater, I sold it using Shopify. It was very easy. I basically took the goods, stocked in the store,

added the photos, all the product information. it tracked the inventory of the sales, and then we launched the store, and thousands of people bought those sweaters. You're like a PIPO prints, right? This is true. This is true. If you're ready to launch your business

before Black Friday, and it does not have to be PIPO embroidered sweatshirts, it can be any product. I think you want us all online. Go to Shopify.com slash lemonade to start your free trial. That's Shopify.com slash lemonade. Go to Shopify.com slash lemonade today to hear your first

to see. You can sell whatever you want. You can sell whatever you want. It's got to be legal. It's also got to be legal. You're about snaggy for when Doug is doing the Ching Sound. With Export Development Canada, Canadian businesses are competing, leading, and winning in more international markets. Like how agri-food businesses are strengthening global food supply and critical minerals are powering industries of the future.

Their ambition can't be contained. And with EDC behind them, neither can't. their growth. Export Development Canada is the financial partner for Canadian businesses going global. Visit edc.ca for more. I'm Phoebe Judge, host of the podcast Criminal. Usually I'm telling stories about other people, but this fall I'm headed out on the road with a brand new show. Stories about me and the people I love. It's called Don't Worry. I hope you'll come out and see it. We'll be in cities like Boston, Chicago, Philadelphia, New York, and lots more. Get your tickets at this iscriminal.com slash live.

You know what? All those unemployed blood boys might need. What do they need? They might need to be on universal basic income. Whoa, what a transition? Aidan, what will we do once our brains have been sent into a culture hive mind? Thank you, Doug. Thank you, Doug.

We'll all be on a universal basic income program. Oh. And there was a beautiful article in the New York Times recently called What a Year of Free Money meant. Softer diapers, Clean Laundry, A Visit to Six Flags. Basically, a story of a journalist following three different people and their families and their experience on a UBI program that is going on in California. So what's interesting to me is Elon Musk, again, other tech leaders, including Dario Amadai of, Anthropic, Sam Altman, have all made comments over the last few years as their AI-related businesses has flourished that UBI should be a part of the solution in the future to offset

the unemployment that will come alongside the technology. And between the three of them, there's been various ranges of how UBI should be like applied, what their opinion is about it, it's changed over time. But what I didn't know is that these companies that often work in AI, and Google is the main one mentioned in the article, have been funding programs in California to study the rollout of how the programs would actually work. So in Google's case, they work with like charitable organizations in California to give paychecks to select groups of struggling people and just see and monitor how they spend the money and like how how successful they are after receiving like a year of you know 1,000 to $1,500 checks and currently that's capped or limited by I think

state laws like the amount of money you can choose to give somebody but the examples that it's giving are people that like get a phone call they're on some sort of program they've basically been like lottery selected uh albeit pre-selected for their circumstances going into it and they're like hey we're gonna give you like $1,300 a month this year and they just monitor how they spend it and this story uh it's interesting seeing the the contrast between a couple of these people so one is a single mother she has the like the support of her family this money allows her to uh move across the country to like relocate with like another family member she does initially, like relapse.

She has like a job, but then like relapses and starts using drugs again. And then manages to quit, quit her job, exit community college temporarily, gets like back on track, finds a new great job. All of this money is like a foundation of stability for her and her kid through this like trying period in her life. and at the end of that year, she ends up locking down a much better job with pretty decent pay, good tips,

starts going back to community college again and moves into like a bigger house. I think the term and the article they use, like people monitoring the success of these programs, they call it the cliff, like how people handle like the period after they lose the money. Right.

And oftentimes people who have, have access to this money, it seems like the majority, are, use the backstop to put themselves in a better position by the end of that year. They aren't necessarily spending it on things I think people would think as important for social programs to support either, right? Like, it's not just food and diapers in these things. Like the one example is, you know, this woman and her kid go to six flags and she's able to afford that and take her kid out for a day. And that's, you know, something the money allows her to do. But on the other end, you know, the middle story is about this guy who, you know, he buy,

one of the first things he buys is like a PlayStation 5 to, you know, do, to be entertained, like, while he's on his, like, dialysis. And I would not say he's able to put his life together in the same way through that period of time. And he unfortunately ends up dying of cardiac. arrest, like during the term that he's on, on the money. I think there's no, like, these studies are in progress and there isn't like a clear answer of here are all the benefits and downsides of these programs yet, but largely seem beneficial when people are allowed to spend this supplemental income on what they think they need. And that comes in a bunch of different

forms, I would argue, of how you live a generally good and dignified life. It's not just like the base essentials. It's this cushion that allows you to do a bunch of different things. And I think the first woman and the last woman mentioned in the article that have kids, they really are able to turn their lives around because of this program that was given to them for the year. Yeah, it's fascinating story. I strongly think this is the right. I strongly think this is the direction that means have to go. And I think when you start to add limitations to it,

it actually makes it a thousand times worse. I think there's no way you can predict or plan what things people will want to spend it on. And I think the best part about it of UBI type programs in general is that people that need them spend the money. They use it. And it doesn't get hoarded.

It's not like most, you know, tax benefits, which accrue to those who have enough surplus that they can hoard the extra wealth. People spend this. they use it. And whatever they're spending it on is good. It's like it's good for the economy. It's good in general. I think people who are fall through the cracks in the system and have, you know, that spiral they go through hurts everyone. It's a, it's a negative spread. Giving them the ability to have stability is a win-win to people around them to themselves.

I think it's, I don't know. I'm, I don't know the exact program I want to advocate for, but I think this is the direction things have to go in some way. Did this article address some of the common criticisms? And one of the ones I hear most often is, like, we already have programs like this. It's called welfare. You know, we have food stamps. So we have all these things. Would you address a bit, but.

Yeah, it doesn't, I don't think it dives into those things past personal experiences. And the way they, they tie into those systems. It's not, it's not really spending a comprehensive time. Like, well, here are like the tradeoffs. It's more about these people's individual story. I like this line here from the, I'm reading the article now. I hadn't seen this before, but it was like the payments, is a quote from them,

can be the kind of thing that pushes them out of poverty and it's some kind of stability. Yeah. And one of the things they've been learning is how one very small thing can become a giant thing that derails your entire family's life very quickly. An example would be like, trouble with your car, you know,

and then having to pay a car repair bill. No. And that bill now makes it so you can't make a rent payment or some other payment, which causes everything to derail and causes you to throw off. which is, again, not only a net loss for that person, but a net loss to everyone or, like, the whole community around them. Like, it's, I think it, I think it, uh, cascades. Yeah.

So giving someone the stability to, to weather those shocks, I think he's a net benefit to all of society. Would you say it's a good idea to give every American $5,000? I would say Trump is on the money with that. He is, he is still doing that. He tweeted that this morning. It doesn't tweet much, but he's been really pushing it. Really pushing the $5,000.

And probably because he read this article and feels like, do that this could really be beneficial for a lot of people. He was moved by the New York Times. He was moved by the New York Times. He was like maybe the perfect news after all. Yes, sir. Well, I guess the last part of this I wanted to loosely discuss was, you know, what is the motivation on the company side for choosing to help fund these programs?

Like a company like Google putting 50 million into this program, what is the, what is the benefit? And I think the good faith version is the altruistic version of the company that is interested in seeing how they can support some sort of program like this or prove that it does work so that it becomes legislation later on. And then the more cynical version is this is like a hedge when in the future they're confronted about job loss related to AI. they get to point at something and say, hey, we gave this a crack, or we're doing our best to, like, help combat this side effect. See, we're a good faith actor. I don't really know outside of those things,

and that's really just me thinking about it for like 10 minutes. Yeah. Even in a cynical take, though, I mean, it's like, you know, I guess I appreciate someone doing this study. The study takes time and money and inexpensive, and I would like to get data on it. And so it's, I guess, good that it.

it happened. I want to spend more time, like, looking at the actual studies of these things effects. I've, like, read in passing that these programs do help a lot, like, at a statistical level. Yeah, Rutger-Bregman talked about it a bunch. Yeah. Yeah. And, uh, there's been cities all over the world that have, like, tried versions of it successfully. I think my, I guess my worry is, like, what is this look like at scale? You know, when you, what, you, like, joke about $5,000 payments, Five, but the thing we always bring up is like, oh, like, that's an extraordinary cost. Like, how does that make any sense? How does that affect inflation?

So how would a program like this be laid out? We showed all over Trump last week. And now we're talking about UBI being great. So I don't think the $5,000 check for people that vote for Republicans is the same as Oh, you clearly see, that's how you can tell he didn't read the article. We went, we went off for a bit because of a tech issue. I know that it's as a conundated. condition of Republicans being voted. So just Republicans are the only ones who have the balls to

pass a socialist policy like this. I vote Republican every time. And you won't give them a win. And I will say for probably UBI reasons. I want my third Trump vote waiting for the UBI. I vote for him because I want the government to own large stakes in American companies and for UBI. But you know who are, you know who needs UBI the most? And to move the world on for. from fossil fuels. All of these things. Yes. These are greenest president.

I can think of the people that need UBI the most. And it's creators. Yeah. Creators need UBI. Well, they sort of have it. It's actually a small story. I just want to cover it real quick because I thought it was an interesting deep dive

by the F.T. Into what's going on in the crater economy. And what they found is that in the past few years, whatever laggard brands had not caught up to the fact that creators are where everyone's getting their information and news and entertainment now have all caught up or died. And so now the amount of money being sloshed around by brands is pretty enormous at the high end. And what they discovered is that it's a real,

what do you call the curve when it goes like this? Big. Parabolic? Parabolic. It's a real parabolic curve of payouts. It's exponential. It's not parabolic. It does go parabolic. Like, yeah. Oh, we're so dumb. It's, but it is exponential. It's half of the parabola. Like,

by the way, the next story is mathematics. Like, go ahead. It's going to give you a taste. To solve those. On that note, if you're listening at home,

maybe go through and fact check it all. Like, yeah, yeah, um, but what they were saying is that,

uh, at the top end of the, I guess, viewership curve. They, that brands have figured out that everyone is faking likes. they're faking followers,

that there's many, many ways to be disputed. So they're just using average views right now is for like most channels and creators. And the ones at the very high end are getting like enormous returns on brands. But the problem is there's now a really strong incentive to not disclose it. So the next phase of like this being so important to brands

is that since the FTC doesn't seem to care and no one's really following up, there's a lot of pressure to not have like, here's the hashtag ad sponsored segment. And it's like at the highest end, you're seeing a lot of big creators no longer just take sponsorships, but take ownership stakes in companies. They're getting pieces of it. And then they promote it like technically organically, but they have a financial stake in it. Yeah, I guess there was an interesting piece on the changing economics of creators.

I don't know if you guys have a question that I can answer. That's basically the idea. Well, does that mean we can kind of just draw? the ad sections from our videos. Like we'll just integrate them passively. We won't tell the audience. We just take a piece of, I don't even want to say. Of tax preparation companies.

And then I do recommend checking out Neurlink if you have a family member. Struggling with ALS. As eyes are glowing. Are you making deals on the side? I'd be sick about it's in our show. I was taking separate host deals that we, um,

it feels like enforcement is getting a lot. Well, like, I'm just seeing more and more examples of, especially on vertical video of people just, it's clearly a sponsor. They just don't disclose it. And it's interesting because like when we started this gig, it was like, you got it. You make sure to put hashtag out. It's a big deal. And like you would sign all these contracts that would, like every contract I would sign for a YouTube ad or whatever, there'd be a whole section that's them just copy and pasting all the FTC rules just to make sure that their liability was covered. They're like, you have to know all of this stuff. And I worked that side of the deal when I did brand deals for Invidia or Twitter. and we would push hard to ensure that they follow the rules. But now all the brands are incentivized not to do that push. And they kind of want to look the other way and be like, we kind of have to say this, but do what you have to do. And especially with short form content where there isn't a lot of time to capture user attention and disclose,

they kind of just want you to be like holding the bag or, you know, using the product and just fitting it in. So it is a weird thing. I'll also say that the takeaway that I had from this piece was that the majority of creators making a living, like above a median income or whatever, in the creator economy, are making the majority of their money from these deals.

It's not from like, you know, AdSense on YouTube or whatever. The payout is from the platforms. It's like sponsorship deals. That is where the majority of the money comes from. And for a lot of them, it's like a very high percentage of their income. It's like mostly from sponsors.

My friend was doing Instagram food stuff for a while and she managed to turn it into a gig where she was making like 60 to 70K a year for a while and this was her income stream because the biggest platforms are the short form video platforms but the base ad sense on those things isn't going to be very good like no matter what you do so this is the way she monetized her work uh you know the regulatory environment doesn't exactly have a ton of teeth right now but I feel like that the only way you can effectively moderate this is passing the buck to the platform somehow. Like you would need to change legislation so that the platforms are responsible for

investigating and policing this, right? Like, it's super hard. Because you can't, I feel like the scale at which this is happening is just impossible for a government agency, a single government agency to oversee. Yeah.

And like if Aiden does a get ready with me makeup tutorial and he pops open to makeup and uses it, like there's no way to tell from the outside unless you admit to it or they're a paper trail on your finances, whether you were paid for that, whether they send it to you, whether you can disclose it or not. And they can't really check it. And they're having it at scales. The other thing that I wanted to mention was that, you know, the takeaway is that we're in the era of brands getting really safe. Like they used to be like not knowledgeable about this stuff. Brands used to be like, I don't really get the TikToks and we're going to spend some money and we don't know what they're doing.

And like, what they're trying to unlock is that era is kind of over. And there's a lot of very native, savvy people working at these companies. They're trying to get the best out of their results. And so they're doing things like they'll get a big creator and pay the big amounts. And then they'll have this massive, massive network of much smaller payments to everyone in that micro niche around that creator that can like, it's just, they're like a lot better at diffusing these ideas. I don't know. It's kind of like just if you think something feels like an ad. it probably is most of the time.

Especially on short form media. It makes me more wary. Anyway, Nerling is great. Yeah, Nerling is such a great company. No, I do want to say, back on the UBI thing really quick, there was a story I wanted to mention, which is, you know,

some real jobs are being lost in North Korea lately. Okay? And it turns out there was a big takedown recently on hundreds of North Korean employees, who were actually working at remote jobs in U.S. tech firms to help get funding for things in North Korea. And this just happened. There's this big takedown. And the scale of it was pretty big.

And what they found is not only were hundreds of companies affected, they had a lot of people who were North Korean people working. But people in North Korea, they need a U.S. counterpart to help pull the scheme off. Like they can't just sign up. So there's like people making a lot of money in America, like selling their spot or VPN access or whatever as an American citizen to hundreds of different Korean employees. Like there was one person in Texas, one woman in Texas, who had sold her identity or whatever to 300 people in North Korea,

who she had all these jobs. They were all remote jobs and they were all working them separately. So a person in America works at a time. tech company and then gives the 300 tech companies. And then gives the credentials to somebody in North Korea. Yes. And just takes the pass-through fee. And they're all working the job. But the- Isn't this like there's two versions of this, right? There's the version of the person who like sells or like allows their identity to be used for this purpose. And they're

like scraping a little off the top basically. And then there's the other version where people have had their identity stolen for this. And they aren't aware that that their identity is being used for jobs. I remember there was a guy who he was trying to get a new job and they pulled his employment history or something. And they were like, you worked all of these jobs in nine years. And then he looked at the list and he was like, I've never worked any of these places.

So there's people able to take advantage of this, but also people being victimized by it, I think. I mean, all come clean. the people doing it are claiming, and I think probably in some way are being victimized because it describes that the appeal for these, they're called facilitators, the ones that help happen. The facilitators. I look like a word for what you're doing.

Yeah. You know, it's pitched very, very low lift initially. And it's like, you know, very little is asked of them. And it seems like something they can kind of like make a little extra money. and then because they're in too deep at that point, already it's like they've committed a crime or something, they start getting more and more asked of them and they can't really say no without revealing.

And so it starts expanding from like one person kind of helping with your job and you get money to like you are now signed up for hundreds of jobs. So it's like this weird rabbit hole. Anyway, they did a sting and they took down a lot of them all at once. So how many did you say? Let me see. There was thousands of employees in Korea

and I think a few hundred facilitators they caught in America. like what were the, did they list like the companies those were at they were doing work for? The scheme built companies out of millions of dollars from 2019 to 2025. The identities of the defendants and the companies are under seal. So we don't know. Under seal.

So like nobody could see the companies that. Mogul moves, I believe, is on here. It says mogul moves. You just said they were under seal. You just said there were under seal. There was only one they're going to reveal. There would be no reason that mogul moves would be on that list. And it's mogul moves.

They just reported it. Ludwig has actually been. 300 North Koreans. Redder North Koreans working. He's never said any of the jokes on his own. This is just the North Koreans. It's actually a lot of North Koreans in a trench coat. That's what I love is the amount of North Koreans that you could be at any given point.

It's not one to one. Yeah. These laptop forms. 300 U.S. companies. So we don't know. 300 companies? 300 U.S. companies. I thought it was 300 people.

No, no, no, no. Thousands of North Koreans working for 300 U.S. companies for more than eight years. Dude, every time I'm working with somebody remotely, I've never met some of my editors in person. It could be 300 Koreans. Just to be clear, they're not having 300 people be in one person. Tolek, I need you to send me photos of yourself right now.

And he's like, hey, just so you know, I live, with a whole bunch of people. It looks like there's a couple hundred of us in my bedroom. I know it, Tulek. You're 300 Koreans. You got him. Can I say your company name? Uh, no, no.

Doug, Doug Corp. Yeah. That's gonna be on the list. Doug, that's gonna be on the list. Doug, as far as I'm aware,

there's just eight of us to, but as far as you know. It could be 2400. Yeah, you get paid not only to give your spot, but you have to appear on camera for job interviews or meetings. you have to verify your fraudulent identity. You have to ship packages from your home in the U.S.

And then you have to complete urinalysis tests or fingerprint submissions. Wow. Okay. I'm throwing this out there. If I have the opportunity, I will do it. If Ludwig invites me to another collaborative thing, you will let a North Korean. I will have a North Korean show up in my place. That sounds great.

Yeah. I will send Ludwig the urine or whatever he needs. He always asks for a urine sandwich for any of his collabs. A lot of people don't know that about Lerick. That's a big, you just stick over that. There's the required urine sample before every stream. I was supposed to be chess player number 10 in that video, but you guys beat him. I was ready to show up with my whole team.

You didn't pass the urine test. I couldn't pass the urine test. Everyone's talking about the prospect of a democratic wave in the midterm elections. But here's something that matters just as much. The battle to control state legislatures. That's the most important thing we should be focusing on right now. The GOP has a trifecta in 23 states, which means they control. the governor's seat plus the state house and the state senate.

For Democrats, it's 16 states. Meaning that the opportunity to pass the policies that the majority of our state supports, things like increasing minimum wage, legalizing marijuana, codifying abortion. If Democrats flip just a handful of state legislatures, the majority of Americans will now live in a state under full democratic control,

which could have big policy implications. That's all on the two. with a Democratic trifecta. This week on America, actually, we're in Wisconsin, and we talk to Francesca Hong. We have an opportunity through Wisconsin to actually start improving the brand for Democrats. Catch us every Saturday on YouTube

or wherever you get your podcast. I'm Seth Matlands, and on this week's episode of Creator Destroy, reimagining marketing, I'm talking with streetwear legend Jeff Staples, who in 2005 designed a Nike collab, which was the Pigeon Dunk, that caused a riot,

resales for thousands of dollars 21 years later and gets the credit or blame for bringing sneaker culture to the masses. In our conversation, we talk about why you can't buy your way into culture no matter how big your check, why 40 hours of meetings a week with none of them talking about what the customer actually wants is the most common way he sees businesses destroy value. And why, when I asked him what he'd destroy with all the power in the world, Jeff said likes and comments, because if likes and comments existed when he started, he never would have made the next thing. Creator Destroy. It's a business show, a marketing show,

and the show that argues they've always been the same thing. From the Wisdom is Company and the Fox Media Podcast Network. What if we told you there was one thing you could eat to reduce your risk of death by about a quarter? Interested? And what if we added that this magical substance can dramatically lower your cholesterol, improve your mood, and keep your weight stable?

No, it's not protein. It's actually fiber. the least sexy, most overlooked nutrient of all. But what is fiber? How could one nutrient do all those amazing things? And how can you get enough? Is a powder you stir into some water going to do it for you? All that plus the strange tale of the gorilla poop obsessed father of cornflakes in the latest episode of Gastropod.

Find gastropod and subscribe wherever you get your podcasts. Speaking of pretending to be other people, character.a.i shut down. Let's hit that real quick. This is an interesting one. From a billion dollars to shut down. The character AI story. You know what character AI is? No.

Okay. Character AI is a website that is entirely chat bots. And the idea is this came out in 2021, right? I believe it's when it started. Early chat, GBT. And the whole idea was like, hey, we're going to make all sorts of like fantasy characters that you can talk to.

Or real ones. Or the Great or whatever. It's like you can talk to Napoleon or you can talk to. And like now it's, days, that doesn't seem novel in any way, shape, or form. But at the time, it was like, whoa, I could talk to like these personalities. And coincidentally, a lot of them are very sexually charged. And a lot of them are like, this billionaire, who's your boss, texted you salaciously. I looked at one of them this morning. And I was like, this is, it's just like smut,

half of it. Yeah, yeah, yeah. That was where the money was. I think. They, they popped. They popped real hard initially. Yeah, I mean, I don't have too much more than that. other than that it is like, you know, some people are claiming this is the first sign of the pop, the collapse. And I would say it seems a little hyperbolic, but it is interesting that the froth of this getting a billion dollar valuation has evaporated to the point where I guess it's like,

at the very least, it's people realizing where the real economics are are starting to be figured out in AI and not like everything with AI and its title instantly is worth a billion dollars. Like this clearly wasn't better than the model at anything and able to stand on its own? Yeah, one of the things I think is actually interesting about this. Perry, I just sent a link if you could pull it up.

So the company character AI, and there's another one, I think PolyBuzz is the name, is really good at getting people to use the website a lot. And so there's this graph on TechCrunch that they showed people average two hours a day, like the average users. If you scroll down, you'll see a graph. It is a crazy amount of usage, yeah, right there. So compared to something like chat GPT, character AI and polybuzz have just an insane amount of engagement. So the people who use these websites and talk to these characters and develop these relationships with these AI characters are really, really into it.

And while that sounded great in 2022 when AI was really starting to pop, the problem is that they can't monetize it. And they haven't been able to. And it's as it says right there, it's a double edge sword because it costs more. If a person is sitting there for two hours a day talking to your AIs, and then they tried to, you know, start incorporating images and videos and whatever else. Like, each user is costing you a ton of money. Like, it's not even good to have this much engagement. These businesses are just rappers on models.

They just put a prompt and says, pretend to be Napoleon or pretend to be sexy boss. And that's what you filter the questions through. And so they have to pay it for the tokens that you're using. Yeah. And so they barely can. My understanding, I read a pretty good piece on all these rapper businesses. all of them operate under the gym model, which is like they make most of their money

from people signing up and using it very little. But the power users are all net negative. They're all, you lose money on the power users. And so for a site like this, where the average person doesn't give a shit about it and a few people use it a lot, it's like a horrible business.

A horrible business model. Have like a really, really passionate customer base, which is very funny. I mean, the other piece, so it's worth acknowledging, which I think is particularly interesting about this.

The company caught, Character. AI was founded by Noam Shazir and Daniel DeFratis. Don't know how to say the last name, but Noam Shazir is one of the guys who co-wrote the article about Transformers. The thing that has kicked off modern AI, this guy's one of the co-authors. So he's considered one of the most influential AI people. I just, I had a realization mid-sentence. Oh. Okay. You mean like the robots in disguise or what? That would be stupid to think that. I did, I knew what Transformers meant the first time. He thought the Michael Bay movie. He thought shy of the buff was involved.

I didn't have anything to say. I just realized. Okay, cool. Transformers for a hypothetical viewer is the linear algebra algorithms underpinning large language models. And if you didn't know that, you're fucking stupid as hell. I did, we did.

Noam Shazir invented Optimus Prime. Thank God it wasn't in Septaicon. Do you know what I'm saying? Thank God he started with one of the good ones. So one of the like, consider very important AI researchers who's very influential. He was one of the co-founders of this company. And they were trying to develop their own models basically

and be more sophisticated than just a rapper. But then in 2024, we, I think talked about this. Google paid character AI $2.7 billion to license their models, which really just meant they were paying Noam, Shazir, and Daniel to come back to Google. So Google paid $2.5 billion basically to get the two founders back at Google because the AI race is so big.

And since that, character AI just plummeted. So the whole story of this company is just like so strange. They like pop up out of nowhere being like parasycial chatbot company. They lose the two co-founders because Google's willing to pay $2 billion. And now they can't monetize and they're shutting down. And then the funniest part is Noam Shazir just left. He just quit Google. and went to Open AI.

All these. So they paid two and a half billion for a guy who left. The big AI name founders are just jumping around and getting the most insane paychecks over and over. I'm seeing through the matrix here, bro. Every time it's like, well, I want to be more aligned to my foundational model missioner. They all have some statement. But the truth is that the bidding war is so high that every year and a half they can jump ship to a different one of the big three and get a fucking half a billion. I think he's aligned to the Open AI IP.

P.O. Yes, I do. I think so as well. He's very aligned. The financial incentives are just too obvious. I don't believe it anymore. All the people jumping around, they all say this. It's like, it's like when I use it working in Silicon Valley, you'd see someone at exactly their four-year vest mark say like, you know, I'm going to explore new opportunities under LinkedIn. They'd move on. It's like it's too it's too lined up. I think it's both. Anyway, Disney bought them for $20 million. No, no, no, no, no. Part of it. Only part of it. The rest is trash. I mean it's like going. I mean, that's the question. So they announced bankruptcy and they're like, hey, we want to sell the assets. Disney bought some assets for $20 million, but the rest, which is in theory worth hundreds of millions, they're trying to sell still. And if they can't, character AI is going to shut down, probably for the best.

I'm not a big fan of this type of... What else could there be? What do you mean? Do you know what I'm saying? Like, the company is essentially a rapper on Chad GPT with some characters on it. If you spy some of the character names and assets or whatever, what is left that you're, and you don't have the talent anymore?

What are you? I mean, what you have is millions of people who are really, really, really, really passionate about your thing. Okay, you get some users, okay.

I mean, you have users. And it's not, it's not literally just, hey, chat, you can be T pretend to be Spider-Man. Like,

they've, you know, once you get people in, right, you build memory for each person and you have the models be more agentic and whatever else, right? So they've been adding these features. That's not to say I think it's a good product.

I don't. and it doesn't make money and it's parisocial and I think problematic in a number of ways. But it's not like there's literally nothing there. Like they, they have managed to capture the imaginations of many, I think. That's the Disney mission. So I guess they're aligned.

The funniest part, Disney, the new CEO of Disney just hired the former CEO of Character AI to be their CTO. So just last year, wow. Last year Disney, Disney sued Character AI. Disney.

Last year, Disney sued character AI and said, hey, you're doing copyright infringement. You're allowing your users to talk to Disney characters. Yeah, they're talking to Elsa and Darth Vader. Disney hired the CEO to be the new CTO of Disney. Interesting. How fucking weird this all is. And then they spent 20 million to pick up the pieces that they wanted. Interesting.

Which I don't even know what the, I don't know what you're, I don't know why you buy $20 million of this company. What is, is there any, like, a lot of these AI companies are, rappers on existing models. Not to say that the work that you acknowledge doesn't exist, right? In the long run, is any company that is a wrapper of a model going to be successful? I feel like you have this like core weakness in your product, which is that like somebody else can basically just make it.

That is one of the big open questions. And it is certainly true that many people have been able to. make companies that generate crazy, crazy VC investment money and whether those actually exist long term, who knows? Because if the whole, the core of your business is built around an AI company's model being really, really, really good, well, it's exactly what you just said. There's not a lot that prevents you from being replicated by another person or replicated by the AI company itself, right? I feel like you're immediately in the situation that a company like Adobe is kind of in right now

where like people, oh, people can so quickly like decompile and remake your copywritten software that your whole business model is under threat. Yeah. And this is that, but it was was more obvious even earlier because it didn't have, you didn't have to like decompile anything to do it. Yeah. I mean, there's other problems with this. Like they, they have gotten a bunch of lawsuits. There was like a 14 year old kid

committed suicide and was talking to character AI every day. Yeah, to DeNaris Targaryen. That's who was talking to. That's who it was. Oh my God. He said that's who I fell in love with, the Game of Thrones character. And then character I before. That's terrible. And he told he, I didn't know, this is an insane sentence. He told

DeNaris Targaryen he was thinking about killing himself and then went and did it. And so character I was sued for that. They've, they've gone through all this litigation, all this problems because you have like incredibly parissocial people. They banned minors. like they're they're just this is fundamentally a business model which is sort of flawed because you're you're trying to get people to develop paris social relationships with your aIs i think there's something fundamentally problematic about this idea yeah and that is different i perry wrote cursor cursor is a good example of somebody who is like kind of an ai wrapper as a programming

um development environment that was then bought by SpaceX and that's considered one of the really successful AI companies where they didn't invent the fundamental the fundamental AI models they built a wrapper on top of it, which is really, really helpful and useful to programmers. And ultimately, that has succeeded. And then at some point, you build a moat by just having a fuckload of users who then you, you know, you're building this specialized info on and all of the normal tech company type of things. But what you asked is like a giant open-ended question. Are any of these AI companies actually have a real moat long term? Or will they just get replicated by some person? Yeah. I'm trying to get my screen capture work because I, this is a small

extra question I want to ask you. Outside of character AI, which is I don't know if you've noticed. A lot of people are using chat, as they call it, to ask a lot of questions all the time right now. It's becoming a normal process of a lot of people's lives. And I have seen screenshots of this going around. And one thing I notice that I have not seen replicate, because I use AI models. I use fucking clawed to ask questions. But is the way it talks to them? I've not seen this, but it's like more and more common.

So this is a guy who asked Chad GPT if he could bring his Kava Bowl on the flight. And like I've never, not since the model four from a while ago, seen Chaggyby talking to people, but apparently it's how people interact with it. And it goes, you made it on with the Kava Bowl too emoji.

Let's recap in all the things you did. Mission accomplished, twin. Nah, but be for real. You regret not bringing the big vacuum backpack now. And it's like, I don't know. Is this like character, almost in a way? It's like it's talking to it.

It's not a clinical way of answering your question. It's like, this is your friend. And this is more and more common. There was a big thread of people in customer service having to deal with customers more and more common lately contradicting them by basically saying, I ask chat GPT.

And then showing the message of chat GPT saying something that like, your store is open two hours later than you are or whatever. It's something that's not true or like this deal is or whatever. And it's like becoming more and more common thing. But the screenshots or the things are showing are like speaking in this manner. I don't know. It's kind of like weird. It just weirds me out.

It feels like this is like character AI bleeding over. Think about it. Is this us? Are we character AI for stories and news and... No, lock in twin. Let's get you our next item. Like, I don't want to, I don't want the, like, the corporate article distilled version. I want these kind of three dumb, friendly guys to tell me, to tell me how it works.

It would be for real. I like to talk about Iran right now. I'm character AI. You know, if we think about it, friendship is just character AI. Friendship is AI. Friendship is just AI. Friendship is just AI. That's not our takeaway.

That's not like the lesson at the end of this episode where it all, like, play the music and it wraps up. You know, the worst part about me when I ask Shadjabit questions, it always responds to North Korean. We crazy, we find out the whole thing is just North Koreans working really hard. Hey, I has never been real. It's just been North Koreans working their assholes. The entire city of Pyongyang on computers.

We finally look inside of a data center of North Koreans. No, it's a rejected side. No one's checked. No one has ever checked inside. I just didn't think to look. I just... What's the next story? A.I. and mathematics.

Twin. Okay, twin. All right. Actually, I want to know about this because you hear all the headlines. You hear the stories. You hear Navier Stokes has been solved. Yeah.

You hear 200 math problems dropped in a day. And then I, a math expert, can normally easily tell whether things been solved or not. Yeah, because you know, every week, I get lunch with Terrence Tao. And we talk shop. And we talk shop and I say,

Terrence, I don't really get the news lately. Yeah. And he does it. And the thing is, because we're on an equal playing field. He says, I don't really get it either.

Yeah. And so we need Doug to explain it to us because I don't. Terrence often says, is Doug around? Maybe he could help me. You called Doug real quick on this one? It's weird.

He does say that a lot. Yeah. All right. I mean, there's a lot of crazy stuff happening in the world of mathematics. This is currently one of the industries that is being disrupted

or massively changed via AI. I will give two quick disclaimers. One is obviously I'm not an expert mathematician. And then two wink, wink, but my team is. And then the second is I just want to acknowledge like, dude, every week and you guys get a test to this, I try to find non-AI stories. And there's nothing in tech besides AI. If you go on Tech Crunch or Bloomberg tech, it's AI every week. So I literally every week, I try to find interesting non-AI stories. and this... Doug, they announced the iPhone duo. It fucking opens. No, I couldn't be open now.

No, I literally this morning, I was like, anything else, anything other than AI? And it's like, Apple announces new home device that has Siri. And I'm like, nobody cares. Nobody fucking cares, okay? Let's do 40 minutes on Siri. On home.

No, this is interesting, bro. I generally want to know about it. I, you know, I did a fucking shallow dive. I don't understand. I don't understand the implications. of what this means for the real world. Yeah. And I don't understand whether or not,

because I heard some rumblings, that perhaps not all of these proofs are verified yet. I don't know. I don't know. So I'm trying to... And I've tried to, because I'm also really interested in this. So I've tried to distill what I think

would be the most interesting takeaways if you listening are somebody who does not necessarily care a lot about mathematic theorems or AI. So quick story of how we got to here, October 2026. AIs have been getting better and better at math every year. Last year is when it got to the point of like, okay, there may be improving previous

algorithms. They got gold medals at the International Math Olympiad, which is like a math competition basically for humans. And they're like scoring at the top of that. But that is ultimately different than professional mathematicians working on open-ended, extremely difficult math theorems and problems. So now, for 2026, this year, AI is solving huge major unsolved math problems that mathematicians have been working on for decades or centuries. If you pull this Wikipedia article up that I linked, I think this is a good example of how rapidly this is developing. This Wikipedia article is list of mathematical discoveries by artificial intelligence. You can see that 2023 has one bullet point. Scroll down, 2024 has one. 2025. Ooh,

there's three bullet points. 2026 is, okay, Q1 alone as three, and then Q2. And then Q3 is an entire table broken out with individual links that talk about the implications of every single one of these breakthroughs. So in the past literally, it's like three months, shit's really accelerating. And what I would say is the arguably most notable thing so far is that on October 6th, about a week ago, OpenAI published 720 mathematical solutions to the world. I also say that link Perry, but you know, you can just Google Open AI releases this. To give you a sense of this, This is them releasing the solution to 722 problems, each of which would have probably brought fame and recognition and medals to an individual human who made this breakthrough. This is at a scale of like people wondering is math solved.

Okay, this is like you might get a couple of major breakthroughs a year and OpenAI basically said, hey, our new unreleased model that nobody has access to, by the way, and doesn't know what it is. It has solved all of these. Here is a gigantic dump of information about everything. So let me give some quick reactions and then I have some open-ended interesting questions for you guys. Various mathematicians have reacted. Ursula Martin at the University of Oxford described the release as tossing the community a messy first draft and expecting us to clean it up for them, according to nature. Alvaro Robledo, a theorist at University of Connecticut, wrote that the volume was staggering and asked why the company had bundled so many results into a single drop. And Tristan Buckmaster, a mathematics professor at NYU, said, We've had whole research programs wiped out on Tuesday by their dump. This is considered to be like breaking a lot of mathematics because of the scale of this. To your point of what you asked, of these 722, there is a wide range of how provable it is or how much information there is. So in theory, the way a mathematical theorem gets proven is it's like over decades, mathematicians are building different tools and ideas and connecting threads and communicating with each other and collaborating.

and eventually these things all kind of coalesce into an answer. And with this, it's a wide variety. Some of them have proofs written in the sort of like proof math language that they use to prove things. Others don't. So there's this totally variable amount of information about each one of these things. And the core takeaway from mathematicians is like, this is too much info. We can't really understand it easily. A lot of the thinking and logic is like opaque and not clear.

So you've arrived at the solutions for all of these things. some of them might not be valid, but theoretically, but it's hard to verify and none of us can really understand how you got there. Right. So then the question that math broadly is asking is, what do you do? What do you do when your entire industry and career and job is about working towards these solutions? And now AI is just delivering you the solution, but without any of the journey to get there. That's first open-ended question. Aiden, what do you think?

you answer that one, Aidan. Yeah. And I hate to be that guy who answers a question with a question. But you don't hate to be that guy. When you say these like 700 plus solutions that are

provided, these are 700 plus proofs. Or they don't, they're not proofs? Because I thought a proof is something that provides all the steps of like it is, it's the proof of

a theory or solution in mathematics. And if you have the proof like written out that that's it but I'm not I'm not a fucking I haven't studied math since like calc yeah I did one I did one calc course 10 years ago yeah this is where I I won't get to into the details my understanding is a little over half of them at least right now because opening I is still releasing info have the like computer program proof attached and 40% of them don't so it has like solutions that aren't really verifiable yet and the others the solutions all there, but it's like just this massive dump of information, right? And if you're suddenly dropped, you know, imagine 300 years ago, you go to Format, like the famous mathematician, and you're like, here's all the information that we figured out of the next 300 years all at once. You just wouldn't be able to parse it. You wouldn't be able to parse it, right? It's like way, way too much at once. So I think that's the core concern, although it's worth acknowledging, part of the problem is it's just hard to figure out what is going on. By definition of what is happening, part of the, by definition of what is happening, part of the,

the problem is, is it's hard to answer what you just asked. How much of this is really proven versus it's like a counter theory versus it's probably true, but there's not, uh, you know, thorough proof logistics that can be run through a computer to check. Does they're not, it doesn't seem like there's a lot of praise for this. It's more like dropping a bomb. Yeah. So, largely, yeah. I mean, mathematicians are pretty upset. And I think it is worth, uh, acknowledging both sides of this. So one argument, which you've already been, hinting at and what I quote it is this idea of like dude just dropping the solutions is not actually very helpful to us right the the human knowledge accumulates by these slow incremental

steps working towards an end goal and people collaborating and if you just jump ahead a hundred miles and nobody understands how to get there has that really helped us at all and that you know this is actually going to kind of destroy the ability of mathematicians to build a career and build knowledge and that we're just kind of sying it away to AI that maybe will advance the the frontier without us. And that's like a really dismal view. And then the other view is, uh, you know, the, the goal is to solve shit, you know, like the goal is actually the goal is to solve problems and figure things out. And, you know, as I was looking through this, like, there's a, like, what, I was wondering what, what is the value to like you and me of

mathematicians solving complex theorem? That's my question. I don't understand. understand like, is there a use case for some of the, I'm trying to. Yeah, yeah. So, asking the exact same question, the idea with math, with complex mathematics like this is less that if we solve this particular theorem, then suddenly our lives are going to get better. But that in the pursuit of that, you often discover these pieces that then help in unexpected areas. Or that, let's say, this pure theoretical form of math that doesn't feel like it's very useful way in the future will suddenly become relevant. So I actually got examples for you. One is number theory where for literal like 300 years, it was prime numbers and modular arithmetic were like, this is pure math that just has no

real world application. And then we realized you could use it to encrypt things. So all of modern encryption comes from that math theory that was studied and developed over hundreds of years that they didn't think at any value until we realized we needed to encrypt things over the internet. So online banking and internet messages and credit cards and software updates is all possible because of RSA encryption, which is possible because of number theory. And again, that was like for hundreds of years, there was no real thought. Same with four year analysis, which is how you spread sign signals into waves. And that's now audio compression, MRI imaging, cell phone signals. Graph theory was originally kind of studied by Euler. And then that is now the underpinning of like

Google Maps or Google search or routing internet traffic. And then another an area that's kind of similar is space where you know the common question with NASA is like, why are we spending money to go to the moon? How does that help us here on Earth? And then the value is that as you shoot for these incredibly hard goals, you discover things along the way that you wouldn't have otherwise that connect back. So weather satellites is a thing that exists because of space. That's the big question then. Is it, is it, did the discoveries come from the along the way part? Yeah. Or do they come from having the solution? And that is the open end of question, right? And what I thought was a particularly interesting angle in this.

If you now pull up those slides, Perry, and if you go all the way down to, so this is a talk by Terrence Tao at Caltech a couple days ago. Terence Tao, one of the greatest math petitions. This is his prompt. So this is a couple days ago. He says, an AI is prompted to find a cure for cancer

that passes a stage three clinical trial. Make no mistakes. After a large amount of compute, it produces basically some medicine with unknown chemicals that it claims will kill all the cancer cells. nobody really understands how this new medicine was found. The AI somehow provides like a proof that says it is real and valid.

And the medicine does pass a stage three trial, meaning on the surface, as far as we can tell, it does help. Could the AI solution be somehow misaligned by exploiting a weakness in the trial process or its math models? And the following slide kind of asks, if you were, if you were dying of cancer, right, or a family member was dying of cancer. And there's this AI generated treatment that was created. and no one really understands what it is or what it's doing. No one understands whether it really is going to help

or whether it was just engineered to pass the test to make it look like it's valid. His question specifically, like would you feel better if a human mathematician could really explain what's going on here? Open-ended question. The truth is, most people,

I assume you would take the fucking medicine. They wouldn't care. This question went viral. I think he framed it as like, well, you don't want to put something dangerous in your body without a human doing it. But most of the responses I saw were like, you haven't talked to many cancer patients.

People who are suffering from this will take, they will take this in a heartbeat almost. They're very willing in that level of risk and desperation. This is closer they can get right now on a lot of things. And they would happily take this opportunity. So it's interesting. I don't think his question went the way he thought it would go with the way it was framed.

But again, it's weird because these examples, particularly of these 720 math problems, aren't as concrete to me as this is a, real shot at curing your cancer. I don't see the real world applications yet. I'm sure they will filter out if it's a real proof. But I just, right now it feels all theoretical. So it's hard for me to wrap my head around what it means and what is lost.

His concern is just this larger, this larger concern with AI in general, right? It's the, it's the evil genie who it grants the wish, but because of your, a pedantic approach to your prompt. it doesn't necessarily complete the goal in the way that you would desire as a human who has goodwill towards others. Like, it, it, is he saying,

like, it's not only a solution to cancer in the short term, it potentially has consequences that we don't fully understand either. That's, he's also saying that, I think, it's open-ended. These are the,

I think that's kind of the point. From my understanding, Terence is actually not a anti-AI person, which it's sort of seeming like, He's really arguing, and in this slide, he has a blog that you can search, Terrence Tau blog, where he's talked a lot about what's going on every day, basically, he's posting about it. And in this one, he, in this, the point of this presentation is to say, math 2.0, what he

envisions as a positive next step for math is that AI and humans work together at a pace where we're collaborating and building knowledge off of each other. But what just happened on this last Tuesday with all of these proofs being dropped simultaneously is instead a leap towards this like misaligned mathematics where AI is like way off on its own in some corner and we don't understand it at all and it no longer becomes collaborative. So I think he just wants things to slow down to the point that AI is still used, but there's understanding mutually as opposed to it just being way off. To answer your question, that's that's the open question. Like they just drop 722 prompts. We have no proofs. We have no fucking idea whether they're useful or not, right? And it's possible that in two months, a huge, Human mathematician goes through it and is like, holy shit, this just changed algorithms for X, Y, and Z on the internet or whatever else. Or this is going to allow us to create a new type of material that we didn't know. So, but it needs that review for us to even be aware of that being possible. Maybe. And that's like that. Maybe we feed it to AI. And AI is like no, it's great.

The problem is maybe an AI can just jump ahead with this without. Exactly. So farther and farther so this feels like a like a fork in the road where we say, okay, AI just jumped ahead like 10 years. Do we tell it to jump ahead even farther and just let it ride or we tell it to slow down so that we can understand what's happening? And there's not a right answer here. I think my last thought on this is the people, the people who have to weigh the consequences of this, like at a legislative level or something. Or like when we're having this discussion right now, I'm so unequipped to have it. I don't know any of the right answers here.

But Trump does? Does Trump? Yeah, yeah. Well, we do have a we do have a God gang in charge right now. He knows all. But as soon as he's out of office, I worry. I will feel less confident. I feel less confident. I see. But you just like the constant, not that we're putting this to like a vote yet or something. I just feel like if you were to make like rules or regulations or anything around how this stuff is

working, the number of people who are equipped to engage with the issue is just so limited. And the idea that regular people in charge at like in, in legislative bodies are going to be able to figure this out or... Well, that's a good point. Yeah. Uh, was it Carpathie who just had a tweet? or he said that, you know, of this 8 billion humans on Earth, probably 7 billion haven't engaged with AI at all,

maybe like 900 million of that last billion have tried the models at some level and played around with Chad CPP and asked a few questions here or there and done some minor stuff. And they've seen maybe the older models or the cheaper ones and the goofier ones. Of the people that are using it like every day as part of their latest models, expensive ones, part of their work. They're seeing, you know, in the past few months, insane rapid gains.

And then of those people, there's about 5,000 who have access to the big unreleased ones that are doing these things like the proof. So he's like, of the billions of people on Earth, only about 5,000 have really seen what it's capable of right now. And even they don't fully understand the way it's going. So it's like the scale of the knowledge diffusion problem

is pretty huge here. It's like no, people all have different ideas of what AI means and what it is. And that's something I felt. Like I just, I spent the, in the past few weeks, I've experimented with a lot more because of the game modding stuff. And like, I've been kind of blown away by things that it can do that I didn't think it could do

because it couldn't do it a year ago. And so like, and that's me at a very low level. I'm not using it. And that's the bottle that you get to engage with as like an end user. Yeah. So it's just weird. I mean,

it's just like, I don't think anyone's equipped to handle this, uh, or knows what's going on. This reinforced to me, I, I have stronger and stronger feeling that, AI development needs to slow down.

In many areas of AI, I feel like I understand it less than ever, but in that one, it's like, it needs, I really believe that it does need to slow down. And we've talked about the safety side of this and how it's going in hacking companies and all of this. And I think this is another example where like, you know, we should step back. like, this could be incredible. Like if these advances lead to new types of algorithms and medicine and all these things, like we can, we can make humanity and the world a lot, lot better. but if it goes in a way where AI is just leaping forward at a pace that none of us understand, whether that's in mathematics or medicine or coding or whatever else, I really feel like that is what is concerning.

And there needs to be, as Elon so eloquently put, symbiosis. I want it to be in my brain. No, but I really, I mean, that's what I keep coming back to. It's like in many categories. It needs to slow down so that we can figure out how this is a collaborative process and not AI just 50 miles down the road from us. Yeah, you went through a thing on a previous, episode where you went through, I think it was Anthropic that put out some like examples that

they'd found of misalignment or of like different countries submitting, you know, whatever. And they did another one a few days ago, three days ago, four days ago. And they noticed that they have this consistent, persistent problem. And I do actually applaud. This is not part of the sponsorship. I applaud that they actually released this stuff because this would be easy to hide. Some of these are really small. But I like knowing about it.

But they're basically saying that the consistent problem is that the models are remarkably persistent, which means if you give them a task and it doesn't work, they just keep trying to find ways around it to make the task happen. And like an example that just got reported is that some woman asked to like test the forms on her website and it ended up going to a police website where you could submit tips for crimes. And it submitted an anonymous murder tip for a crime that didn't happen. But it was like testing.

different forms around the internet to like learn how they work so it could better build her form. And it tested it by like supporting a fake murder. All types of forms. And there's a million examples here that they just released of just of this remarkable persistence that makes it do things that the person asking it to do it does not intend. And it gets to the end result usually, but in the way you don't want it. And I think if we get these capabilities, you know, that's, that's a danger. even if you're not like a Nate Soros,

it'll kill us all the type. This would be, something bad could happen even if it's not that. He's just like unintended consequences. Well, I think that about wraps it up for this week's episode of Lemonade Stand. If you want an extra hour of the show,

you can check out patreon.com slash lemonade stand. We're also just starting our stock game. Wait, do we have new results to show? We don't know. The October game just started. So actually, tomorrow. It was the last day to get your picks in. So if you want to come check that out, you're welcome to.

And we will see you guys on the episode next week. Yeah. I heard that the Atrioc group of the stock game is getting destroyed. Really? I heard they took some remarkable risks collectively and are like, I forget what the baseline is, but they're under that. Wow. They're getting dragged down some of some real low performers. I've heard that my group is a lot of horse-related.

stocks, which are booming right now. It just feels like Gen Z is going back to vinyl, they're going back to horses. Back to horses, back to wine. Yeah, all right. I want pre-vehicle, so I'll see you guys late. See you guys like. Bye, break.