TBPN - DOGE Zoomers, The De Minimis Loophole, Supersonic Corridor, A Healthier America

Episode Date: February 4, 2025

TBPN.com is made possible by:Ramp - https://ramp.comEight Sleep - https://eightsleep.com/tbpnWander - https://wander.com/tbpnPublic - https://public.comAdQuick - https://adquick.comBezel - https://getbezel.comPolymarket - https://polymarket.comFollow TBPN: https://TBPN.comhttps://x.com/tbpnhttps://podcasts.apple.com/us/podcast/technology-brothers/id1772360235https://youtube.com/@technologybrotherspod?si=lpk53xTE9WBEcIjV

Welcome to Technology Brothers, the number one live show in tech. We have some breaking news on the cover of the Wall Street Journal, business and finance section. Masa Yoshisone and Sam Altman were pictured in Tokyo on Monday. Masa was holding a crystal ball and dropped it. And Sam Altman is watching him pick it up. And it's so funny because they don't really address it in the context. Yeah, yeah, yeah.

They don't really address it in the context of it. this article, but it's very clear that the Wall Street Journal was like, this is a metaphor of some sort. That is a great omen. That's actually, you need to go. Dropping a crystal ball. I mean, these things can't, people always read into these things. People were like, oh, the cyber truck, he broke the glass. Like these live demos, sometimes they go wrong. But, yeah, Masa, you got to work on that grip strength. The Lone Ranger will send you some grip, grip strength trainers. And maybe you won't drop the crystal ball next time. This article is pretty interesting.

That's actually wild. Yeah, I'll read some of it. In his newly built palace near Tokyo, lined by stone statues of Roman emperors, and surrounded by an 18-hole golf course, Masayoshi Sone was stewing. After declaring for years the imminent arrival of artificial intelligence revolution, this chief executive officer of SoftBank Group had missed out on it. I haven't been able to do anything, he thought, according to a speech he gave to SoftBank investors last year, can I just get old like this?

die. As it turns out, all he needed was a new golden boy. And now Sown, who has his history of who has a history of latching onto charismatic startup founders, has found one in Sam Altman. In what would be the largest yet investment in a startup, Sone is preparing to put as much as $43 billion towards Altman's open AI in a pair of transactions. SoftBank is in talks to invest between $15 billion and $25 billion in the chat GPT maker as part of a blockbuster $40 billion funding round the Wall Street Journal reported Thursday. The round would value OpenAI at up to 300 billion, nearly double its valuation in October,

and under an undeniable sign of Sohn's confidence in its prospects. Jory, what you got? Okay, so my reading into this is this is separate from Stargate. So not only his son levering up, levering up his arm position to go heavy into Stargate, he's also, it's not enough to just own the infrastructure layer. He's got to own the developer tool and the consumer product layer in OpenAI. And honestly, it's absolutely insane to put this into context,

one of the biggest software IPOs ever Snowflake, we've talked about it on the show. They raised in their IPO like $2.5 billion or something like that, you know, somewhere in that range. And so just so Sam is really, raising, you know, such an egregious, if he gets his round done, it's truly, he's definitely the greatest fundraiser of all time, putting Adam Newman to shame, hit the gong, a preeminent gong hit. But honestly, absolutely insane if they pull it off. And the only two people on earth they could are Masa and Sam. And from what I know, this kind of, Sam was really just out they're testing the deep research product.

He posted about this that he was using the deep research product to find a new, I think some type of vintage Accura. And he must have just run into Masa while he was doing that. And they just hit it off. And now they got a potential deal. So I love to see, you know, a little coincidence like that turn into a big funding round. Yeah.

I hope they got out to the track and put those NSXs to work. I'd love to see Masa behind the wheel. That's a good. For a $40 billion round, throw in, you know, why not throw in a vintage, you know, Accura, right? Yeah. It's a rounding error. So. And so in addition, the Japanese tech and investment conglomerate has committed 18 billion towards Stargate.

So 18 billion for Stargate between 15 and 25 for chat GPT for OpenAI. So that's a combined, you know, we're in the 30 billion territory. Fantastic. but they got the money. And remember, the $30 billion number was Masa's initial target for Vision Fund. He went, he was going to Saudi Arabia to meet with MBS. And on the plane, he had written out that he wanted $30 billion and he circled that. And then he crossed it out and wrote 100 because he liked, he liked the idea of saying,

give me $100 billion. I'll give you a trillion. And so Masa continues to lever up to go bigger than ever. and, you know, we'll see how it pans out. It's way too early to have any judgment. But we just, we like any opportunity to hit the gong. So thank you to Mosa and Sam for providing that. For sure.

Yeah, it's unclear, you know, the metaphor of dropping the crystal ball, that clearly happened because he didn't go turbo-long AI three years ago. But maybe he's picking the crystal ball back up and he's putting the money. in the right place. I just, I just, what was Mosa doing in every, they had a little bit of one of the earlier rounds, right?

This is not their first time investing in the company. So Mossa's getting the updates on every, the first of every month. And he's like, I got to, I got to own a lot more of this bad boy. He's like, I don't care if the model layer is getting commoditized.

I just, I love Sam. He's my guy. I'm going turbo, turbo long. Yeah. Well, hopefully it works.

out. We always love seeing these big deals get done. And overall, I'm excited to see a lot of capital flow into an American AI company. I think this is exactly what we need right now. So, yeah, bad day to be, bad day to be deep seek. Bad day to be deep seek. I'm glad the money's not going there. So close to Japan yet so far away. Truly, truly. Well, let's go to Luke Ferretor. We broke the news yesterday that he is over at Doge. He has read access to the U.S. Treasury and payments flow and has been doing a whole bunch of data science and machine learning to understand where money is going in the government. He's part of this crack team of zoomers that Elon has hired and staffed on a project to understand how the government works, where money is going. It's been

extremely controversial. A lot of people came out in support of Luke. We did and the rest of the team. And a lot of people were very upset. And so it was very controversial. But we want to give a deep dive on Luke and his previous work on the Herculaneum scrolls today. And so we will go through that in just a minute. Jordy, which got? One quick note, I think it's worth mentioning for the audience that hasn't been following Doge maybe as closely as some other people.

But it's important to understand that Luke is actually not an employee of the government. From what we know, he's an employee of Doge, which you can think of as operating as this sort of personal vehicle for Elon and Elon's team and this sort of initiative to try to have an impact on the government. So think about it more as like a, almost like a consulting shop, think tank. Yeah, think tank. So that's a right way to look at Luke, but we'll kind of get into where he got started. Yeah. And so let's go through some of these posts. Of course, he famously won a portion of a

$700,000 prize for figuring out how to read ancient scrolls that were carbonized by Mount Vesuvius, which we'll go through today. And people are having a lot of fun with this. Armand Tully says the doge zoomers preparing to fire half of the federal government. And it's the bobs from office space, but with zoomer haircuts, of course. And so people are having a lot of fun with this. You know, everyone, it's become a massive political football, even though I think Luke is pretty much just trying to understand where the money is actually going. which is a very reasonable thing for either party to want to do.

I like that everybody had to clarify that he only has read access. It's not like, you know, he's very, you know, he's not just in there, you know, just like he's using it like Venmo or cash app, just whipping money around. He's just trying to get it. He's just trying to understand the scrolls that are the U.S. Treasury Department's database and incredible paper trail. It really is like the scrolls project all over. It's so funny the similarities.

And so Jarvis writes, like five people who understand how computers work show up in D.C. And within two weeks, restructure the entire system that's been in place since World War II. That's wild. I love this post. It's so funny. But it is a testament to AI and data analysis that you can actually do this extremely high leverage work and comb through all these payments. Somebody else was saying that one of the early Doge projects was just get a dump of payments,

just get a whole bunch of like just data on where things are going and just look for the crazy stuff that pops out that seems like really weird. And if you look through the, the PPP loans, there were a lot of those that went to reasonable restaurants that were closed down by mandate because of COVID made sense that they would that they would need PPP to bridge. But then there were some people that set up LLC is called like Ford Raptor LLC. And we know what that guy was doing. And so, wait, you don't need to be a genius. It's not a left-right issue. Like, you're leaving through that.

You're like, why are my tax dollars going to buy someone a truck? Like, this should not be happening. And so hopefully, you know, the Doge team can find a bunch of stuff like that. Yeah, and part of it is understanding, okay, let's understand where the money went, and then let's understand how it's being spent. And yesterday we highlighted that out of $700-ish million that was going towards just one university in California from the federal government, So this is U.S. federal taxpayer dollars going to a California university.

Yep. We both live in California. I'm a product of the UC system. I'm nothing against the UC system, but they were spending 40% of the budget, roughly $300 million a year purely on administrative costs. So it's not, you know, the first step is understand, let's understand where the money's going. Yep.

Second step is let's understand how it's being spent. But there's a lot of work to do just on the first part, because this is one of the most complicated systems that I can even think of, right? Yeah. And there's just been generations of Croft built up. Well, let's move on to Ashley Vance's article, which touches on the Herculaneum Scrolls project. He wrote this back in over a year ago, actually a year ago tomorrow or maybe today. February 5th, 2024, Ashley Vance and Ellen Hewitt over at Bloomberg, right? A few years ago during one of California's steadily worsening wildfire seasons, whoof, that hits hard. Nat Friedman's family home burned down.

A few months after that, Friedman was in COVID-19 lockdown in the Bay Area, both freaked out and bored. Like many a middle-aged dad, he turned for healing and guidance to ancient Rome. While some of us were watching Tiger King and playing with our kids' Legos, he read books about the empire and helped his daughter make paper models of Roman villas. Instead of sourdough, he learned to bake Pannas Quadratus, a Roman loaf pictured in some of the frescoes found in Pompeii. During sleepless pandemic nights, he spent hours trawling the internet for more Rome stuff. That's how he arrived on the Herculanian papyri, a fork in the road that led him toward further obsession. He recalls exclaiming, how the hell is no one ever told me about this?

The Herculaneum papyri are a collection of scrolls whose status among classicists approaches the mythical. The scrolls were buried inside an Italian countryside villa by the same volcanic eruption in 79 AD that froze Pompey, in time. To date, only about 800 have been recovered from the small portion of the villa that's been excavated, but it's thought that the villa, which historians believe belong to Julius Caesar's prosperous father-in-law, had a huge library that could contain thousands or even tens of thousands more. Sorry to interrupt. We have to really bring back the word prosperous into today's vernacular. It's a fantastic word, and everybody should aim to be prosperous in their own way. Absolutely. So he says such a hall would represent the largest collection of ancient texts

ever discovered and the conventional wisdom among scholars is that it would multiply our supply of ancient Greek and Roman poetry plays and philosophy by many fold. So we're talking about like a five or 10x increase in the amount of text from Greek and Roman literature. High on their wish list are the works by the likes of Sappho and Sophocles, but some say it's easy to imagine fresh revelations about the earliest years of Christianity. So it's been a massive goal of, you know, all these ancient scholars to unearth these and we'll go through how they did it. Jordy, you got anything for me? Now that's it. That's it. Okay. So some of these texts would completely rewrite the history of key periods of the ancient worlds says Robert Fowler, a classicist

and the chair of the Herculaneum Society, a charity that tries to raise awareness of the scrolls and the villa site. I love that he's just out there being like, I got to spread the word of these scrolls, I got to get the word out. Someone is going to figure this out and he landed on it. He found that and that put together crack team and actually figured it out basically. I mean, it's not 100% down. And that's not, for anybody that's not super familiar, he was, was he, he was a co-founder of GitHub or he just joined his CEO. He was CEO, yeah. He was so he's CEO of GitHub, a tool that literally every developer that I know uses every single day. Yep. And he has a thing for aggressively pursuing passion projects and side projects with no, not necessarily

any sort of direct monetary incentive, more so using his wealth to try to uncover knowledge. And so the scroll is one project that that's really exciting. And the one that he went pretty viral for recently was called plasticlist.org, which is similar to what another founder named Cormack is working on, a friend of mine. And what Plastic List did is he went and he spent probably roughly half a million dollars, so similar to the to this sort of Scrolls project and tested hundreds of different popular food products because he realized that this one, he had realized this one protein shake that he was drinking had just like egregious levels of microplastics.

And he was sort of traumatized by that. So he's saying, all right, let's test a bunch of different products. And so he's doing exactly what we want other other beaners. to be doing. So not just building companies and investing, but saying, hey, let me, let me spend half a million bucks to sort of unearth some, you know, important knowledge for humanity and, and pull together some great people in the process. So, yeah, I'm already excited for, for Nat's, like, next project because he's keep going. Yeah, 100%. So we'll pull up a picture of the scrolls. You can see these. And, John, I actually can't see this on my screen, by the way. Oh, really?

Yeah. You should be able to see the screen. I'm sharing the Bloomberg Business Week. This is an ancient Roman scroll. I see Luke Scrolls. So I think you might be in the wrong tab. Huh. Oh, did it never?

Let me try and share this again with you. Let's go share. Let's go here. Is that good, Ben? Can you see that? Bloomberg Business Week. This is an ancient Roman scroll flash fired by a volcano. and we got pictures of Nat Friedman here.

And so you can see from this scroll, people have spent hundreds of years trying to unroll them, sometimes carefully, sometimes not, and the scrolls are brittle. Even the most meticulous attempts at unrolling them have tended to end badly with them crumbling into ashy pieces. In recent years, efforts have been made to create high-resolution 3D scans of the scroll's interiors,

the idea being to unspool them virtually. This work, though, has often been more tantalizing than revelatory. Scholars have been able to glimpse only snippets of the scrolls, innards, and hints of ink on the papyrus. Some experts have sworn they could see letters in the scans, but consensus proved elusive. And scanning the entire cache is logistically difficult and prohibitively expensive for all but the deepest pocketed patrons. Anything on the order of words or paragraphs has long remained a mystery. So can you imagine you've discovered this ancient Roman scroll and you're just absolutely thrilled. You're like, okay, I'm going to start, I'm going to start unrolling it and just disintegrates like immediately.

Every single time. And they've tried so many different like bring in the, yeah. And Mike Salana posted something earlier today that was that was pretty interesting. He basically said blogs from the early 2010s are actually less recoverable in many ways in these ancient, you know, multi-thousan-year-old scrolls. which is a whole other issue of like how do you properly archive the internet and all the content we put out so that it can be you know preserved for uh you know people you know 5,000 years from today yeah but anyways looking a couple chads here just a couple roman empire enthusiasts yeah we love to see it he clearly like became actually actual buddies with the professor on the project and they're just like hanging out now um and so um Friedman isn't your average Rome-loving dad.

He was CEO of GitHub. And most importantly within GitHub, he developed GitHub co-pilot, which is their AI coding assistant. And he's been AI-pilled for years and years and years because this was one of the first LLM implementations that actually got traction. And I remember back in, I think it was built on GPT3. And back in 20, I guess it was like 2020, 2019, 2020.

He was saying like every other. model like like gpt three at the time people would kind of use it as a search engine or to like you know do copy editing and stuff but it wasn't quite dialed there GitHub copilot was the one product that didn't have super high churn obviously that's changed now with things like perplexity and chat gpt and a lot of these apps but GitHub co-pilot was the first canary in the coal mine for these LLMs these AI systems are robust enough to be used by people for for a long time yeah now now I'm sure people, a lot of people just know him as an investor because he's now deploying a ton of capital and his ability to have an idea and then pull together super talented people and just ship it.

He did this with another Nat Grossman. So there's Nat Friedman and that Grossman. It's Daniel Gross, right? Daniel Gross, sorry, Daniel, Dan, Dan Gross. And they shipped an entire cluster just for their portfolio companies because they have another program called the AI grant, which is like a similar to YC type of structure. So I love that just like having an idea, sharing it publicly, and then just like quickly actually acting on it, just not letting it stay in the timeline is great to see. Yeah, and so he launches the Vesuvius challenge,

offering a million dollars in prizes to people who could develop AI software capable of reading four passages from a single scroll. Maybe there was obvious stuff no one had tried, he recalls, thinking. My life has validated this notion again and again. So it's this VC investor mindset applied to like ancient history research, I guess. And so as the months ticked by, it became clear that Friedman's hunch was a good one. Contestants from around the world, many of them 20-somethings with computer science backgrounds, developed new techniques for taking the 3D scans and flattening them into more readable sheets.

Some appeared to find letters, then words. They swapped messages about their work in a Discord chat. And often much older classicists sometimes looked on in hopeful awe and sometimes slagged off. and sometimes slagged off the amateur historians. So there's a little bit spicy. Oh, these AI kids are going to be able to make an impact here? They don't really know the true stuff. I love that there's like these little subplots and battles going on.

You're not a real historian unless you've had gloves on and you're sort of trying to piece apart, you know, disintegrating this ancient scroll. This is very disrespectful to the guys who just go Indiana Jones mode and get in there. I post it off snake. You've got to be the only edge, the AI models have eaten the entire internet.

So the only edge now is to go Indiana Jones mode and go looking for rare esoteric knowledge and, you know, discontinued books, right? Yep. But they really did here. And so you can see a picture here. Before Mount Musuvius erupted, the town of Herculaneum sat at the edge of the Gulf of Naples,

the sort of gateway, wealthy Romans used to relax and think. Unlike Pompeii, which took a direct hit from the Vesuvian lava flow. Herculaneum was buried gradually by waves of ash, pumice, and gases. Although the process was anything but gentle, most inhabitants had time to escape, and much of the town was left intact under the hardening igneous rock. Farmers first rediscovered the town in the 18th century when some well diggers found marble statues in the ground. In 1750, one of them collided with the marble floor of the villa, thought to belong to Caesar's father-in-law. Senator Lucius Culpernius, Piso Cessonius, also known today by historians as Piso. During this time,

the first excavators who dug tunnels into the villa to map it were mostly after more valuable artifacts like the statues, paintings, and recognizable household objects. Initially, people who ran across the scrolls, some of which were scattered across the colorful floor mosaics, thought they were just logs and threw them on a fire. No. Terrible. Eventually, though, somebody noticed the logs were often found in what appeared to be libraries or reading rooms, realized they were burnt papyrus. Anyone who tried to open one, however, found it crumbling in their hands. Crazy.

Should I go on? Let's see. Today, the villa remains mostly buried, unexcavated, off limits. And so there's levels to this project where they get a few scrolls. They scan them. It's very expensive. and then that and then the more that they can unearth, the more they can decipher, the more they can invest in more digging and more scanning because each one of those is very expensive. I think even just

just taking the scroll to the CT scan, I think they flew it on like a private jet and like needed to like really like make sure they couldn't check this. You can't check this on United. It's going to be a disaster. It's going to fall apart. Yeah, at least you throw an air tag in it. If you're going to. Yeah. So we have about 800 scrolls from the villa today, but there could be thousands or tens of thousands more. For the past 20 years, Brent Seals, a computer science professor at the University of Kentucky, has used advanced medical imaging technology designed for CT scans and ultrasounds to analyze unreadable old texts. No one else was working on it, and no one else thought it was even possible.

Progress was slow, sales belt software that could theoretically take the scans of a coiled scroll and unroll it virtually, but wasn't prepared to handle a real Herculaneum scroll when he put it to the test in 2009. The complexity of what we saw broke all of my software. The layers inside the scroll were not uniform. They were tangled and mashed together, and my software could not follow them reliably. And of course, 2009, this is before the Deep Learning Revolution. There's a lot like the, I'm sure this was more of a programmatic approach, like more algorithmic approach to like mapping the geometry and trying to unroll it. a terrible job for like hand coding, perfect for deep learning.

And so by 2016, he and his students had managed to read one scroll, a charred ancient Hebrew text by programming their specialized software to detect changes in density between the burnt manuscript and the burnt ink layered onto it. The software made the letters light up against a darker background. Sales team had high hopes to apply this technique to the Herculane and papyri, but those were written in a different carbon-based ink. and their imaging gear couldn't illuminate it in the same way. So massive setback in 2016.

It is crazy how long this story's been going on. Yeah. And relevant too. I'm sure Luke is already dealing with this. You know, he had been posting last week. Hey, does anybody have any way to process huge volumes of, you know, PDFs through any type of programmatic way? It's going to figure out how to do it with one document format.

Then he's going to get, you know, a trillion new documents in another format and be having it do it in the same way. So he's prepared. Yeah. And this next paragraph is the biggest indictment of current Silicon Valley, rich guy, billionaire VC culture. So Nat Friedman finds this guy sales who's working on these scrolls. And he had set up Google alerts for him. He's, he's, Nat Friedman's just getting obsessed with Rome. A year past, there's no news. He starts watching YouTube videos really understands that this guy sales needs money. So in 2022, Friedman was convinced he could help. He invited Seals out to California for an event where Silicon Valley types get together and

share big ideas. This is like the perfect thing to send this guy to. Sales gave a short presentation on the scrolls to the group, but no one bit. And Nat Friedman says, I felt very, very guilty about this and embarrassed because he'd come out to California and California had failed him. sure what happened was like, oh yeah, we just wanted to hear about this crazy stuff, but we really want to put our money in like B2B SaaS. Like I could possibly throw a 20 in the in the tri-corner hat for this guy. It's so funny because you know many of the people that were pitched there on this idea would go on to donate millions to VCs for Kamala. Yeah. And meanwhile, that goes, you know, goes long on the scrolls and becomes literally a myth himself. Totally. Or,

uncovering this kind of thing thousands of, you know, it's a legendary move. Yep. And so Friedman proposed the idea of a contest. He said he put up some of his own money to fund it and his investing partner Daniel Gross offered to match it. Seale says he was mindful of the tradeoff. The Herculane and papyri had turned into his life's work and he wanted to be the one to decode them. More than a few of his students had also poured time and energy into the project and planned to publish papers about their efforts. So now suddenly a couple of rich

guys from Silicon Valley were barging into their territory and suggesting that internet randos could deliver the breakthroughs that had eluded the experts more than glory though seals just really hope the scrolls would be read and he agreed to hear freedman out and help design the AI contest they kicked off the vesuvius challenge on the aides of march treatment announced the contest on the platform we fondly remember his twitter and many of his tech friends agreed to pledge their money toward the effort while a cohort of budding papyrologists began to dig in to the town at hand. Incredible.

So they lobby the Italian and British collectors for years to scan the first scroll. Suddenly the Italians were now offering up two new scrolls for scanning to provide more AI training data. With Freedman's backing, a team set to work building precision-fitting 3D printed cases to protect new scrolls on their private jet flight, I was right, from Italy to a particle accelerator in England where they scanned for three days straight at a cost. of about 70K. That's a lot of money, but it's such a good investment. It's so interesting. And yeah, the whole thing is like- Way less than the cost of flying private from L.A. to London,

though. So- Yeah. And so- Good bargain. Seeing the imaging process and action drives home both the magic and difficulty inherent in this quest. One of the scroll remnants placed in the scanner, for example, wasn't much bigger than a fat finger. It was peppered by high-energy x-rays, much like a human going through a CT-scan. except the resulting images were delivered in extremely high resolution. For the real nerds, about eight micrometers. These images were virtually carved into a mass of tiny slices, too numerous for a person to count,

along with each slice. The scanner picked up infinitesimal changes in density and thickness. Software was then used to unroll and flatten out the slices, and the resulting images looked recognizably like sheets of papyrus, the writing on them hidden. And so these, like the files from, the output of this scan are massive, so it's a perfect case for data analysis and AI.

An algorithm that can detect tiny amounts of ink on each little piece of scroll fragment can then combine that data into a unified legible simulation of how the scroll might have appeared back in 79 AD. It is crazy. It's 2,000 years old. That's so, so old. Like even finding a book from 200 years old, that's 200 years old is difficult to deal with.

And like the paper might be dissolving. imagine 2000. Yeah. Hard enough to find an original copy of Mary Meeker's internet report. Yeah, it is. Yeah, we got to get that on you. Original hand printed by Mary Meeker, staple. You know, that's priceless, right?

Yeah, yeah. There's only a few copies left. It's our scrolls. Yeah, yeah. Those are the real dark scrolls. Yeah, yeah, the drafts that didn't make it to the printer. And so this is where Luke Ferretor enters the story. Luke Ferretor was hooked at from the start Ferretor, a bouncy, 22-year-old Nebraska

who often exclaims, oh my goodness, a herd Friedman described the contest on a podcast in March. I think there's a 50% chance that someone will encounter this opportunity, get the data, and get nerd sniped by it, and we'll solve it this year, Friedman said on the show. Ferretor thought, that could be me. I love that. Yeah, he's, this is a guy who's looking to get nerd sniped by, opportunities, right? Yes, yes.

You're not on the level unless you're actively looking to be nerds night by opportunities. Yeah. And so there's a, there's a photo of Luke here with his, with his hand-built piece. He's got some graphics cards in there. He's grown up a lot in the last year. Yeah, it's looking great, looking great. And so, and this is an interesting twist. Do you know Casey Hanmer?

I've actually met this guy. He has a, he has a castle in Los Angeles that he uses for his company, Terraform Labs. Very cool company. They're converting, they're trying to build, I think it's natural gas conversion from the air using solar. So they're going to put out a ton of solar panels and then basically create, like suck carbon out of the atmosphere, use hydrogen and get natural gas for energy. So you can use the solar panels to basically generate gas that then you can use and it's all carbon neutral. So very cool. And he has a ton of really interesting blog posts about his thoughts on global.

and energy and stuff and where we go from here. So then Casey Hanmer, an Australian mathematician, physicist, and polymath, scored a point for humankind by beating the computers to the first major breakthrough. Hanmer took a few stabs at writing scroll reading code, but he soon concluded he might have better luck if he just stared at the images for a really long time. This is so funny to me. Eventually, he began to notice what he and other contestants have come to call Crackle, a faint pattern of cracks and lines on the page that resemble what you might see in the mud of a dried-out lake bed.

And there's an image here of all the letters deciphered, and the crackle became a big thing in the discord. To hammer his eyes, the crackle seemed to have the shape of Greek letters and the blobs and strokes that accompany handwritten ink. He believes it to be dried-out ink that's lifted up from the surface of the page. The crackle discovery led Hanmer to try identifying clips of letters in one scroll image. In the spirit of the contest, he posted his findings to the Discord. In June, at the time, Ferretor was a summer. I love, by the way, I love how this is a contest, but everybody's working together. It reminds me of the, the Russ Ulbric case where, yeah, everybody wants to win, you know,

everybody wants to, like, pin Ross and, like, be the guy. Yeah. But at the same time, it's as much as it's a contest, you know, especially in the case with this girl, everybody's sort of sharing their discoveries along the way. And it's not, it's not a, they're not seeing it as a zero-sum game.

They're playing for humanity, right? And also the challenge. You see, I actually disagree with you because in the, in the, in the Ross Ulbert case, all of the different groups, they weren't working together well at all.

No, they weren't working together well. And doing fraud. I'm not saying they were working together well, but there was some collaboration. Totally. In this case, you could imagine a world with the scrolls

where nobody wants to share anything in the Discord because I want the 700K. Totally, totally. And there's also this interesting thing where in the Ross Albert case, everyone thought it was going to be this crazy hacking data analysis that would win the day. But just going on Google

and looking back at old posts was like a very key part of the story. That was key, yeah. And it's the same thing where Casey Hanmer just looking at it and staring at it for a long time actually helped break the case wide open, which is awesome. So Ferretore was a summer intern at SpaceX, he was in the break room sipping a Diet Coke when he saw the first post, and his initial

disbelief didn't last long. Over the next month, he began hunting for crackle in the other images. One letter here, another couple there. Most of the letters were invisible to the human eye, but 1% or 2% had the crackle. Armed with those letters, he trained a model to recognize hidden ink, revealing a few more letters. Then Ferretor added those letters to the model's training data and ran it again and again and again. The model starts with something that only a human eye can see, the crackle pattern, then learns to see ink that we can't. Unlike today's large language models, which gobble up data, Ferretor's model was able to get by with crumbs. For each 64 by 64 pixel square of the image, it was merely asking, is there ink here or not? And it helped

that the output was known. Greek letters squared along the right angles of the crosshatched papyrus fibers. So it's like the perfect case for using AI. And fortunately, because it's not this massive large language thing, he can run it on his computer that he built by hand. He just needs, he doesn't need a graphics card and he does need to write some somewhat complex code, but he doesn't need a massive data center. You don't need to build Stargate for this. Which is great. It's a really cool, cool opportunity. So in August, Ferretor received an opportunity to put his software to the test. He'd returned to Nebraska to finish out summer. and found himself at a house party with friends when a new crackle-rich image popped up in the contest discord channel.

As people around him danced and drank, Ferritor hopped on his phone, connected remotely to his dorm computer, threw the image into his machine learning system, then put his phone away. An hour later, I drive all my drunk friends home, and then I'm walking out of the parking garage, and I take out my phone, not expecting to see anything, he says, but when I opened it up, there's three Greek letters on the screen. I love this anecdote. It's so good. amazing. Amazing.

Somebody was posting that this is, this is the counterpoint to Vivek saying that, you know, Americans, you know, just are jocks and, you know, don't have the work ethic. He's locking in remotely from his phone to try to inch forward and create more progress, you know, from Nebraska, right, solving this incredible, almost impossible challenge. So it's just such an amazing story. I'm not going to pretend I understand the same. by the bell references, but he's clearly the good character in that show. Whatever that means.

So around 2 a.m. Ferretor texts his mom and then Friedman and the other contestants about what he'd found, fighting back tears of joy. That was the moment where I was like, oh my goodness, this is actually going to work. We are going to read the scrolls. Soon enough, Ferretore had found 10 letters and won $40,000 for one of the contest's progress prizes. The classicist reviewed his work and said he'd found the Greek word for purple. Isn't that amazing?

That's so amazing. So one, we got to start using, oh, my goodness, because it's a great, it's a great phrase and a great substitute for an alternative. But I just wanted to put it out there for our listeners. This is an example of a great side hustle. So next time you're thinking to yourself, you know, I like my PM job. I make a couple hundred thousand dollars a month at Meta, but I'd like to do something on the side. why not just sort of try to crack the code of some ancient scrolls to make a few an extra

10 you know 10,000 dollars on the side you know it's a good it's intellectually stimulating and just a great sort of not it's not passive income but it's a nice you know second income stream yeah I mean there there are a lot of stories like this I'm pretty sure great George Hatz the the hacker who broke the PS3 the original iPhone uh jail broke the iPhone and eventually went on to develop comma AI and a bunch of other cool stuff. But fantastic hacker, early in his career, he would do hacking competitions and bug bounties. So he would go to Facebook and say, I hacked your entire system. Like, I'm a white hat hacker. I'm not going to do anything bad, but can you give me a reward? And they'd say, absolutely. Like, you just saved us billions of dollars. Here's 100K or something like that.

And then was he part of the reason that a lot of these companies ended up rolling out more formal programs around this because probably generally yes part of a try yeah and then i think there was even there were even startups created that that were just marketplaces for bounties yeah uh at some point or another i don't know if any of them really broadly caught on but um it's cool yeah cool strategy and i think at one point he went a car because of it or like i i think he he hacked the iPhone and uh uh steve wasniak sent him an email being like this is awesome like you're your your you're a legend hacker. And then Steve Jobs was like, this is unacceptable. This destroys my business or something like that. I'm not exactly sure the story, but it was very funny that he was just like, you know,

as a kid, he was very young and he was having an impact and making money. Love it to see it. So Ferretor continued to train his machine learning model on crackle data and post his progress on Discord and Twitter. The discoveries he and Hamer made also set off a new wave of enthusiasm among contestants and some began to employ similar techniques in the latter part of 23, Ferritour formed an alliance with two other contestants, Yusuf Nader and Julian Schillinger, in which they agreed to combine their technology and share any prize money. I love this.

This is the crew that they're building. Love it. And so in the end, 18 teams submit entries for the grand prize. Some submissions were ho-hum, but a handful showed that Friedman's gamble had paid off. The scroll images that were once ambiguous blobs now had entire paragraphs of letters lighting up across them. The AI systems had brought the past to life. It's a situation that you practically never encounter as a classicist. You mostly look at texts that have been looked at someone by someone before. The idea that you are reading a text that was last unrolled on someone's desk

1900 years ago is unbelievable and you are on mute. A group of classicists reviewed all the entries and did in fact deem ferretor's team the winners so he actually won they were able to stitch together more than a dozen columns of text while entire paragraphs all uh over their and their entry uh still translating luke got the email telling him that he won and he just goes oh my goodness oh my goodness i'm i'm fully taking that term and using that constantly it's so good oh my goodness Oh my goodness. It's so nice. Still translating the scholars believe the text to be another work by Philodemus, one centered on the pleasures of music and food and their effects on the senses. Peering at and beginning to transcribe the first reasonably legible scans of this brand new ancient book was an extraordinarily emotional experience. I love that. They're just like so stoked. It's great. Yeah. It beats slightly changing the color of a button. to a different shade of blue and getting, you know, 1% increasing conversion.

You know, it's close, but it's definitely slightly more emotional. Everyone involved in this story is just so earnest, and that's what I absolutely love about this. There's just not an ounce of cynicism here. And it's just beautiful. And that's why I'm so optimistic about his work at Doge. I think he's just genuinely. Yeah, and that's why the people that are actually sending ridiculous messages, hating on him, trying to dunk on him, making him turn his account private.

It's like if they just took literally five minutes to try to understand his story, it's hard to say, how do you position this guy's a bad person, right? Yeah. Well, I mean, a lot of those people that are, that are, you know, like, you know, aggressively anti-Luc, they are not reading this coverage. They're reading some clickbait, sensational, sensational viral article about how, you know, He has right access to the treasury and he's going to send money all. And he's just going to steal all the money for Elon's thing.

You saw there was this, I forget which politician he represents somewhere between San Antonio and Austin, Texas. So Texas-based U.S. representative. He was posting, we need to figure out how to fire Elon Musk. like he's we need to use every bit of political power and pressure and every law to get this man fired and it's just so funny because he's not an employee of the government you can't fire him and he's working under the direction of the white house and the chief you know the chief executive officer of the country right now right and so um and and luke's in the luke's in the same boat so good luck good luck firing luke as a point of clarification it came out yesterday Elon is a special

of the government. He does have a dot-gov email address. I think the the exact structure of Doge is evolving. And I think we'll go through a couple different maybe maybe congressional or legal challenges, but but it is is evolving. But yeah, I mean, obviously there's a lot of people that are anti. A lot of people are pro. Even on the left and the right, even there's even some people on the right that are like, I don't like that the tech people are in charge. So it's developing. But, But we're here to talk about Luke, the brother of the week. And so let's close out with the goals going forward. Barring a mass relocation, Friedman is working to refine what he's got.

There's plenty left to do. The first contest yielded about 5% of one scroll. A new set of contestants, he says, might be able to reach 85%. He also wants to fund the creation of more automated systems. They can speed the process of scanning and digital smoothing so that you don't have to 3D print the case, bring it over here. like bring the CT scan there, dig out better. You need a lot of different people here.

It's not just about the AI. The AI is just one piece in the value chain here. He's now one of the few living souls who's roamed the villa tunnels. And he says he's also contemplating buying scanners that can be placed right at the villa and used in parallel to scan tons of scrolls per day. Even if there's just one dialogue of Aristotle or a beautiful lost Homeric poem or a dispatch from a Roman general about this guy, this Jesus Christ guy who's roaming around,

he says, all you need is one of those for the whole thing to be more than worth it. What a fantastic story. What a fantastic story. Amazing. I love it. I'm so excited to see where the Scrolls project goes. Yeah, the only bummer, small bummer,

is that Luke was kind of somewhat of, you know, he's infamous already, but he was a bit more of a secret. it. And now when he does raise his pre-seed round, it will be 100x ever subscribed. And, you know, we'll all be lucky if we get a thousand dollars slug. But, uh, just send the wire now. But it's good. Yeah, yeah. Here's a check. Cash it when you, when, when, when you incorporate at any terms. Yeah. At any terms. Yeah, we need it. We need a pre-incorporation safe that

that just says yeah. This goes into your first, your first round. Yeah. If you start. a company at any point. Yeah, yeah, exactly. Well, let's let's go to some, uh, some fantastic analysis from Ben Thompson over a shatnery. He's talking about the day minimus loophole and he opens with a quote from Bloomberg. The US president Donald Trump's tariffs on China and other countries are rocking markets and casting a pall over Chinese firms seeking to go public. Take Sheen, the fast fashion sensation, the online retailer founded in China, but headquartered in Singapore, is said to be working on what could be one of the biggest listings in London in years, potentially valuing the company at $61 billion U.S. dollars as soon as this year.

Trump's executive orders specify that the de minimis exemption for packages worth less than $800, a loophole used by the likes of Sheen for years will no longer apply. Though Sheen has been diversifying its shipping bases ahead of Trump's trade policies, the full scope of the de minimis changes is still unclear. It's unlikely that the company would be immune to the change. changes. Sheen didn't respond to a request for comment. Yeah, John, can you can you break down for the audience? So sheen manufactures their products in China and then instead of sending a single

shipping container that would have a value far higher than than $800, they send the individual packages. Yep. From China direct to consumers. Yep. And that's what allows them to get around that that that that's the loophole, right? Yes, that is the de minimis loophole. And the reason that the day minimus loophole exists, it was not, this is not something that Sheen and Timu lobbied for. And then, oh, they got lucky and they knew that this was going to be a thing. No, the day minimis loophole makes a ton of sense.

Because if you're some Etsy seller and you want to send someone a $50 package, a t-shirt from Italy to America, it makes sense that having to do all the different customs and tariffs and all the different oversight on that small package is, cumbersome. It actually imposes a massive cost on the government and the ports, and there's all these extra filings. And so the government kind of wisely at the time said, hey, let's just focus on the big stuff that's coming in, the massive shipments of pallets of TVs or the cars that are coming in. Let's tax those. And you know what? At the time, I'm sure they looked at it and they were like, you know, if we draw the bar at $800, that's less than 1% of trade volume. And so sure, we'll lose, we'll lose, if we're taxing something at 10%, we'll lose 10% of 1% of our trade, you know, revenue doesn't matter, not a big deal. But then, once this loophole started to exist

and became popularized, people figured out how to take advantage of it. And the answer was, let's ship directly from the factory, an individual package directly to the American consumer. It takes a long time. And so when you go on Timu, I actually went on Timu, and it is the funniest e-commerce flow you've ever seen. You go to Google images and you search for something. You land on the Timo website. You're on mute. And you land on the T-Mu like, you know, web view.

It really wants you to install the app. But it does one of those spinners, you know, the spinners to be like, you know, what's your percent off? And I'm like, okay, like, I guess I have to click this to get to the product or whatever. And it spins. and you always win 100% off. They just figured, they realized that like giving someone 10% off isn't as good.

So just tell them they got 100% off. And then it pops up another spinner that's like, now you're eligible for a multiplier. And it's like 2x, 3x, 6x. You always win the most, the highest like multiple. So then all of a sudden it's like, you have won $600 in free stuff, like install the app. And then of course, once you go down that funnel,

it's like, oh, well, like you got to pay shipping. and all this other stuff and the shipping is the only thing that costs money. So, you know, is this whole like kind of scammy, like ultra optimized thing. But I was just cracking up being like, oh, really, Timo? Like, you're just going to give me all this stuff for free? Like, it's such a scam. Well, you remember when we launched our trailer for the podcast. Yeah.

At the end of last year, somebody, one of the negative comments that I saw was, oh, tech brothers do Timo's succession. and I was like, yes, this is T-Mu succession. We filmed this video in basically 12 hours with a $2,000 budget. You nailed it. But it's interesting how T-Mu became such a... Timu exploded. And I don't know if Ben has the data here,

but it did become popular and it did make it really difficult. If you made custom T-shirts at a factory in L.A., and you could go on Timu and get a shirt that's, you know, oftentimes something like a t-shirt, you're not as worried about the quality if you're just some company that plans to give them away. Say you're an ice cream shop. You'd rather pay $5 for the t-shirt than $25 locally.

And so it really made it almost impossible to compete if you were a localized manufacturer. Yeah. And it's been part of this value chain erosion that's happened where it used to be, you know, like Apple, manufacturers in China, they have a massive facility there, a huge partnership with Foxcon, there's Apple employees on the ground. And so it's very much like an Apple designed and made product with the help of Chinese manufacturing. Then you go to the next level, which is like, oh, an entrepreneur

has an idea. And they're like, I want to just sell T-shirts and they just call a Chinese factory. They basically do everything. It's very turnkey. Then, and then they list that on Amazon. And it's like China's providing like all the manufacturing. Then Sheen and Timu kind of cut out the the middleman and just went direct. And so all of a sudden it went like products come out of the Chinese factory and they go directly to the door of the American consumer. And to be clear, manufacturers like Apple aren't benefiting from this loophole, even if they're selling a $500 product because Apple is sending, you know, a shipping

container with with 10,000 AirPods in it. And so the total value, even though the individual product might be a couple hundred bucks, ends up being far beyond the sort of loophole based on because the, the, loophole is based around the shipment itself. Exactly, exactly. And so T-MU is owned by PDD Holdings, a company that I'm sure will do a deep dev on soon. They rapidly expanded in the United States by offering steep discounts on a variety of products for people willing to wait a week or so for delivery, because it comes literally from China, from the factory. This is another Chinese-linked firm that's in the U.S.'s crosshairs. The popular marketplace, which e-marketer estimates,

will sell $30 billion of products to U.S. shoppers this year. became an alternative to Amazon as well as retail chains such as Hobby Lobby, Party City, and Dollar stores. Like Sheen, Timo has largely anticipated Trump's decision by expanding networks in the U.S. and moving to bigger bulk orders. Timu didn't respond to Quest for comment. American shoppers and companies imported about $48 billion of shipments from the world under the day minimis loophole in the first nine months of last year, according to U.S. Customs and Border Protection estimates. And so we will show.

You can think of what we're doing for technology podcasting. Timo is doing for consumer products. Just flooding the market. And the scale is honestly unbelievable. And the obvious critique of Timo and the long-term critique of China has just been, you know, they're manufacturing products at incredible scale at incredible prices and just flooding the market. and oftentimes specifically in the case of TEMU, you can, if you want to buy an air filter, you can go buy a TEMU air filter.

I'm sure it costs like 20 bucks. Yep. And the critique is that, yeah, you would have had to spend $200 to get a comparable product that was made by a more reputable firm. And the TEMU product's going to just basically evaporate over time, break down and things like that. And so a lot of these products have a pretty short, not shelf life, but just life in general. general, right? Totally. And so there are two charts in here that will show. One is day minimis

exports from China over the last six years from 2018 to 2023. And the top bar there is going, or the bottom bar is going to the United States. But you can see China has wrapped up, has, has ramped up these day minimis exports from about $5 billion in 2018 to $65 billion. dollars in 2023. And so this de minimis loophole was discovered and then exploited to great effect. And you've seen, you know, this, this chart. So this, the blue bar is the United States, but the other, the other section and the rest of the bars there are, presumably there aren't. It's, it's maybe, maybe Malaysia, Taiwan, United Kingdom, Vietnam, et cetera, adopted similar de minimis. Oh, yeah, yeah. Dameanimus is a common construct,

because it increases efficiency of trade. And so other is probably all of Europe, Africa, Latam, like all these different countries combined. And you can just tell that, like, this chart looks like a startup graph, but you don't see types, you don't see graphs like this in global trade. Like global trade typically grows at like 2%, 3%, 10%, not, you know, not 6x or 10x over, you know, 10x over six years. Like that's very, very rare.

and it's all driven by just finding this loophole and then just spamming it, basically. And so let's go into the Stratory article. Ben writes, now that tariffs are front and center in the news cycle, you may see some commentary referring to the Smoot-Hawley Tariff Act of 1930, which dramatically raised tariffs on over 20,000 imported goods in an attempt to protect U.S. industries. So interesting, this stuff is not new. Many commentators feel that it contributed to the Great Depression.

One challenge in implementing that level of tariffs, however, was the sheer amount of manpower necessary to inspect every import and levy the appropriate fee, particularly for small items mailed to individuals to alleviate the challenge. The act included a de minimis exception. And so this is what we were talking about. And here's a quote, that the secretary of the Treasury B and is hereby authorized in his discretion to prescribe regulations for the administration free of duty articles, not exceeding a value, $1,

imported by one person one day. So that $1 figure, it was originally $1 in 1930, it's been expanded. In 1962, they took it to $10. In 75, they took it to $25. In 78, there was a lot of inflation in the 70s. They took it to $50. In 1980, it was $100,288.

And then finally, in 2016, it went up to 2018. The lesson here, go back in time and buy Bitcoin in 1930. This currency to basement is concerning. Yep. And so this has been a boon to businesses like Timo and Sheen, which sell Chinese-made products to U.S. consumers and then ship them directly. So now China dominates de minimis shipments. A key part of China's global e-commerce growth has been the expansion of PRC and PRC-tied e-commerce firms into the U.S. market.

the U.S. retail e-commerce market constitutes over half of all global e-commerce sales. And so imports under Section 321 of the Tariff Act of 1930. And Ben, I don't know if Jordy's not on screen, but you might just want to go to me. Section 231 allows for U.S. imports under a de minimis threshold to enter free of tariffs, fees, and taxes. In 2016, Congress raised the threshold to $800 per shipment. and that dramatically changed the landscape. And so let's go to as China is dominating,

this is obviously important, particularly for U.S. e-commerce and its associated ecosystem, but this has also played a significant role in the fentanyl epidemic. Reuters has an excellent in-depth investigation of the relationship here. We will have to read that on the show soon. And so the big question is, how is this affecting Amazon and META? And I don't know what happened with Jordy.

Let's see. I'm still here. I think my camera battery does. You can probably just switch it over to the computer camera. And I think you'll be good there. And we should be able to see that. So Stratectary, however, is a tech newsletter. And the angle I am most interested in is the effect on U.S. tech companies.

That list starts with Amazon, which will likely be affected by this rule. And so the information reported previously that Amazon plans to launch a section of its shopping site featuring cheap items that will ship directly to overseas consumers from warehouses in China. So this de minimis exception was such a big deal that Amazon was like, well, we have to compete. We can't just get left behind here. So we need a TEMU competitor within the Amazon ecosystem. And so those orders, they're not going to be prime. They're not going to be next day.

They're going to take nine to 11 days to get to customers. but they will stem the bleeding because it's a directly competitive product with TEMU. And so it's not clear if these Amazon shipments will be made using a U.S. trade provision that exempts individual packages worth less than $800 from U.S. custom duties. Sheen and Timo are among sellers that use the provision, which some U.S. politicians and trade groups have criticized as a tariff loophole. So it's hard to understand the point of this new offering if it's not leveraging the de minimis loophole.

Like that's the only reason that Amazon would do this is because they're interested in leveraging the loophole. The program did launch last fall, by the way, and it's called Amazon Hall, which I think is so funny. What a great one. Amazon says, and that's probably a reference to people ordering 100 items on T-Mu and then doing these sort of unboxing videos, these TikTok videos. I'm sure that the thing is TikTok has been using, they do this. warming feature, which we talked about on the show yesterday, where they will show your videos to more people if you talk about specific topics. So over the week, over the last week, it's been if you talk about deep seek, you're going to get a lot of views. And so I'm sure

they did a number of things where if you do Timo halls, you're going to get a bunch of a bunch more views. And so just blatant proof that the Chinese companies coordinate to use TikTok as a channel to influence activity in the U.S., which is, again, why is the company still operating? I don't know. I'm sure there's deals being done, you know, behind the scenes. That's the big reason why Ben Thompson was advocating for Amazon or Walmart to be the logical acquirer for TikTok if there was a divestiture,

because that is the ultimate monetization for TikTok more so than just running ads. And so if you sell it to meta, is this crazy? It is kind of like anti-competitive. They're building this crazy monopoly. Google doesn't really need it. Microsoft, it doesn't really make that much sense. But Amazon or Walmart could be perfectly positioned because TikTok shop has been so successful. Well, we learned yesterday from somebody in the chat, too, that Walmart had it, took a shot at social media with their WM Connect product back in the AOL days.

So Walmart is no stranger to, you know, social media. Yeah. And so Ben is throwing some shade here. Amazon on the truth. He says, I suspect when Amazon says, we also hear from customers, they mean we look nervously at Timu and Sheen's rising sales numbers.

Hall, interestingly enough, is only available on mobile where Timu and Sheen derive their sales. In other words, I still believe in my original thesis. Amazon would love for this loophole to be closed, a world where everyone pays the same import duties is one where Amazon's delivery network can be much more effectively brought to bear.

Yeah, it is interesting to think about how these companies realize that even though U.S. consumers basically now demand two-day shipping max, like your conversion rate will suffer if you're not offering competitive shipping. Yep. They realize, hey, if we can get somebody 50 items for what it costs to order, you know, five items on Amazon, they'll wait, you know, they're happy to wait seven days in exchange. Yep. And so Timo and Sheen are not responsible for all of META's growth,

we're going over to how this is the effect Facebook and Instagram. But the two companies combined, it's estimated spent just under $3 billion a year to date. Meta's revenue is up around $11 billion year over year. And so the growth of Timo and Sheen, certainly they're spending more. They're driving a lot of growth, but META's growing independently. So it's less of this existential issue. If Tee Mew and Sheen go away, it shouldn't be that much of effect on META. Moreover, this spend is more palatable and potentially sustainable than TikTok's spend was.

So when TikTok was spending on meta, they were like directly stealing, you know, users. But e-commerce apps are not a strategic threat to meta, quite the opposite, in fact, and also lack of network effects to make themselves sustaining entities like TikTok has become. So the subsidies could go away. Macro events could intervene. Meta still deserves credit for figuring out ATT, but Timo and Sheen do make the story a bit more. complicated because while ATT was kind of throwing all this curveball towards meta and they were figuring out how do we get the small mom and pop e-commerce engines going again well timu and she came in and were like here's three billion dollars to like you know keep the revenues up into the right which is great

yeah and so this quarter did turn out to be the peak of timu and sheen's share of meta advertising around under 7% on facebook and just under 4% on instagram and they're down to around 4% and 2% respectively now still to the extent their businesses are impacted by these measures and notably an important part of their hedging has been building distribution centers in Mexico there may be some impact on meta's ad prices i don't think it'll be substantial is what ben says and so very very interesting uh very very interesting discussion of how the trump tariffs and changes in regulation are are affecting tech he ends with a little bit about AI and abundance it's worth pointing out that the de minimis exception exists for a good reason.

We mentioned this. There are a lot of items that are shipped into the U.S. every year, and it's understandably difficult to examine everything. It's also worth pointing out that any item that qualified for the de minimis exception also gets to use what is known as informal entry, which means much more limited paperwork. No needs for a custom bond. Informal entry applies to shipments worth less than $2,500.

To that end, it remains to be seen just how much of a difference these Terricks make. and exactly how the U.S. Customs Authority handles the onslaught of packages. So, you know, you're doing doge. You're trying to cut down on the amount of government employees. Well, you know, you might have just 10x the amount of work that's going on if you're trying to investigate and tax all these packages. Yeah. But he ends by saying, I would also note that this seems like a very interesting problem

on which AI could be brought to bear. There's a story here about how the Internet scale simply crushes a loophole created for an analog world. perhaps there is a solution in which AI's ability to ingest and understand vast amounts of data could make the chaos, could make sense of the chaos in a way that preserves the convenience while solving a problem of abundance. Of course, the chasm between theory and reality is a massive one, particularly in terms of government bureaucracy. But I do think that this is one of the many examples of where AI will have unexpected benefits. And so we got to get Luke Ferretor into the

Is it Department of Homeland Security? I don't actually... Yeah. I don't know. Yeah, yeah. At the ports. He's got to go to the port. Yeah, yeah.

He's got to open his laptop and set up some LLMs to look at every day minimums import that's coming in. He has U.S. Customs and Border Patrol is the one. It's the customs department. And so let's move on to our third story of the day. A wonderful deep dive. So that last one was from Ben Thompson, Everett, St. Highly recommend that you go subscribe. It's a fantastic, fantastic newsletter in your email

three times a week and you also get a bunch of podcasts and much of great stuff with it. So highly recommend Ben. Let's go over to Ashley Vance, who now, that first piece we read about Luke Fairtor, that was at Bloomberg, but he's no longer with Bloomberg. Ashley Vance has moved over to core memory.com, his new media company, he's independent, he's going direct, and he has a fantastic write-up of the boom supersonic test that happened in the Mojave. And he says, take me to bed or lose me forever in the supersonic corridor. The city of Mojave has a population of about 4,000 people and sits 100 miles north of Los Angeles. Some people might describe it as a dump, I guess, and they would not be altogether wrong.

As the name indicates, the city of Mojave is in the Mojave Desert. It gets about two to six inches of rain a year. Experiences extremes of heat and cold and has gusting winds. California State Route 14 rips right along the city's edge. And if you're not on it or not paying attention, you could drive by Mojave in a flash, missing out on the delicacies offered up at a Denny's Wiener Schnitzel or Subway or a fine stay at the two-star America's best value in.

Core memory has a need, a need for reeds. Become a free or paid subscriber and support our battle. against Nameless Faces enemy. Let's go. I love reading the calls to action in the middle of a substack post. So go subscribe if you're not subscribed already. All of this is to say that I love Mojave and go there as often as I can. What drew me this time is this boom supersonic launch.

And so the airport is flanked by Edwards Air Force Base and the Naval Air Weapons Station, China Lake. China Lake is a fascinating story. we should dig into at some point. That's where they do a lot of like crazy military tests. That means the Mojave has an exceptional, sometimes eccentric aircraft doing wonderful things and that the flying stuff is complemented by stuff that goes, boom. The event is catalyzed.

The event that catalyzed the Mojave's position as a special place incorporated all these characteristics. It occurred in 1947 when Captain Charles Yeager became the first human to go past the sound barrier in the Bell X-1 plane. In the years that followed, the military bases did most of the work pushing the Mojave's reputation as the forefront of aerospace adventure. Tom Wolfe captured the scene well in the right stuff where courageous manly pilots were turning up in the desert to flex their cahones, and young women were apparently all about it.

Wolf describes both groups mingling at the happy bottom riding club run by the iconic Florence Low Pancho Barnes with language that I'm not sure. mainstream writers would get away with today, although labial piping little birds is evocative. His writing is this is why when Ashley Vance started core memory, legacy media died, because they lost. They lost Al Star. I think this type of writing makes a comeback with Ashley Vance in the wild. He doesn't have an editor now. He can say whatever he wants.

Let's get the flowery language out there. Let's differentiate from the LLMs, the slop. We're not going slap when we're reading Ashley Vance. It's good stuff. Yeah, OpenAI is hitting refresh on CoreMemory.com waiting to ingest this incredible flowery language. Yeah. And so he's there to talk about Boom Supersonic. Last week, Mojave, the city, returned to its full glory, at least for a moment.

Boom Supersonic flew its XB1 jet past the sound barrier for the first time. And I got to witness the whole affair. And he's making a video about it. so go subscribe to core memory. Boom was founded in 2014 by Blake Scholl, and he had nothing resembling an aerospace pedigree. Scholl, a computer scientist, had worked at Amazon and Groupon,

dear God, before starting Boone. Yeah, so his LinkedIn was getting shared around because he was a product manager at Groupon right before starting boom. Yeah. So what an outlier bet. He would have been nicely pattern matched to go start some type of retail, you know,

app SaaS company, but not the guy that's getting pattern matched to start a supersonic aviation aerospace company. Yeah. And so even with his pedigree, he still, he went to YC. They gave him a lot of coaching on how to get, how to pull forward like contracts and kind of like not like real revenue, but like if you make it, we'll buy it and that de-risk for investors. And so all of a sudden, I think I think the combination of the background, the YC strategy, the demo day, all that stuff got the business off the ground and got the initial funding. And I love that when people talk about him working a Groupon, he's like, yeah, now I'm just trying to do 50% off coupons for how long it'll take you to fly from L.A. to New York. Yeah. He has a great one. He's got a bunch of zingers based on it.

And so here we are 11 years later. What a journey. And Boom is still at it. And the XB1 is not even the aircraft around which it hopes to fly regular folk. Rather, the XB1 is a single person jet that Boom has used to test out some of its ideas, materials, and team when it comes to handling the pressures of supersonic flight. The plane that could end up taking people from London to New York in a zip is called Overture and will need a few years and tons more money to create and clear for safe travel. We've obviously been here before with the Concord, but it had the weight of the British and French governments behind it. The XB1 has been built on the back of venture capital, and when it went supersonic last Tuesday, it marked the first time an independently developed jet

had done so. So it was grab with a with a 50 person team right. It was a 50 person team on the on the XB1 project. They do, I think they have 140 employees. A lot of those people are working on overture and other stuff. But still the fact that you can fit 50 people in a shot with the plane is incredible. And crazy.

Coming off of the back of a recap that, you know, a lot of their investors had to, you know, they had to say, hey, we're going to, you know,

basically double down, triple down on this thing to make it work. Incredible timing. And I'm imagining this success of this latest launch, our flight should attract some more capital and allow them to continue to hopefully thrive. 100%. 100%.

So after the flight, Scholl groused on X that neither the New York Times nor the Wall Street Journal had bothered to cover the event. That's warranted grousing. calling them out, Ashley. Don't let up. But then again, you know, I'm glad that we have core memory. You know, I don't know if I'm going to read a New York Times right up.

Oh, it's annoying that they went fast. I don't like it. Slow down. Slow down. Too fast. Okay, so he says, yes, okay, boom had already flown XB1 about 10 times from Mojave at subs supersonic speeds. Shoal will tell you that the company has learned most of what it needed to during these test flights. and there's also a degree to which the XB1

busting through the sound barrier was simply symbiotic. But also, WTF. There are pragmatic reasons this test flight was such a big deal. Boom is one of our last best shots at getting supersonic commercial travel. Concord retired in 2003 because it was too expensive. Afterwards, people were inclined to deem affordable supersonic flights as an impossibility, which is ridiculous.

Yeah. We have a couple major supersonic bets going with Boom and Hurstall. firmius, big fans of both companies, but they're both battling away in a horrifically hard and costly business. And if they fail, we might not see anyone take a swing at supersonic commercial flight for decades. Maybe you like being wedged in a seat for 11 hours going from San Francisco, San Francisco to Tokyo. I don't. I'm with you, Ashley. Yeah, no, it really, it really is wild to think about just cutting, cutting, even cutting in half, being able to get to Europe. from from L.A.

in the time that it takes to get to New York would be incredible. You're shrinking at a time when in many ways, looking at the trade war story and things like that, Peter Zihon was right in his prediction that we're going through this period of de-globalization, but supersonic flight would have an opposite effect in that everything would become closer, right? Being able to get to Tokyo in a couple hours, very cool.

Sam Altman would have, would love that. He'd get to go hang with Moss. a lot more regularly and go car shopping as well. I love it. I love it. And so Ashley's waxing poetic about the impact of this. He says on the far more visceral front, boom, doing this test flight was not just cool, but important for all the historical context reasons I laid out earlier.

Some of the U.S.'s best engineering achievements have taken place in the Mojave because of the culture built up in the Mojave. It's a special place and it requires tending and fuel to keep on thriving. The U.S. is a rare country to, the U.S. is that rare country to even have a supersonic corridor and to have supersonic ready objects to shove through it. It's good to test such engineering limits and be reminded of what we're capable of and to maintain the need for speed. One of my favorite Mojave Aerospace types is Ben Brockert,

who went on to work at Astrospace after a stint at Maston. I wrote about him in my book When the Heavens went on sale and include a tribute to Ben from that book here. Having pine to get into aerospace since he was a kid, Broker bought a $500 very, very used 17 passenger van and drove it to the Mojave Desert to begin knocking on the doors of space startups. It took months to find a gig, living out of his van and trying to fill the days as best he could. I managed to get a library card and read all the books I had not read, he said. I also spent a lot of time bombing around the desert and trying not to die. through a stroke of luck, a position opened up at Maston, and Brockert soon established himself as an essential part of the zombie team. He continued reading aerospace books and learned a ton

on the job over the course of three years. Mojave is like a new space penance. He said, it sucks, but you put in your two to three years and then you go do something somewhere else if you're good. And so what a fantastic anecdote about just like, if you want to be an AI, you probably got to go to SF, if you want to be in defense tech, you're going to be in the If you're going to be in supersonic stuff, you've got to go to Mojave. And I love that. It's not so bad out there. I'm sure you've driven through the desert at some point.

It's got a kind of a, you know, a super galactic vibe out there. I'm bullish on the Mojave. We should do some on the ground, on the ground reporting in the Mojave. Ashley, next time you go, please invite us. We'll tag along. And whatever news you're covering and breaking, we'll front run you a little bit. We'll break it live. But then we'll still cover your article the next day. Yeah. Well, let's move on to some back and forth we had on the timeline with Patrick Collison,

the founder and CEO of Stripe. We reacted to his fantastic post about two different types of phone booths. You might have seen this on a previous show. And he followed up by showing us an image of a beautiful building, which I think he has his eye on, maybe. I don't believe this is the actual Stripe headquarters. It looks somewhat AI generated. I couldn't find it on a Google image

reverse image search, but I think he got the right idea. Patrick, I want to see this, I want to see this in action. And keep making everything beautiful. Like, what you're doing is so important to tech. And I just, I mean, everything from like the blog and the fast,

He has a whole documentation of all the different projects that have been done very quickly. He's an incredible thought leader, I guess you could call him in tech, and just an inspiration to everyone who's building and wants to make beautiful things. And so happy to have a back and forth with them on the timeline. Yeah, and this is a great example of a founder with incredible taste, who now has the resources to go out. And instead of building the Apple Loop, whatever, what do they call their campus? He can build this structure that looks like, you know, what maybe a bank would have looked like during the Roman Empire, right?

Yeah, yeah. It is very funny. It is. I mean, they're not technically a bank, but they do a lot of, you know, FinTech is bank. Financial institution, right? Financial institution. And yeah, we love Stripe. Stripe helped us incorporate technology brothers LLC, which was very helpful. Certainly did. So shout out Stripe Atlas. Give Patrick. call us in a call, tell them the technology brother sent you. Go incorporate your next business on Stripe Atlas today. We love Stripe. And thanks for engaging with us, Patrick. We love to have you on the show.

You're welcome anytime. Anytime. To some breaking news, Jason Carmen has launched his new company. He's back. He's back. He took a little break. He was working on just one 90-minute movie.

No big deal. Stop posting the weekly documentary. which were already insane and crazy high production value. In many ways, let's just be honest, in many ways, John Coogan walked so Jason Carmen could fly, right? For sure. He kind of came in. He was like, hey, John's a little distracted with the podcast, you know. I'm just going to launch a series of highly produced films that will go on to live on, I'm sure, Netflix and Amazon Prime and all these platforms.

So it's amazing to see. The guy sprints everywhere and incredible energy, unrelenting drive. He says, this is the company I've wanted to build since I was nine. And I'm so glad he landed here and doing this. Obviously, he had a ton of different opportunities. He was in-house at a startup for a while. This is such, it's just a perfect fit for him. He's going to be able to make science fiction films as well as documentaries, corporate work, you know, stuff on Netflix.

I saw his camera. He's able to post behind the scenes now. He's running the Sony Venice 2. This rigs like 100K. He is beyond the requirements to deliver to Netflix at this point. Like all of his gear is ready to be shown almost in IMAX. Like it's that level of production. And tech has never seen anyone who's this much of an artisan and craftsmen care about this type of media in this way.

It's always been a side project. But this guy, this is a. life's work. This is clearly what he's doing. And he's gone full send. And we're happy to support him here at Technology Brothers. Yeah. And it's amazing to see this rise. I think when you think about legendary filmmakers of today, you know, that built their careers across 20, 30, 40 years. Yep. You would get to see their films, but you wouldn't get to see behind the scenes in the same way of seeing their journey. Maybe it's an interview with the New Yorker or, you know, they do a talk at Sundance, but you're not getting this sort of daily,

seeing the progress and seeing somebody build a career, which is amazing to see. So congratulations, Jason. Congratulations, Jason. Keep up all the hard work and keep making incredible content. Keep posting. Keep posting. We have an update on a post from yesterday. Mateo over at eight sleep. Absolutely dog. We both sleep on eight sleeps. We highly recommend eight sleep. The pods were delivered to Doge. I hope this will supercharge Elon, and the incredible team to shape the future of America. Let's go. And so they packed these up.

They ran them through the metal detectors, through the scanners, the X-ray scanners, and they got them in the building. And incredible, quick turnaround. I didn't even know how they got them out of the factory that fast or out of the distribution center. But they were there in D.C.

in less than 24 hours, delivered. And so enjoy it, Luke. Enjoy it, the rest of the Doge team. I hope you guys get high sleep scores. Luke, for the four hours, Luke is sleeping a night, it's going to be the best four hours of sleep of his life. And yeah, if you don't have, if you don't, if you can invest a lot of hours in your sleep, you definitely got to be on an aid sleep.

Yeah, I mean, A8 sleep also sponsors F1, these elite athletes. You think about Luke, he is an elite athlete of the thing government. The thing too that I don't think, I don't think people realize about eight is, is that the cozy element is crazy. So a lot of people think about the chilling, which is nice, you know, maybe you get into and you want to be a little bit cooler, relax. But that morning, you know, turning up the heat a little bit in the morning, it's like being next to a warm fire.

So it warms my heart knowing that Luke, you know, is going to be in the depths of the federal government, you know, bureaucracy. And he's going to be able to be cozy if he wants and maybe even use it as sort of an alarm feature. It's like as you start sweating, it's time to hit, it's time to hit, you know, hit the tree. Put out a new CSV. Pull out a new stack of PDFs.

Start organizing. Digging through the Treasury docs. Let's see. We got some more massive breaking news. Callie Means has posted about RFK. Let's be very clear. RFK is the most important cabinet confirmation in American history. Bobby and President Trump have gone up against the largest and most powerful industries in the country and our winning.

That would be the food and drug industries. The result is a chance for a healthier future for our kids. And a good friend with Callie, love Callie and everything. that he's done to popularize health and wellness and spread a lot of just very interesting and contrarian ideas. I'm sure you've met him. Can you give us some background on Callie and his business partner, Justin? Well, John, Callie's an absolute dog. Co-founder of TrueMed, co-founders with Justin Mares, former brother of the week. TrueMed makes it easy to use your HSA FSA funds to buy products that you're already buying, which is super important because

that HSA, FSA benefits are already cool, but it's not, historically, hasn't been easy to spend that money on consumer products. We got a chance to hang out with Cali and Justin at Hereticon a while back last year. And Callie even then was spending hours every day working on the Maha movement, which is an interesting, make America healthy again is something that we should all be, get behind. It clearly was politicized in many ways. Totally. But I'm, I'm glad to see this get through. And there's so much more to, you know, this is not a political move. This shouldn't be political at all, right?

We're trying to make the average, you know, American healthy. There's a lot of aspects of our sort of way of life that simply aren't. And, and yeah, I mean, what they went up against here is insane. The, you know, the pharmaceutical, the drug industry is spends billions of dollars on advertising. With legacy media, they spend tons of money, you know, lobbying pretty much every single politician in Washington and at the state level. And so what they went up against to get this done is absolutely incredible. And, you know, personally, you know, we don't talk about politics on this show, but I firmly believe it's not a political issue at all. It's just this will be good.

You know, Americans will get more information. And hopefully we can create some. guardrails to help everybody be healthier. So I think it's a huge win. For sure. For sure. Congratulations. One interesting note.

I'm going to potentially butcher the details here, but just I didn't realize, you remember seeing like we pretty much spend all of our time on X, but if you've ever watched, you know, CNN, you'll get like three ads back to back for drugs. And you're thinking to yourself, this doesn't seem like. a very good ad, right? Because it's like quick shot of a guy like driving his truck and then like wave into his wife. And then it says, this drug will kill you. This drug will put you into a long-term coma. Like this drug will do all this. And so, so I was always so confused. I was like, how is this actually good advertising? Because I come away from this being like this could,

this could maim me for life. You know, like it doesn't seem like an ad. And one of the one of the reasons that the pharmaceutical industry has had such a chokehold on legacy media is because they're simply spending so many dollars, it would be almost financial suicide for someone like a CNN to ever put out negative coverage of the pharmaceutical industry. And the pharmaceutical industry does many, many important innovations that help us live longer, better lives. But for a really long time, the legacy media just hasn't been able to talk about some issues with the pharmaceutical industry and some of the downsides. And so the fact that they were able to get this through

despite having the, you know, the pharmaceutical industry having so much incentive to not have this happen just shows that the time to make America healthy again has really come around. So congratulations to RFK and Cali and everybody involved in the movement. And I mean, that point that you just made sounds very political, but it's been that that exact.

exact point has been a major talking point of the left for generations. Like, Noam Chomsky has the concept of manufacturing consent. And it's exactly that. And so, like, you, you could be quoting RFK in the modern era where he's a Republican.

You could be quoting RFK from a few years ago when he was a Democrat. You could be quoting, you know, one of the biggest socialists in Nome Chomsky. Yeah. So this shouldn't be as much of a political football as it is. But, unfortunately,

it has become one. I hope people withhold judgment. and understand that there's a very positive intention with this movement around health. Totally. And let's see how things go, right? And at the end of the day, every American should be able to make decisions for themselves around their health. It's like one of the greatest personal freedoms that we should always have. So I'm excited to see what comes to this.

Yeah, it's undeniable that lifespans have flatlined. America is getting unhealthier, decade by decade. Everyone kind of agrees on that. Clearly, that's a reason to try something different. This is different, and hopefully it's the right move. So I'm optimistic. Let's move on to Nico over at default. He had some press for his company written by Brandon Short,

Brendan Short. And I love this because included in this is a fantastic quote. He says, Nico has moved to New York City and has an office with a bunch of crack engineers, a hungry team of go-to-market folks who are at the bleeding edge. And to use Nico's words, enough Lucy and Celsius to topple a sub-Saharan warlord. The team is building their company in a way that I wouldn't want to compete with. And so thank you, Nico, for the shout-out. I love default. We need to put it to use on the show. Get some outbound going. Yeah, we need our own outbound, you know, guys just dialing, you know, on the folks.

all day long, cold calling people say, hey, have you heard about the technology brothers? Exactly. We're asking about you. They want you to listen to the show. Yep. Give us, you know, just three hours of your day every single day. We don't ask for much. Just three or four hours a day.

Yeah, exactly. Totally free. It's hand-to-hand combat. Every single listener matters. We need SDRs that are convincing people to... Pounding the RSS feed. Hey, we'd love to upgrade you from just a listener to live stream watcher, well. Have you looked into this? Oh, have you left five stars yet? You know, really develop a

relationship with them. It's a relationship business. This podcasting is a relationship business. Yep. Well said. It's fantastic. Let's move on to the timeline. We got a funny post to kick it off with. Apparently, over the last decade, the purity of cocaine has increased significantly, while the price of cocaine has remained the same. And the question is, what other industry slash products has this happened in and Hunter says it's worth noting that private equity has largely avoided the industry and the interesting thing here is that this this this this has actually happened this this trend has happened in a lot of industries specifically tech industries uh google search has gotten better while remaining free like youtube has gotten more content without changing the price

and there's a lot of TVs are TVs are basically free but i mean so this this reminds me i posted yesterday that Pablo Escobar might be the greatest CPG founder of all time. And I'll just read out some of the stats. So Pablo obviously contributed to this trend. And private equity was actually the counterparty to this trade that Pablo was making in many ways. He was dominating the market in the 80s. He had about 80% of the TAM at the time. So just to get some ideas of the scale,

Pablo bootstrapped to annualized revenue of $22 billion. He didn't really have access to the private markets the way certain entrepreneurs would. Global operations of sales and distribution hubs on nearly every continent developed an Amazon-style logistics division. He had his own trucks, boats, planes, fulfillment centers, ports, et cetera. He expanded margin and improved deliverability and traceability. And this is one of my favorite ones. He actually built a prison for himself so he could lock in and really sort of inspire himself to grind harder. and then he escaped from it just to just to prove that he could.

Fantastic. And he even had a run at politics, too. He became elected to the House of Representatives. So, yeah, I'm sure that private equity is bummed that they couldn't get into that market. But yes, other people took advantage. And Escobar is the one that's chronicled in Narcos season one. Is that correct? Yeah.

Oh, yeah. Great show. The last show that the last good show that Netflix made. Oh, yeah. It's been. a couple of years, but I remember watching that, five, six years ago. It was great. Highly recommended. Yeah. Well, let's move on

to Vittorio. He says, B. Trump, 4D. Chess Grand Master, slap 25% tariffs on Canada and Mexico, secure the border or else drama ensues. Canada and Mexico are losing their minds. Canada's screeching about retaliation. Mexico sweating, concedes first and agrees everything. Tariffs

are paused. 24 hours later, Canada folds two. Total victory. Economy saved. Border secured. America first. how does he keep getting away with it? And it is, it is funny. Like Trump the dealmaker, he knows how to come in and put the hurt on someone and really expose the leverage. And it's interesting that he, you know, presumably chose to do it over the weekend when the markets weren't open. Yeah. And so if that was intentional to try to minimize some of the chaos, you know,

it worked well. It was interesting to see how quickly everybody went from, you know, you can't just stop the flow of fentanyl it's it's impossible all the stuff to trudeau saying yeah we're we're sending choppers and the military to the border and we're going to we're going to stop the flow um and so anyways um hopefully it has a positive impact uh fentanyl is is uh you know our our our version of the opium wars yeah so um very bad so let's get it out of the country but hopefully you know no market turmoil and and less illegal drugs. Like that's a win forever. I think everyone should be happy about that

if that's the outcome. And hopefully it will be. Well, let's move on to Aaron Levy, founder and CEO of Box, cloud storage company. He says, can confirm Open AI deep research is quite strong. In a few minutes, it did what used to take a dozen hours. The implications to knowledge work is going to be quite profound when you just ask an AI agent to perform full tasks for you

and it comes back with a finished result. I, what is your experience with deep research been so far? So, we've used it for a couple projects this week that are relevant. So yesterday we did a deep dive on Mary Meeker, absolute legend, icon of the bubble, and more recently a venture capitalist. I thought it was very competent at researching Mary Meeker. And the reason for that is that the project was basically collect information from the entire internet, try to rank and get the highest quality sources and then organize it in a sort of

linear storyline and help me understand her strengths, her vision, what was her impact, things like that. It did a really good job at that. Today, you tried to use it for Luke Ferreter to try to get a better understanding of his background. And I think it was much weaker the output. And that's obviously because there's just much less information about Luke online and sort of super high quality sources. So anyway, yeah, my thoughts on your area, like, I can't tell if I'm, I'm probably some combination of bad at understanding the correct way to prompt deep research to get great results.

I noticed I was dumping in whole text exports of PDFs and it was kind of just breaking and like bugging out. And so there's probably a better way to interact with it to prompt it. And I'm sure we'll learn these patterns. And I kind of need to unlearn some of the patterns that I learned when I was using 01 Pro or GPT4. Because for a lot of those, with some of them, they have the ability to pull things from links. So I would dump a bunch of links. It would pull all those in. Other ones, PDF uploads. So I would upload a bunch of PDFs. And then it would have that as extra context. With the deep research, it does a lot of it by itself. So I think with just a simpler prompt, you can get a better result, but I think, like, doing too much handholding might actually be a problem for it.

And so there's still the question of, how do I get it a source that it doesn't have access to that's maybe behind a paywall or some arcane scroll that it doesn't know about? And I want it to be in the research report. And then also, there's just a question of, like, for a lot of these, for some of the really popular topics, like, there might already be a deep research stuff. report out there just in the form of the Wikipedia page or in the form of some sort of podcast. Somebody who's posting that it's basically a new form of Wikipedia that you can sort of slightly customize and prompt. But yeah, it's interesting that OpenAI made a huge effort to hype this launch, right? They have a very good network of investors, team members, Sam, et cetera, that can create a lot of hype around a launch. And a lot of people have been working

on the same idea using LLMs for research, right? And so there was a bunch of different product layer companies doing this previously, but just goes to show that opening, I do think they're feeling the pressure around the model layer commoditizing and potentially good models going open source. And so they're just needing to innovate more at the product level and say, hey, we're actually not just the model layer. We're operating at the app, the product layer as well. I mean, the easiest win, and I don't know, maybe I'm playing like armchair product manager right now,

but the easiest win I could imagine would be just there's only a text box. And when I prompt it, in my prompt, I say, I'd like you to deep research this. Or I'd like you to go online. Or I'd like you to use a PDF upload. And then it routes me to the correct model for that. And it intelligently knows, hey, this prompt is perfect for deep research. let's go over here, 03 high. 03 mini is actually best for this instead of putting that on me because now I wind up with, I write the same query in five different tabs with all the different

Yeah, here's here's how, here's how the workflow will actually work, right? As you go to XAI, you say, hey, I want to do this thing. How should I do it? And it says, well, I can route it here if you want and then bring it back and then or I'll try this and I'll try it in a couple different places. and so it's interesting to think about how you might use a model to help you better understand how to use other models and just get the best output 100%. Yeah, there's already a version of this happening with like the open router system where a lot of the programmers who want to use an open source model will they instead of just saying like I want to use Deepseek v3 hosted by deep seek, they will say, I just want to use this model at the lowest possible cost.

all of the different servers and companies that serve these models are out there competing on price. And whatever load balancing is going on, whoever has the lowest cost of energy or GPUs that day, low demand, will price their model dynamically and route you to that. We just need kind of like a new layer, a new abstraction layer maybe. Because, I mean, it'd be very cool to have like, you know, there's one place where I go to prompt, oh, looks like you want to upload a full book, a million token context. Gemini for that. Oh, looks like you want to do some deep research. Chat GPT deep research for that.

Go here, do this and just kind of route that intentionally. But I do think all of this will be solved very quickly. It seems too obvious to not be arguing the works. And this is one of those things where we're talking about it now. They're probably putting the polishing touches on it right now. It'll go out. Like once it's a meme, the company knows, especially if it's like, you know, a fast-moving startup with a founder mode CEO and all this other stuff. Like it's not, it's not like, talking about Google where people will say for years or Apple, people were beating the drum of like Siri doesn't work for years and it took them so long.

Open AI, like they're moving fast. They got, well, yeah, look at Apple today. Everybody's saying your photo app is the worst app that I've used in forever. And they're like,

okay, we heard you loud and clear. We're going to launch a part of full competitor now. It's just like, please just fix the, uh, the photo. app.

Yeah. Anyways. Anyways, let's move on to Nikiel, Krishnan. He says, if there's a market map, you're too late. Do you get this reference? Is this a Kanye? Oh, it's Drake. It's Drake.

Okay. It's a cover art. Okay. Wow. I don't know. I'm out of touch. You just out of yourself as not being a Drake fan, which is not the worst thing to be out of it. Kendrick all the way, baby.

Kendrick all the way. Anyway, it's great, great graphic. I mean, this is, I mean, the original album, if for those that aren't Drake enthusiasts or didn't pay attention to pop culture like 10 years ago at this point, is if you're reading this, it's too late implying. I'm sure he had some sort of beef with somebody and he was dropping the album and, you know, you know, was saying if you're reading this, you know, I already dissed you or something like that. in this case, I do think it's true. A lot of the best companies get started before there's a market map. You look at Anderol and Palantir, you know, these, yeah, maybe you could go back and look around when they were incorporated and make a market map, but it's really more like by the time the market maps are being shared publicly on X, you're, you know, that is not the right point to try to

start the company, right? Yep. And it's not to discourage entrepreneurship or getting excited about a category and making a bet there, but usually by the time the markets map, there's a bunch of super, super talented people that are already tackling different areas of it and maybe have a two, three, four, five year head start, right? Sometimes even more than that. There are two notable exceptions I can think of.

First would be Google because there were a lot of search engines and they were definitely mapped in Mary Meeker's report, X-site, Lycos, Yahoo. It was very clear that there was a defined market, and maybe it wasn't a market map on Twitter because Twitter didn't exist at the time, but it was clearly understood. And then the other one would be Facebook, actually, because LinkedIn had already started, and there, I believe one of the early investors, maybe was Reid Hoffman or somebody in Peter Thiel's network, had actually started a company called like social networking.com or something like that. And a lot of people were circling. And Peter has said that like he knew that there was going to be someone that cracked the code.

And so when he met Mark Zuckerberg, he was like, I'm in. And so there are some exceptions. But but yeah, I mean, once it's a full market map and there's like hundreds of companies, it's a little bit rougher. It's more, it's more just like, okay, if there's a few competitors and they're doing something really fundamentally wrong, there's probably still innovation opportunity, but if you're just like looking at a market map, it's going to be a rough go. Anyway, let's move. Let's stay with meta, actually, and move into a news article from the Wall Street Journal. Meta in talks to reincorporate in Texas or another state exit Delaware.

Potential step by social media giant follows similar move by Elon Musk, who moved to reincorporate his companies in Texas and Nevada. And so, Sheal, Monot says, Delaware probably lost its status as a place where most companies choose to incorporate. Huge hit to the state budget. Forty percent of general fund comes from franchise taxes and corporation fees from companies who choose to incorporate there due to the business friendly approach. And so it'll be interesting to see like YC and Stripe Atlas. They still kind of recommend Delaware. It's like the standard thing. But these things change over time. Yeah, hopefully I'm not lagging too much. It seems like my connection's dropping a little bit.

But yeah, so give some people some backstory here. So obviously every venture back startup has been incorporating in Delaware. I've invested in over 50 companies. I don't think I've invested in a single company that wasn't incorporated in Delaware originally. And yeah, you can register in other states where you kind of operate. But the actual incorporation happens in Delaware because there's so much case law there. It just makes sense to do it. Sort of been the default.

and Delaware's really benefited from that. Now, what happened and why META is making this move is that there was a very high-profile lawsuit. I think we covered it at different points where there was a Tesla shareholder who had something like seven shares or something like that, or maybe it was even,

it was like very small number of Tesla shares, and he basically partnered with a law firm or a group of law firms to bring a very aggressive case against Elon's pay package saying that it wasn't the best interests of the company and ended up being, I don't know where it's shaken out today. I think it was the judge, like, ruled in favor of the shareholders and said, no, this is not right,

even though it had been approved, you know, back in the past. And then they did a vote to overturn it. And Elon was sort of rallying the troops and saying, like, hey, I had, and originally the pay package was set up in a way where Elon had to hit the world's, most aggressive, you know, sales and growth targets to achieve the pay package. And he basically did the impossible to do that. And then you had this hyper sort of political judge in Delaware who struck it down. And what that means is every CEO now is saying, look, we, I don't feel comfortable operating out of Delaware if judges there are going to get involved in basically the internal

decision making at the company, right? It just doesn't, it really doesn't make sense. And so now you have people moving to Nevada, some people moving to Texas. And we'll see what shakes out, but a huge cell phone from Delaware. Yeah. And so I just looked up with the story was Elon calls his shot and says, I want more equity in Tesla. And you will benchmark, you will benchmark my performance to the stock price, which is the purest essence of shareholder value creation for a CEO. It's perfectly aligned. He's not talking about revenue or deliveries. He's just talking about if the value of the company goes up, this will be accretive for everyone. You will have, you'll, you'll, you'll dilute yourself because you're giving me more shares, but your shares will be worth so much more. And so during that time,

he like 10x or 100x the stock. It went way up, generational run. And so he hit all of targets. He was ready to get that package. The Delaware court struck it down and then Elon did a shareholder vote. All the shareholders were like, yeah, this is great. Like, this guy is putting up insane numbers. Let's give it to him again. And then a judge struck that down again. So as of December 2024, the pay package has not been approved. But if they reincorporate in Texas, they can try and go approve it again. And then the Delaware judge would not have the authority to block it. And so... Yeah, and to be clear, the real devil in all of this is the law firm that I believe made

hundreds of millions or billions of dollars in fees as part of this lawsuit, even though the original plaintiff in the case, the shareholder who brought the case, had had something like seven shares or something like that. I need to dig into more of where that money actually changed from. terms of the person who... Yeah. I need to dig into more of like where that, where that money for that law firm came from because, you know, the money had to come from somewhere.

It clearly couldn't come from the plaintiff if they just owned seven shares or something like that. So maybe the Delaware court awarded it. I wonder if that's on hold as Elon's fighting back. It's all very unclear, but certainly worth a deep dive and certainly reshaping how large tech companies think about where they're incorporating because I don't know if Zuck wants to do some sort of similar pay package. He already owns so much of meta. I don't think it's that big of an issue, but it's certainly scary thinking that the shareholders collectively could want to do one

thing and then a judge could say, hey, actually, we we don't want you to do that, even though it's massively popular amongst the shareholders broadly. Anyway, let's move on to Gary Tan GTs on the timeline line again with a low-tam banger that I thought we should highlight here because we love cars. He says, if you like cars, watch this space. And he's quote tweeting Bubbs, who says the car community for the car community, it's going to be fire. And he posts a logo for something called rides. I couldn't find any more information on this. But I wanted to track it because it seems like maybe it's a YC company.

Maybe it's some sort of ride sharing or car community. maybe there'll be the ability to buy and sell cars on there. All seems very interesting. I've always thought there was an opportunity to build like a Strava-like app for car enthusiasts. Interesting, yeah. People are constantly doing, you know, I'll meet up with friends to go for a drive. I'll share that drive with other people. There's little rallies that happen.

There's car meets. There's things like that. And then so I'm just totally sort of extrapolating what something like this could be, but I do think that you could use a social network to bootstrap a actual, like, you know, secondary marketplace for cars. And some of these car marketplaces can be extremely lucrative. And so, anyways, cool opportunity there. We'll have to, Gary, send us some more information, and we will, we'll use that and talk about it on an upcoming show. Let's move on to Ryan Peterson. Can't believe this app is only $8.00. Absolute banger, 40K. L.

likes. He says, Dear Doge, please look into the U.S. Mint in San Francisco. It occupies an enormous block of prime commercial real estate. And all it does is produce special commemorative coins for collectors. The coins aren't even cool. And Elon says noted. Noted. Damn, bad day to be a commemorative coin. I got to say, if you're seeing that post get 40 kids. I might have to take the other side of this. I'm very pro-physical commemorative coins, generally anti the meme coins. I think these are the commendery of coins are Lindy. Maybe the designs aren't cool.

That's an easy fix. Let's get a flex port coin. Let's get a SpaceX coin made from the mint. But yeah, I was pro-crypto until I bought the client, the official Kleiner Perkins meme coin yesterday. I think I lost it all. The market cap last time I checked went from 10 grand to something like zero. And so I'm going to have to talk with Everton.

about this and see, you know. You should send him your tokens. Yeah, yeah, yeah. No, honestly, there's like a, you know, commemorative token for me. Leave it in the wallet. Everyone will know. You bought Kleiner. I bought her.

I was early. But the other thing. It was a hacked project, by the way. In case you're just tuning in for the first time. This was a hacked meme token on pumped out fun. We do not support this and it's not financial advice. That is correct. anyway let's move on and and it's worth noting that that the app uh would cost negative eight dollars

uh you know negative like like uh ryan peterson is actually printing money if you're getting 40k likes on a post totally definitely making your eight bucks back by being a premium subscriber so shout out to jacobi and tyler and all of our friends at x for making x usage profitable for posters yeah for posters uh let's go to willam he says our children will will romantic the last era of pure human creation in ways we can't imagine. And he posts a screenshot that says, cultural memory has a certain rhythm. Every generation romanticizes the era

roughly 30 years before its own. Gen Z clings to the pixelated naivete of the 2000s. Millennials buff the 1980s and 90s into a yuppie golden age. Gen X yearned for 1960s idealism, but the 2020s will mark a rupture in this condition. rupture in this continuum. When an AI generated song hit the billboard charts in 2023, the prior century collapsed into legend. The last time creativity wasn't a dialogue with silicon. I think that's very well written. Very interesting. Well, lock in brothers because

this show is not a dialogue with silicon in any way. Yeah, it's it's cool because even so So Willem is an investor in a company called Visual Electric that makes the, so far to me, it's an amazing alternative to stock photos because you can just go in there. It's similar to mid-jurney, but it's more oriented around designer use cases. So I've used it to create a bunch of assets. I used it to create an asset for PMF or die. It's super powerful. But ultimately, I don't, I think it's still like the idea that we've been romanticizing every art

form forever. We romanticized, you know, hitting hitting a stick at the right tempo against another stick, right? And we moved sort of up the chart. And ultimately, the power of combining the human mind with, you know, new tools creates beautiful things. So it reminds me of that, I think it was, was it Grimes that was just absolutely dunking on the guy who was doing the AI song generation tool who said, oh, it sucks. learn music, you know, why even learn when you could just generate songs with AI. But ultimately, I think these new tools are just going to introduce more people to the process of creativity. And then as you fall in love with the pure process of creation, you realize that the pure

act of creation is so enjoyable. Totally. And so stimulating that you'll use the tools and then you'll take them to new places that they haven't been, then you'll combine them with your own genius and both things happen. Yeah, study chess. Chess has been solved by computers for decades. It's more popular than ever. People like watching people do things. And this is not going away, even as the computers defeat humans in all of the evals. The entertainment will continue on. Let's move on to David Oolovich over at A16Z. He says, for the 12,589th time, revesting is actually, is usually the founder CEO's friend.

The email below from counsel explaining to CEO why the departing co-founder is basically fully vested. I explained this all to no avail in the series B. And the email says closing the loop on this for the rest of the thread. The revesting concept was proposed early on in the series B negotiations but dropped in the final term sheet, so it was not included in the series B docs. Best. Rough. Best.

Brutal. Clearly some sort of co-founder dispute, somebody leaving, and the vesting, they did not revest, and so there's nothing to negotiate. Yeah, just for some context,

if you go on Stripe Atlas today or you do a standard incorporation with a Goodwin or a Cooley or someone like that, they're going to set the founders up on a four-year vest. Yep. that's four years is a long time, but oftentimes it takes many, many, many, many years to build a

company far beyond four years. And so especially now when a company can go from zero to a series B or even further in a few years, you end up in these situations like David is highlighting where a founder's able to leave after three years and they've fully, you know, almost fully vested their shares. yet there's so much more work to do. And in that situation, they could have the same amount of equity or close to the same amount of equity as the CEO and the team that's going to continue to work on the business. And so, yeah, I think it's always an opportunity to extend out the vesting period.

And it's usually in the benefit of the person who's going to continue working on that business day to day infinitely. Yeah, yeah. And I mean, that's what the negotiation during Series B was clearly framed around was, hey, look, we're going to go bigger. We're raising money. We are going to be in this for another four years, another eight years, decade. So let's revest to make sure this is the team that's like staying on board. And because like refresh grants just aren't going to be anywhere near what the initial founding grant is.

And so, but it can often be seen as a very adversarial between the co-foundary. team and the investors. And so the co-founding teams is like, we don't want to revest. Like, what if you fire us all or what if we leave? Like, then we won't have our equity. We started this company. And it becomes this big debate. But even I think the happy medium here can be just extending the unvested portion at the Series B. And so maybe you're halfway through a four-year vest. You take the second half that has invested and you extend that to a new. four-year vest. So you're not actually losing the shares that you vested on the original schedule. You're just extending the unvested shares. And that's kind of a happy medium that can

work out sometimes. But it's always tricky because, yeah, I mean, you could not be doing your job, get fired. And you're like, ah, I didn't get my full stake, which I wanted. And that would have been awesome. But then, you know, like, what's fair to the team? What's fair to the people that are going to go build this company for the next decade. You know, you got to have equity for those folks. You got to have them aligned. Anyway, we will end it there. Thanks for tuning in.

Thanks for dealing with all the technical difficulties. We are still ironing things out. But that's what comes with being the number one live show in tech. You know, you just got to figure things out sometimes. You got to wrestle with tech. Yeah, you got to. It's not always autonomous. Sometimes it acts up and you got to just figure it out.

Thanks for bearing with us. We're trying to get 100% better every day. A lot of people say, I want to be 1% better every day. It's not enough for us doing our best to hit 100 every single day. And soon we'll be at the top of the charts. I love it. Leave us five-star review, Apple Podcasts, and Spotify, leave an ad in your review and we'll read it on the show. Thanks so much for tuning in and we'll see you tomorrow. Thank you, folks. See you tomorrow. Bye.