TBPN - Starlink Cellphones, Open AI Critique, Ross Ulbricht Founder Story, The Best of X

Episode Date: January 30, 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(02:00) - Starlink Cellphones (11:35) - Open AI Update (34:20) - Ross Ulbricht Founder Story (02:02:45) - Valentine's Day (02:07:24) - The Timeline

Welcome to Technology Brothers, the most profitable podcast in the world. Today we are breaking down Ross Ulbricht. He was released from prison because he was pardoned by Donald Trump. We will give you the full deep dive on his life and career and the drug in a marketplace he built called the Silk Road. Sharing economy startup. Yeah. Another way you could put it that way. Yeah, exactly.

Tons to learn from him. Tons to, you know, avoid mainly all the crimes. Mistakes were made. Mistakes were made. But that should be a fun one. We're going to take you through Ross Ulbricht, get you up to speed on that because he's out in the wild now. Yeah.

He could do anything. Yeah. I think the part means. There was, we weren't involved with the rumor. Yeah. But there had no, there had nothing to do with it. There may have been,

there may have been a rumor or some misinformation shared about him, you know, potentially outraising, but it's just a rumor. Yeah. And so also in, in economy news or Fed world, The Fed. We love the economy here. We love the economy. We love business. The Fed had a meeting today. People were hoping for a rate cut. Ideally to zero. We had. So we had the Dom ready. We picked it up this morning in preparation for a celebratory moment. Yep. And I feel a little bit rugged by Powell, to be honest. Yeah. I mean, I don't think the market was pricing in a cut. I know. But it didn't happen. But there were also rumors. There were rumors that Trump was saying. It sounds smart. Optimus, you know, get paid. Well, maybe next time, but I'm sure there will be a reason to break out the Don pairing on soon. So enjoy.

Leave it on the table. It might be this episode. You know, you never know. Yeah. Surprise. Trump might override it. Override it. Send out a post.

Hey, talk to Powell. We fix the issue. We're going down to 2%. I love that. That would help us recover from Monday's, you know. Monday's blood bath. Blood bath. It's terrible.

Well, let's go to SpaceX, New Jersey. Starlink is coming to iPhones. They're teaming up with Apple and T-Mobile. There's an article in Bloomberg from Mark German, God who covers Apple better than anybody in the game. And he says Apple has been secretly working with SpaceX and T-Mobile to add support for the Starlink Network in its latest iPhone software,

providing an alternative to the company's in-house satellite communication service, which, I don't know if you've used it. It's simultaneously amazing and absolutely terrible. T-Mobile? No, the satellite internet on your... your phone. Oh, yeah. Like, if you get caught with no, yeah, no connection during the fires. Yeah. So you have to go outside. And you're like, it's amazing. It still works, but it's so slow. And you can only send a few text messages. It's like being back in like, being on a plane. I don't know, like

1990s. It's like being on the phone, the phone, like planes now, they offer the free messaging product. And then they have, like the paid version. And they all kind of suck. And it's like, okay, well, you can't receive images, but you can send text back and forth. And so they've been testing iPhones with Starlink service in an under the radar move. The smartphone's latest software update released Monday now supports the technology. The tie-up comes as a surprise. T-Mobile had... So that's interesting.

I didn't realize they fully had the tech. It was just their... The hardware software. Yeah, just the software. Yeah. T-Mobile has previously only specified Starlink as an option for Samsung phones, such as the Z-fold and S-24 models. Do we have any listener?

Yes, we do. We have one. One? Lone Ranger. The Lone Ranger. I'm pretty sure. I'm pretty sure. Yeah.

He said that. Shout out to the Lone Ranger. In a post on X responding to the news, Musk said images, music, and podcasts should be supported by current Starlink technology and future upgrades will add video support as well. So it does seem like when you see the daily podcasts. Yes, fantastic. You'll be able to listen to this anywhere. I mean, there are some like restrictions because you have to be able to see the sky.

You can't really use the satellite receiver in. side. And then also, if you just look at the size of the Starlink dish that I have, I have the mini, and that gets pretty good speeds, but they make a bigger one. As you scale that down, it does seem like you're scaling down the bandwidth and there's some sort of, you know, like kind of like power law function there. An Apple, an Apple spokesperson declined to comment while T-Mobile said the test. Yeah, Mark is really just throwing this news out there without he has so many moles and sources inside Apple. Clearly they weren't exactly. ready to announce all of this.

But Elon's no stranger to it. He just reacts. Yeah. Yeah, yeah. Text messages are already going out from T-Mobile saying you're in the T-Mobile Starlink beta. You can now stay connected with texting via satellite from virtually anywhere. So I think Apple...

It's interesting, too, because I never followed this closely, but I always imagine that Starlink would eventually come out with their own phone plan and actually threaten the telecom providers. Yep. But maybe telecom companies. just find a way, you know, to really, I mean, I think it really comes down to, the cell phone is tied to service still.

Yeah, and I think it comes down to like, reliability. Yeah, yeah, and using it in buildings. Like, you would say, like, why would you pay for that plan? If it doesn't work in a building, that'd be weird. And then I guess they could team up in kind of like white label, like, you know, like Mint Mobile that Ryan Reynolds company? Yeah, yeah, yeah.

Like that is, they don't have their own towers. They piggyback on the, they roam on the other towers. And so I'm sure there's some way. to build something. But Apple was originally working with Global Star to offer satellite features in 2022. There's been speculation over whether it might support competing networks. The day after Apple's 22 announcement, Musk posted on X that his company had some promising conversations with Apple about Starlink connectivity.

It's so funny how he just like rips like these private conversations on X. He's just like, oh yeah. It just walked out of Apple. Like we're definitely getting a deal done. Like, yeah. It's in the bag. Do not do that if you're not Elon Musk. Yeah, exactly.

So when a T-Mobile phone is in an area without cellular connectivity, devices that are part of the Starlink program will first try to pair with Starlink satellites. Users will also be able to activate texting via the satellite menu for the Global Star Service or contact emergency services through Apple. The initial version of Starlink is exclusively for testing, but Starlink and with SpaceX and T-Mobile have said that they plan to expand into data connections and voice calls in the future. The program is also only available in the U.S. best for now. And that contrasts with Apple's Global Star Service, which works in several countries. So when I am getting the opportunity to use satellite on my iPhone, it's still Global Star. Yes, exactly. And so Global Star shares fell because people are thinking, okay, they clearly aren't on the level of Starlink and Elon's going to make a bunch of deals happen that directly

threaten their business. Yeah. And so the history of like satellite communications for cell phone and internet, that's been going on for, I mean, 30, 40 years. It's also crazy. Think of the cost structure for Starlink because they own the rockets versus Global Star, who presumably has to use SpaceX to get their satellites up. That's not a great competitive dynamic for global stars. Yeah, a lot of this started with, you know, the guy who built the, he owns Madison Square Garden, he built the sphere in Las Vegas, James Dolan.

So James Dolan at one point was working with Madison Square Garden on like all their sports and entertainment thing. And he had a satellite TV startup that didn't, that he was like obsessed with. He was beefing with his dad about it. It was this crazy story. They wind up selling it, I think, to DISH networks. And DISH is the traditional like satellite TV provider. So you put the.

satellite dish on your, on your roof, and then you're getting a constant stream of TV. So it's not two-way communication, it's not bespoke. So that satellite, all it has to do is just say, okay, there's 100 channels of data. I just got to send those out to everyone all the time. Whereas Starlink's much more complicated because it has to send you, your, you know, Instagram, your specific Instagram images. When I'm using deep seek. Yeah, yeah. It has to send like your specific data, right? Internet's more complicated. And so then there were sat phones, satellite phones. and then Viasat was a company that was like really promising for a while, putting up big, like huge satellites.

Viasat still does a lot of stuff with the airlines, right? Yep, yep, exactly. And so, and then Astronis is also doing geostationary satellite internet. So they will put one up over Alaska and then that will just control just Alaska, but it's just one. Starlink has this mesh strategy, so they put up like thousands and thousands of them. And that's, of course, like, driven by the fact that they have, you know, all these rockets that are going up all the time. So they can afford to just put up tons.

And then they're also in low Earth orbit as opposed to geostationary orbit. So they're, I think, like, it's like 3,000 miles instead of 30,000 miles or something like that. So much closer to the Earth. And that enables lower latency. So instead of, like, you could sometimes with satellite internet or satellite TV, you could download, like, a big file, but there would be latency. So you could never do, like, a zoom call. or play a live game, and that's the advantage of Starlink. And so currently, if you're using an iPhone with Global Star,

you have to point your phone at the sky to find the satellite, whereas with Starlink, it's designed to work automatically, even when the phone is in a customer's pocket. So maybe it'll work inside to some degree. I'm not exactly sure. But both the Starlink and Apple satellites features are designed to work in off-the-grid areas, such as hiking trails that don't have cellular service, the capabilities can't be used in places

where a mobile phone network is within reach. So it will just say, okay, we're not even bothering with the satellites if you're connecting to a cell phone tower. And the support for the Apple feature is available on most current iPhone models and the company plans to bring it to its ultra smart watch later this year.

That's cool. So like if you go hiking, you just bring your watch, you don't even need the phone. If something happens, you see a bear, you can still send a text message saying, come rescue me. Glock's coming out. Yeah, maybe you don't need the Starlink connected Apple Watch. You just need the Glock.

T-Mobile updated its website this week. SpaceX requested authority to begin beta testing the service starting Monday. The FCC commissioned granted SpaceX conditional approval for its satellites to supplement T-Mobile cellular network in November. So very exciting to see more Starlink news, more Starlink rolling out. Yeah, it turns out owning the heavens as SpaceLink. does is quite valuable. It's free real estate. You can do a lot of stuff with it.

Yeah. There's a lot of alpha in that. I mean, that's the whole, that's the whole like mega bull thesis on the boring company. It's just like there's only so much land under the earth. And like if you can justify a way to like own it, like that's extremely valuable. Yeah. I love that. Yeah.

And we've talked before about land yachts and bean airs, you know, beaners looking to develop large drills attached to vehicles that allow them. them to burrow deep beneath the earth and pop up in Saint-Trope, Borschavelle, Aspen, Arbazole, you know, like being able to just pop up pop up is a great feature and something that is in high demand. Fantastic. Should we talk a little bit about AI? Yeah, we love talking about organic intelligence and artificial intelligence.

Today we have an article here. I wanted to cover some news on OpenAI. for those that don't know, this is an American company that competes with a Chinese company called Deepseek. Do you want to... Yeah, people call it like the American Deep Seek, basically.

Yeah, it's the American Deep Seek. The American Deep Seek. Very similar to chat GPT except it struggles to talk about Tiananmen Square. Yeah. But of course, if you've been listening to the show, you've seen, you know, you've heard many deep dives. We've probably spent three hours

talking about Deep Seek at this point. But Ben Thompson dropped a new update on Stratory and we're going to go through it. So to review his conclusion on OpenAI, he has some criticism of the company specifically around speed of innovation, which is interesting. And his thesis is basically like all the crazy nonprofit,

effective altruism, all the fear mongering around AI doomerism, that really held us back. And this is what we get for that. Meanwhile, like, no one is causing an AI panic over deep seek. People are like, oh, cool, they use floating point eight instead of FP32. Like, it's like, oh, it's so much more efficient. Like, it's going to be able to, you know, like, review my, you know, homework faster.

And that's it. It's not, like, everyone's saying it's a incredible jump. No one's saying it's more dangerous now. And that's just a complete vibe shift from two years ago when everything was, I was reading my son, the boy who cried wolf. and nothing, like, the AI Dumer's really need to read that story. Because they've been like, GPT1, this is going to kill us all. GPT2, we're done for.

GPT3 is completely over. It's never been more over. And now we're getting a new AI model every couple months. It's like, yeah, it could kill me, but I'm not going to listen to you. Now people widely agree that we have AGI. Yeah. And it's ASI that we need, artificial super intelligence. Yeah.

which so the goalposts keep moving. And so let's read from Ben's, let's read from Ben's previous day conclusion. He says, that leaves America in a choice we have to make. We could, for very logical reasons, double down on defensive measures, like massively expanding the chip ban and imposing a permission-based regulatory regime

on chips and semiconductor equipment that mirrors the EU's approach to tech. Alternatively, we could realize that we have real competition and actually give ourselves permission to compete. Stop wringing our hands, stop campaigning for regulations. Indeed, go the other way and cut out all the cruff in our companies that has nothing to do with winning. If we choose to compete, we can still win. And if we do,

we will have a Chinese company to thank. I like that because he's just so focused on just like, go, go, go, competition and just actually win there. So some folks read this as being simply about government regulations. And I did specifically mention the president Biden executive order on AI, which Trump has already repealed. But that misses the deeper. point I was diving at, one of these intense frustrations of the last few years of the AI debate has been the Motten Bailey aspect of the safety question. And this is funny because there's this AI Dumer guy on X who always uses the Motten Bailey like meme whenever someone's arguing with him. He's like, no, that's the Motten Bailey. And now Ben Thompson's kind of like flipping it around.

And so we'll break down like what that is because it's actually kind of odd. But I understand and grant the very real concerns about AI safety, which I think everyone. does. There is some safety concerns, but that's the MOT in the safety argument. But the way that AI safety has too often manifested has been through the imposition of various cultural mores that roughly align with San Francisco politics. That's the Bailey. Indeed, if you look again at the excerpt from OpenAI's GPT2 announcement that I linked to yesterday, it's right there in the first sentence due to concerns about large language models being used to generate deceptive, biased or abusive language at scale, we are only releasing a much smaller version of GPT2 along with

sampling code. We are not releasing the data set training code or GPT2 model weights. So this is the moment that OpenAI decided to stop open sourcing everything. And the reason was we believe that our release strategy limits the initial set of organizations who may choose to do this and gives the AI community more time. to have a discussion about the implications. So they were worried about generating deceptive, biased, or abusive language. And then Ben says, first off, deceptive biased or abusive language is not a human extinction

level danger. So you've said, hey, we got to lock this down because it's going to paperclip us. And historically, we've dealt with bots trying to influence public opinion at scale. Yeah. Yeah, they're not. Are we good, Ben? Cool. Cut for a second.

But historically, it hasn't been this intelligent bot who's having a conversation. It's more these bots that are spamming comments being like, I think America's bad. Yeah, yeah, yeah. It's literally just copy pasting. That's deceptive and biased. Totally. You know, doing the same things. I think one thing that Open AI has always balanced is clearly they want to win.

Yeah. Sam enjoys the competition, but loves winning. He said, he even said on television, I think about a year ago, this point nobody could ever compete with us like he's on the record being you know saying i don't know nobody is uh i i saw it was it was shown on like entertainment tonight like some random television but it was from from some event he was at so he said nobody can compete with us that's obviously not true they also have used this pending looming doom or creating this all power intelligence as a marketing and capital raising strategy to basically create this aura around the company of we are creating

intelligence. We're creating this powerful being that's more powerful than humanity. And you have to trust us to control it. So historically, I think some of this stuff due to concerns about, you know, large language models playing into that narrative, but then also saying, well, we have the power. Yeah. I do, I do wonder, like, there are so many different people around the table. Like, a lot of this language was clearly driven by the previous Open AI administration essentially, like the nonprofit board. Some of those people were really, really crazy. And for a while, it felt like, even though Eliezer Utikowski wasn't tied to Open AI, it felt like he was almost like an informal puppet master. It felt like he was like an informal spokesperson for a while. And now it's like he's

completely out of the picture. But it was like, oh, if Eliezer says something like, Open AI is going to have to respond because it's really critical that we like take him seriously. And now it's been kind of, you know, we've moved past that. And so second, who decides what is deceptive, biased or abusive? And third, whatever your position on these questions, I think we can all agree that a tremendous amount of time and effort has been devoted to arguing about these cultural war issues and implementing one side or the other of them. By the, and by definition, that time has an opportunity cost. More broadly, I think it is completely valid to argue. you that a lot of time and energy and tech has been devoted to many things other than actual

tech. And that's the justification for that lack of focus has been rooted. I would actually be interested how many tens of millions or potentially hundreds of millions of dollars are spent on their safety team annually. Like it could easily be somewhere in the range of 50 to 100 million that's just being spent on people that are a part of the safety apparatus, which Ben is basically putting in the culture war bucket. Yeah. Like almost classifying them as deep, like, making them like a DEI org.

Yeah, what's interesting is like, so he goes on to talk about the 01 chain of thought and how they, when they launched O1, in comparison to R1, R1 tells you everything it's thinking and just dumps it out and it's open source. So even if it didn't, you could still see every single token that it's thinking about. And that's turned out to be a great UI pattern.

We've talked about this. It's just really reassuring to the user. But when OpenAI launched 01, they said they believe that a hidden chain of thought presents a unique opportunity for monitoring models, assuming it's faithful and legible. The hidden chain of thought allows us to read the mind of the model and understand its thought process. And this was the thing where I went to 01, asked it to summarize a book. and one of the hidden chain of thoughts sections was copyright infringement. So it thought internally about like, is it okay for me to share the contents of this book?

And it decided, yes. And so there's a little bit of a bad boy. Like, there's that. Ross. Yeah. And then there's also the fact that, so they were also thinking about hiding the chain of thought so that people couldn't steal it and use it to train their model basically. and if they have some interesting thing in there,

they can kind of tuck that behind the scenes and just give you a better answer. But this exposed chain of thought from R1 has clearly won out on the day, and I think we'll see other modelship. Yeah, users like it. That was the only UI change that people cared about or talked about

with the Deep Seek launch was that exposed chain of thought. Everything was either a copy, like even just the fact that you have like the chats on the side, like where the buttons are laid out is a verbatim copy of chat GPT. Yeah, down to the API. Yeah. The way the API works. Many of them are actually degraded features. Like you can't, you can't click a button just to talk to it. There's no voice mode at all. It doesn't have a lot of features yet. But exposed chain of thought is like the fundamental UI pattern breakthrough. And so exposed chain of thought teaches users how to do better

prompting because they can see exactly where the AI gets confused. It increases trust in the model because you gain an understanding of how it's arriving at the answer. And it's just cool and endearing. One thing that was interesting with the normy discovery of R1 is how delightful they found the experience. And so Ben understands OpenAI's competitive concerns, but the existence of R1 has shown that locking down chain of thought perhaps didn't matter so much. More generally, the consideration of competitive concerns gets back to my overall critique that the U.S. approach to competition is too rooted in stopping and blocking would-be competitors instead of out-innovating them. Yeah. And this was the thing where, like, remember when we went back to that Ilya working

on A-Star, and the framing around that was absolute panic. I remember there were fake news articles about Open AI. This was almost two years ago, we said, right? It was a year and a half ago, or maybe a year and three months ago. And they said, A-star is so good. It solved every foundational like math problem, and it also broke encryption. Yeah. And so they were like, they have an AI model that can break encryption. So they're just going to like, you know, Bitcoin should go to zero because they broke an encryption. And it was like obviously not true, but people were just so insane about it. And internally, it seemed like a lot of people were maybe just using it for internal politics and like, you know, oh, is this dangerous? Like, why are we moving so fast? The nonprofit folks having a big problem with it. And then O1 just throws it or R1 just comes out and throws it out there. And it's like, yeah, it's basically fine. Yeah. Very funny. Yeah. It's kind of the, it's kind of the most savage founders that I know, well, when they're fundraising, they'll take a PDF,

export of their deck and they'll share it with anybody that's like remotely like interested in what they're doing or relevant in some way. And then the the founders that are sort of afraid of competition, you know, kind of lock their decks down. They put emails, you know, authentication and passwords on their deck. And then they usually end up just like not being the actual like winner in the category. And so it's the same thing here where we should just be really focused on innovation, winning innovation, being at the edge. understanding that people are going to reverse engineer everything good that gets done in the world. That's been the case for every technology.

And things will get copied one way or another. And this was kind of the weird critique that I saw once the Sam Altman like haters got really emboldened. They would make the claim that Sam Altman is like non-technical and has never done anything good and is not responsible for any innovation in AI. Like the transformer was invented at Google. He didn't invent the transformer. He just like popularized it. But then simultaneously they'll argue that him, they'll quote the Open AI board

that criticized Sam for launching ChatGBTGPT too quickly without informing the board. Do you remember this whole thing? It was like, as evidence of Sam like moving recklessly with regard to AI safety, he just put out, he just tweeted out chat GPT one day. And it's like, well, which is it? Like he can't be both.

He can't both have no impact and also create the most dominant, product and drive it forward and actually get it out in the world. If you're thinking about Open AI as a business, him tweeting it out and getting to 100 million users in however it was weeks. It's fantastic. It's, you know, clearly was the right decision to just go. Yeah. If you're thinking about it from a nonprofit safety org lens, you should have waited a year, you should have tested more. Should have put more guardrails up. You should have rolled it out slowly. There's a number of things that. Yeah. And so I kind of see it as as like there is a world where,

Sam was fighting like an internal, like the nonprofit structure, culture, culture war, like a long house type environment that was saying like, no, don't release products, don't move fast, don't innovate, don't do anything that could be risky because they're so terrified of what they're building. And he's looking at it and he's like, actually, this is fine. It's chatbot. Let's just get it out there and see what happens. And that was the right decision. But there's still like, there's still that impulse internally. to say, hey, maybe we should lock this down a little bit more. And so there's an interesting discussion right now.

Is this a Sputnik moment or a Google moment? And so Ishaan Wong, who is, I think the CEO of Reddit, says Sputniks showed that the Soviets could do something the U.S. couldn't, a new fearsome power. They didn't subsequently publish all the technical details in half the blueprints. They only showed that it could be done. With DeepSeek, if I recall correctly, a lab in Berkeley read their paper and duplicated the claimed results on a small scale within a day.

That's why I say this is like the Google moment in 2004. Google filed its S1 in 2004 and revealed to the world that they had built the largest supercomputer cluster by using distributed algorithms to network together commodity computers at the best performance per dollar on the cost curve. And famously, Google's was massively profitable when they went public. And every VC is like, why can't we have more Googles? Because the financials were so fantastic. This was in contrast to every other tech company of the day during the dot-com boom, who at that point just bought what were essentially larger and larger mainframes, always at the most expensive leading edge of the cost curve.

And so Ben says it's hard to overstate what a favor deep seek did to basically everyone by outsource, by open sourcing their model and explaining exactly how they accomplished what they did. They could have, and from a geopolitical perspective, arguably should have kept all their progress a secret, even as China built drastically more. efficient AI infrastructure. You know the narrative that they are a hedge fund. This was a side project. And so they just released it, open a massive short against a video. That's a funny one.

Yeah. And then made, you know, they could have made how many tens of billions of dollars off of that short, right? There are so many theories here. Like,

it is a fun theory. Yeah. And there is a world where that was exactly the strategy is like they were doing, they were building models internally. Yeah. To run their investing. And then they realized, hey, we can train a model and we can throw up a clone of chat GPT.

And if we pay to get it number one in the app store, we can tank the markets on Monday and like get, you know, make like an entire career's worth of, uh, it's definitely a tinfoil hat theory, but I, but I'm interested. Yeah, yeah. I'm listening. Yeah, yeah.

I'm not saying it. I mean, Palmer, Palmer just cited that in a post that we'll go through later. And, uh, and it's not as crazy as it sounds. There were some people pushing back and he had some very good responses to those things. So then he moves to a tweet from Roon who says, this is not the right lesson.

There is an opportunity cost to research your time and there. And if they're spending it micro-optimizing PtX to make the best of limited comms, like the memory communication bandwidth, that's something else they weren't doing. Nerfing some aspect of the hardware is never a benefit. The most dangerous, and then Ben says, the most dangerous lure about thinking in terms of comparative advantages, how time and learning curves lead you to consequences. that you never considered when you made the decision. I wrote about this exactly last year in a chance to build.

I started the post by describing how U.S. semiconductor companies kicked off outsourcing to Asia and tied that to the fact that Apple can't make an iPhone in the U.S. And so Steve Jobs famously said, you know, it's not even a matter of price. It's just there are not 30,000 manufacturing engineers in the United States. There are not 30,000 manufacturing, engineering jobs to be filled. This is because the structure of the world economy, choices made starting with, in woods in particular and cemented by the removal of tariffs over time made them non-viable. And so connect this to Google.

What if Google had taken Roon's advice and not wasted time, what not wasted the time of valuable researchers in transforming how data centers are built? Perhaps they could have waited until Alibaba did it a decade later and rented capacity. And in the meantime, they would have been a far smaller and less significant company specifically. and the entire U.S. tech industry would have been similarly stunted generally. To be fair to Rune, he does seem to be talking narrowly about the specific optimizations that DeepSeek made to overcome the H-800's Nerf memory bandwidth, but I think the critique holds more generally.

U.S. tech companies, including Open AI, have clearly in retrospect been NVIDIA's willing suckers happy to pay for more capacity instead of wringing out their own. How many years would it have taken them to discover deep-seeked breakthroughs? Yeah, and this is a classic if you're trying to scale anything really quickly do anything really fast, you just end up paying more for it. So you hire somebody and you just say, yeah, we'll pay you, you know, $20,000 a year more than we budgeted for the role because we need the role. Or you're trying to buy billboards quickly and you say, yeah, we're going to pay 30% premium,

but we want to launch this now because there's this competitive pressure. So what these U.S. firms have been doing is not exactly out of the ordinary. It's just the nature of trying to do things fast as you focused on, you just focused on, you just focus on speed versus efficiency and cost. Yep, yep. And so he closes talking about aggregation theory, says, on the flip side, non-reasoning LLMs are still great at producing content, and they also get smaller by distilling the output of reasoning models. One of the big R1 takeaways is that you can infuse reasoning into normal LLMs. More than that, though, you can also use them to generate massive amounts of synthetic data that simply teach the normal LLMs the answer to an effective

infinitely infinite number of questions. And of course, reasoning could end up simply being effectively free. One announcement that I should have added to yesterday's post was Google's announcement of Gemini 2.0 flash thinking, that it appears to be competitive with R1. The China is crushing U.S. rhetoric today forgets about Gemini 2.0 flash thinking, likely cheaper, longer context, and better on reasoning. We're still early in the AI race. So this was like completely missed.

Yeah. But we don't know the prices yet. I believe that Google owns search and YouTube and they basically have every mainstream media company by the, by the balls. Yeah. And they can't even get news out. No, it is, it is crazy. Their product sense is just so bad. And they have this whole problem.

And their marketing sense now. It's like, is it Gemini or is it Bard or they've been renaming models left and right? And it's clearly they're shipping their org chart because they had, they had Google brain and then deep mind and they finally merge them. But they're still just like, like, I know what Google is. Google is a search engine. I don't know what, I know what chat GPT is. ChatGPT is a chat bot.

Like Google Gemini 2.0 flash thinking, like what, how do I even access that? I tried to get Gemini set up. I finally figured it out. It sounds like an internal project. I went to Gemini and then I had to upgrade to Gemini advanced. And then I got access to the 1.5 model.

But I was even asking the model like, what are your capabilities? And it didn't know. Like it was like not telling me like how to do it. That should be one of the new bent. benchmark for models. Yeah. What's your context window?

What version am I on? How do I upgrade? Take my money. Interviewing for a role and somebody asks. So what are you good at? What do you do? You don't have a good answer for it? That's kind of a tough self.

Yeah. And so Google is a company that actually cares about and has invested in infrastructure and perhaps doesn't see the need for software type of margins. Open AI and Anthropic have allegedly been earning on their API. So I wouldn't be surprised if they are very competitive. Of course, they're not open source, but again, R1 is I can get all the thinking I want locally for free. This isn't necessarily dispositive in terms of the long-term prospects for aggregation theory,

but given that the entire history of technology has been defined by processing speed simultaneously increasing, even as it decreases in price, the burden of proof is probably on those who say that zero marginal costs are an aberration. So he's saying that, like, in the long term, all of these models will be optimized to the point where there's zero marginal cost. And at that point, being the entry point and being able to tax that is really valuable. And so there's a couple posts that we'll go into during the timeline that say, oh, this is massively bullish for Apple. Because all of a sudden the inference gets super cheap, gets much better, can be done on the edge, bake down.

They don't need to mess around with any of the actual training or inference themselves. They can just throw it on the phone and it's there and it's great. And same with Google and Facebook. they were saying it's really good for Facebook because they can just chuck this in and they don't need to worry about, you know, playing the model race. Yeah. But fascinating story. It continues to evolve. Ben Thompson staying competitive.

Goated. Goated. Goated. Let's move on to Ross Ulbricht and the story of the Silk Road. We've had a lot of laughs. Oh, yeah. Obviously, it's a dark story in many ways. Yep.

There's sort of this bright spot now. with him being pardoned. We're not, we'll wait till the end to talk about whether or not we agree with that decision. Not that it's on our, you know, not that we like to comment on politics. But yeah, there's, there's so many just insane aspects of this story. Yeah. And I think that as we looked into it, we realized that so many people had talked about the history and the sort of crime story around Silk Road because it's so fascinating.

It's this internet narco trafficker meets hacker meets Anon who, you know, was an Eagle Scout, right? So it's just like wild story. Yeah, we should. And an amazing marketer. Oh, yeah. Even after he still has a ton of fans. I don't know if Ben, you can pull this, this photo up, but he was an Eagle Scout. I was an Eagle Scout.

Really? Yeah. I didn't know that. My God was an Eagle Scout. And I remember when I was in the Boy Scouts, they would tell everyone, no, Eagle Scout has ever been convicted of a felony or something like that. It was like clearly wrong, but it was just to like put you in the mindset of like, well, I don't want to be the first one, so I can't be a criminal. Yeah. And it was very effective. The Boy Scouts and the Girl Scouts fell off

massively. There are tons of like culture war issues. Yeah, yeah. And real, you know, like, yeah, criminal stuff. Crimes and stuff like that. But anyway, so. But he has a lot of fans and dripped out technology brothers like really, really likes him. And has posted him multiple times there's another photo of him uh yeah we'll put that up on the screen uh of him and without a doubt he's he's benefited massively from his you know aesthetics aesthetics yeah the physiognomy is which is why it's underpriced to get diced i don't know that he's diced i don't know that he's actually lifting a lot of people say he looks healthier than brian johnson after 11 years in prison eating prison slop yeah what's the secret what's the secret yeah we'll have to

ask him. He does look young and he certainly played into that and we'll go into kind of the timeline here because a lot of people think, oh yeah, he was 17 when he did this. He, wrong. So he, yeah, the other aspect that we'll talk about throughout the story is, is how good of a marketer he was. And he and his family post incarceration on his carceral journey. Yeah. They were able to spend this story like he was like some 18 year old kid, you know, high schooler almost. What you'll find out is that he was, you know, 11 years ago. And, uh, they were years ago, he was, when he was put into prison, he was 29 years old. Yeah. So he wasn't, you know, he wasn't a kid. No, not at all. He's born in 1984. And he was very aware of his

action. So just to get into kind of like the high level founder mode elements of the story, and then we'll get into the timeline. He's a self-taught developer, first time CEO, completely bootstrapped the company, went from zero to 1.2 billion in sales in less than three years, personally netted over 80 million in fees off the marketplace with a very low take rate, which we'll also get into. Operating a fully remote, global, and basically anonymous team where he was the only one that knew the real identity of his employees, he would require them to show ID, leveraged novel payment rails using Bitcoin and maintain this sort of missionary sort of

mode to building the company that was basically almost like a religious pursuit for him because of his libertarian beliefs, which we also get into. And so the backbone of our research was mostly Nick Bilton's American Kingpin, which is a great book. I highly recommend it. It reads like, you know, it's a page turner. It's like a total thriller. And so, uh, Billton kicks it off by taking you back to the beginning of Ross Albrecht's story, born in Austin, Texas, March 27th, 1984, supportive middle class household, nothing crazy going on.

Eagle Scout, as we mentioned. Friends and family described him as polite, gentile, and unassuming. Ulbricht developed an interest in science, tinkering with computers. He goes to University of Texas at Dallas on scholarship, initially studying physics. Then he pursues a master's degree in material science at Penn State. And he basically gets obsessed with libertarian economic theory, especially in Woodwig von Mises, Murray, Rothbard, and Milton Friedman. Yeah, and so while he's there, so one thing to know,

note he's not studying computer science he's not a programmer at all at this point he's interested in technology but he's spending his time going and debating yeah libertarian beliefs arguing for things like the legalization of drugs his big his big point you know the high level sort of his high level view is that government should not have any decision-making power over individuals and individuals should be able to do whatever they want and he goes on to do whatever he wants yeah For the, you know, basically. Have you seen that post? Did you go on Banger Archive?

But it's like, women will say they want a man who reads, but every man who reads ascribes to some ho-scaring, extremist political belief. And it's like, yeah, he got into reading and started reading like Mises and Rothbard and like these like kind of crazy philosophers. And he really... Speaking of Ludwig von Mises, there's this UFC fighter Money Moikano. Yeah. Oh, yeah.

done this post-fight speech where he just is like hammering against the government. Yeah. And I mean obviously there's a lot of good ideas in libertarianism, but it clearly like Ross took it to such an extreme that he was using it to justify anything. And clearly bending his like libertarians love to,

they have a hammer and everything looks like a nail. And so they wind up not seeing any externalities just viewing every little minor interaction and is like, oh, well, this fits with my libertarian ideology, so therefore it's justified. Yeah. And he really, really twisted the ideology in, like, a much more negative way than many libertarians who are just arguing for, like, slightly smaller government or, like, you know,

slightly, you know, lower taxes or something or more efficiency in the government. Like, you could look at Doge as, like, a libertarian project in some ways, but it's much different than, yeah, we should legalize guns and drugs and all this crazy stuff, which is, like, where he wound up. And so. Yeah, not just legalizing it, but brokering transatlantic. transactions at scale for a number of illicit goods. He dabbled in entrepreneurial ventures after graduate school.

One of these involved day trading. Another was an e-book enterprise. None proved lucrative or fulfilling. So to be clear, he was trying to do day trading, realized it wasn't very profitable. He also, he was a part of this someone else's sort of nonprofit called Goodwagon Books, which they would go around a houses, collect sort of old books if people didn't want. they would sell the ones online that they could and they would donate to prisons the ones that they couldn't sell wow and so he was probably foreshadowing donating books to himself to himself in his later years

and uh yeah one one other thing about this time that i thought was interesting while at penn state a little like odd jobs he was working he saved up enough to buy a rental house and he later shared that that he found it so frustrating having college students as as you know did generate college students as tenants that he sold it. Yeah. And he actually used that money to kind of get Silk Road off the ground because he didn't have an income for some amount of time. Yeah.

So he becomes fascinated with Tor, the Onion router, which is the portal to the dark web. The timing here is so key. Yeah. It's like if he had had all these ideas five years before, impossible. It wouldn't have had Bitcoin. I don't know where Tor was in the development, but it wasn't. All this stuff was like this perfect convergence.

Totally. and Bitcoin obviously coming up at this time. Few people outside of cryptography circles were paying attention to Bitcoin, but Ross saw it as a tool for circumventing government oversight. He once famously stated, I want to use economic theory as a means to abolish the use of coercion. This quote illustrates Ross's self-styled mission of empowering voluntary exchanges. And he, yeah, it's also very funny throughout this because he's not like super technical. there's a lot of times when clearly you can tell that he's he's bought you know like oh nothing can go wrong it's decentralized like oh bitcoin it's anonymous and it's like it's not it's actually not

anonymous like in many ways and there's always ways around this stuff you don't have to um but the what's the word for you don't have to k yc yeah but but it's it's a public ledger exactly exactly so it's very very traceable and uh and we'll go into this later but uh they're like tracking the bitcoin movements actually got a lot more people caught because it was all tracking And so- And there's still companies today, like chain alias, that leverage public blockchains to track down criminals at scale. Yeah.

And so combining Tor with Bitcoin and his ideological calling, he sets up the first version of the Silk Road. By early 2011, Silk Road was online, though still small. And his site allowed buyers and sellers to communicate post-product listing and leave reviews. This is basically an eBay. Yeah. And so just going back a little bit, taking one quick break. So he to anytime you're starting a marketplace. So the fact that he took this marketplace

from zero to 1.2 billion in volume in three years bootstrapped is insane. Very, very, very, very few companies do that. He, in order to get this started, marketplaces have this chicken and egg problem where what comes first, the supply, you know, you need supply in order to have demand on the marketplace. And so he actually starts growing mushrooms in order to create the supply to have supply on on the marketplace so he's growing mushrooms he's building the site and then once he's at a point where he's ready and he has a girlfriend at this time who's like sounds like a very sweet girl and they're very like in love at the time yep and he at one point like takes his girlfriend like blindfolded out to like see the mushrooms and like he's like all proud of it like takes her out

just like damp kind of feel or whatever and there's just and so he's he's uh he's this is like the very beginning of him actually, so he had the intention to commit crimes. Yeah. And he's very aware that it's what he's doing is illegal. And he specifically chose psilocybin mushrooms because he felt like the penalty for that would be less severe than starting by selling harder drugs. Or even something like marijuana at the time. Yeah, but I thought the amount that he was growing was, it was, it wasn't like he was growing like, oh, he's got like one little plant. It was like a whole roomful. And the punishment for that if he had gotten caught would have been years and years in prison. just for that, which is like still kind of good. Yeah, and one thing is in order to kind of kickstart

the first initial sales, he, uh, he went on a site that would talk a for a public forum. Yep. That would talk about mushrooms. He made a post and he was very sort of, this was smart at the time if you weren't committing crimes, but he, he posted something under this username. I forget what the username was called. Was it Altoid? It was Altoid. Yeah. It was like, hey, has anybody tried this site. Yeah, looks pretty cool. Looks cool. And he used that to sort of catalyze some traction because he knew he was holding supply

that he was ready to sell. Yep. And so we'll get to this later, but that ended up being part of his downfall because he had used his real email to create the Altoid account. Yep. And he had gone back at one point and changed the email, but there was a record in the database that it had Ross Ulbricht at gmail.com tied to that username. So that was one of the ways that they sort of ended up.

Yeah, he's one of those things where he's, oh, everything's anonymized and decentralized. It's like, you left paper trail all over the place, bro. Like, you were not that good at your opsec. Yeah. And so it's fascinating how fast this took off. So he's thinking about it in late 2010. In January of 2011, he launches the Silk Road, code named underground brokers in Austin, Texas, aiming to create a free market website where users can buy anything anonymously.

And by June, just six months later, Senator Chuck Schencher, Schumer denounces Silk Road in Washington, D.C., prompting heightened media coverage that forces Ross to Bolter. They also got a big Gawker article, which I think was right around this time. Yeah, this all happened right around the same time. So a reporter who is at Gawker and Gocker's no longer with us due to the actions of some powerful people in tech. But this Gawker article, hears about Silk Road, gets on it and is browsing around, ends up writing this whole piece. Ross talks to them and says, please don't write this piece. That usually doesn't work with journalists.

They know they're on to something. They end up publishing this. And there was even a while that that journalist was suspected of starting Silk Road because it was such a big marketing moment that people were like, did he just like, did he like use this? And he'd written about some other topics. And that was like Ross really felt like it was too early for it to go that public. Like they were starting to do sales in a meaningful way. But at the same time, he was a self-taught

developer. And so he struggled with hacks and stuff like that all the time. And so he would notice every now and then he'd be like, oh, what the heck? Like my wallet's getting drained. And so he'd be freaking out. The Bitcoin at the time, to be clear, we should look up on public what the actual price was. But it would be like, I'm sure it was like a few dollars. It was a few dollars. Yeah. And so you'd be noticing like, oh, my wallet's getting drained. It wasn't. that damaging, but he kept having these sort of small errors because he was not a super sophisticated. He was like a hacker type builder, but not a security engineer. And so he kept constantly getting kind of lightly scammed. And so this launch, like you can imagine,

he's just in Austin building this thing by himself. He had a couple engineers that would help him that knew they were committing crimes too, but they were they were on board with it. Yeah. And then going zero to 60 in six months, zero to 100, zero. to a thousand, right? Like he's getting a senator in Washington, D.C. that's denouncing you is, the pressure was on right away. Yeah, usually these, these like, you know, government federal level pressures take like years until like the first like, oh, is Facebook, you know, leading to misinformation that took like, you know, years and years or even like the jewel hearings. Probably almost 10 years. Yeah, the jewel hearings were probably, I mean, that company was started like

10 years before. And there was being publicly marketed the whole time. Yeah, totally, totally. Totally. Had the Gawker article and some other coverage, but it wasn't. Yeah, I mean, Chuck Schumer had the Zinn press conference in 2024. Zinn launched in America 2014. And so the dark web Amazon, the Silk Road really like just took off so fast six months. And to be clear, the reasons why it was in many ways a much better looking purely at the user experience.

If somebody that's buying drugs in 2011, while this starts. They can go and find a dealer on the street or hit up some random sketchy person on Facebook. And there's all these ways to have a transaction that don't have a bunch of reviews, right? So you're kind of trusting the dealer, which is like people that sell drugs.

I can imagine you wouldn't want to trust them with your life. And so the experience at the time was from a purely like a product experience to be able to go on a site to see vetted sellers, that have a selling history, an activity on the forum and reviews, he really made the experience of buying these narcotics much better. And people were allowed to buy very small amounts. And this ended up being what slowly helped get the marketplace taken down is usually authorities are like Homeland Security, DEA, FBI, when they're tracking drugs, they're looking for these big shipments.

Yeah. And as Silk Road started taking off, they started getting these envelopes that would be, you know, one little pill. So like one person who just like goes on there, they just order one. Yeah. It's kind of cute. It's kind of cute. But, uh, but yeah, it was, it was it, it, it can't be understated, like going back

to the founder story bit, he made the experience, like he really innovated on the experience. Yeah. Yeah. Yeah. There's a couple other things that were pretty innovative here. Uh, they had an. an escrow system, which used Bitcoin kind of smart contracts, but really just it would transfer the Bitcoin out of the buyer's account into a holding account. And then once the transaction

cleared, it would get paid to the seller. Yeah. And so like if you got scammed, you could actually have like a dispute resolution process. It was like a PayPal. Exactly. Type experience. Yeah. And PayPal really unlocked either. And to be clear, so that it wasn't fully trustless, right? No, no, no. Like somebody could have received it. Centralized. And we saw that because everyone was stealing Bitcoin from the site constantly. Well, yeah, and Ross would also have to get involved with these disputes. And there was even situations where he'd be negotiating with cartels saying, you know, somebody would, would steal.

Yeah. And he'd be, the cartels would get mad at him and be threatening him. So, so one thing to be clear at every step of the way, he was doing things that didn't scale. Yeah. And that involved committing crimes. Yeah, yeah, yeah. And so looking at, and just to compare this, so another. company like Airbnb, I got launched out of this era. Airbnb innovated on the vacation rental

because there were some marketplaces, but people would book a vacation rental without seeing any reviews. And then you show up there and just like, well, the toilets don't work and the cheats are all dirty. It's a very, very similar experience. We're just opening it up and creating more trust, really catalyzed some velocity. Yeah. And so as he's growing the site, he starts going by his pseudonym, the Dread Pirate Roberts, which is a reference to the Princess Bride. Yeah, yeah, yeah. Which is like kind of an odd reference. I don't think of that as like particularly libertarian.

It's just kind of like a funny, weird thing to pick because he likes that movie. Really shows you how old he is, though, because that movie came out in like the 80s, right? Yeah, yeah. It's a classic, though. Classic, and he never went by Ross, to be clear. He went by admin. Oh, okay. And so going by Dread Pirate Roberts was a way to build this brand world around the site, right?

It was this pirates themed kind of... It's funny because there's another pirate themed, like, decentralized, like the pirate bay. Are you familiar with this? It's like a bit torrent, right? So for downloading illegal movies and stuff, much less harmful, but still kind of similar piracy theme. Pirates of the internet.

And so people start obsessing over this guy. They revere DPR as a benevolent philosopher king who is giving the world a safer alternative to street drugs. And people really believe, the narrative of like harm reduction that yeah that he you know reduce the amount of violent crime because anytime that there's a drug transaction on the street there's a chance that it goes south and someone gets shot or stabbed or murdered or something like that and that's not going to happen on the Silk Road now the problem is is that well if you're if you're increasing the amount

of drug dealing by a thousand acts even if they're 50% safer well there's still a net negative impact of that and you need to measure that and so sure and and there's sure i'm sure a lot of cartels were like, cool, I'll just go on, I'll just go on Silk Road, buy a bunch of stuff, and then deal it on the street. And so you just wind up with more street deals. Yeah. And at this time, so he, so the whole, this brand world building that he's doing, this is why if you view him purely as a founder and ignore the heinous crimes that he committed over a long period of time, he would do things like they would have a 420 sale, right? Where it was like whoever, if you bought on, if you bought on the site, you got like a ticket that entered you into a, you know, something where you could win like an all-expenses paid trip and stuff like that. So he was following sort of like traditional e-commerce like marketing method. Black Friday Cyber Monday is like huge for him. Yeah. Yeah. Yeah. Let's pump. And he'd be like no fees today on the platform by as many drugs as you want. But meanwhile, the dark side of that is that people would be buying heroin and overdosing and killing like. Yeah. Yeah. It's bad. Like it was very. To him, he he had these libertarian ideas.

deals. Meanwhile, you're enabling addicts to, if you make it almost as easy to buy drugs as going on Amazon, you're enabling a lot of behavior that isn't so great. Yeah. And yeah, I mean, another factor of the growth was that Silk Road really drove early adoption of Bitcoin and it created this crazy flywheel where you would buy a little bit of Bitcoin to go on Silk Road by some psilocybin mushrooms or something. And then because there was so much buying pressure, the price of Bitcoin would go up, you'd still have some left over in your wallet and be like, whoa, I made money. Like, I could buy more now. And then that would drive more people into it. And so it became this, like, virtuous cycle. And I remember reading about some journalists who had tested the site and bought,

like, one, you know, drug pill or something when they were reporting on it. Yeah. And then forgotten about what they left in their wallet. So they bought like, you know, oh, I needed to buy $100 a Bitcoin, and I spent 50 of it on the Silk Road back in 2010. And now it's like 500K or something. And then they have to find their keys. And some of them, some of them did. Some of them lost it.

Yeah. So the crazy thing about this is so Ross, in the very, very early days in 2011, had a girlfriend. Oh, yeah. She didn't agree with what he was doing. And she basically gave him an ultimatum at one point, like me or Silk Road.

Yep. He chose Silk Road. Founder mode. Absolutely. dog. So he goes with that. And then at one point, like later he tells her they like get they they get together again at some point in San Francisco and he tells her that, you know, yeah, I gave the site away all this stuff. His her roommate also during this time found out about it. Because like they got in an argument and she posted on his Facebook wall like I hope the authorities

don't find out about Ross Oldbrick's drug website and then like deleted it. He got her to delete it. And so he had a couple loose ends here. The other thing that's crazy about his experience doing this is he. So anybody who's built a company before, you know that it's the kind of thing that you're thinking about around the clock. You wake up at 3 a.m. You're thinking about it. Yep.

Right as you're going to bed, you're telling your significant other about it, whatever. It's not something you can get off your mind, and it's a very uncomfortable thing to not be able to talk about it because you have all these ideas. Like, right, we're in the gym. I'm like, oh, I need to talk to you about this challenge, whatever. So Ross doesn't have any of that.

He has his girlfriend at first, and then she's like, wow, like my, my boyfriend is like a narco trafficker, like digital pirate. Like, this is crazy. So that, that ends. And then he starts to develop these friendships online, people that he's talking to, people that are users of the platform, people that are sellers. And so he starts developing these friendships. He's got a buddy that's in Vietnam.

Yeah. I forget. He's really developing like two worlds, almost like a split personality thing. Like he actually. Well, he doesn't have. that much of an IRL life. Yeah, it doesn't have much of an IRL life. And then he actually splits his computer hard drive in two and kind of has two wings of the, of

the computer one for his personal life, one for the DreadPriot Roberts. And it's almost like there's like he can switch to just being a completely different person, completely different moral framework. Because even during the trial, a lot of the like, you know, witness testimony character witnesses were like, yeah, even during this period, like he was just totally, totally normal guy. And it's like he was able to put on one face and then. go online and then be chatting with people about like, oh yeah, let's kill this person. Let's bring

in a million pounds of math. Send me the photo. Send me the photo when you finish the job, you know. Yeah. So in this whole time, so law enforcement already by this point has started to develop, like use anonymous accounts to develop relationships with him. He already has a law enforcement that are buying drugs on the platform to like build cases and they're trying to. He has this guy, do Yegi, what's his name? Oh, I don't remember. No. So this is a random guy in Chicago. Chicago, who's with Homeland Security, who's painstakingly collecting thousands of Silk Road orders, which he had figured out. Historically, apparently, people would send drugs in the mail, and they'd have like these

handwritten letters. Yep. So that was like one way you could potentially identify. Sure. But then people on Silk Road wanted to look like they're more professional. So they'd type of that. So they would print out these sort of like, mailing levels, mailing, return labels. And so there was this, this guy, De Yegi, I forget his name.

name. And what's crazy is many of these law enforcement people themselves end up committing crimes as part of the process. It's so chaotic. It's like the most chaotic story. And I mean, there's just so many different organizations involved in this. The DEA, the FBI, the IRS, Homeland Security, the Secret Service. And then they all have varying levels of, of jurisdiction and abilities. And they all want to win. And they all want to win. And they all want credit for it. And so it's just chaotic. Let's go through some of the, some of the posts. just to kind of break it up. I have this one from Roon here

who says, okay, but where are the crypto conferences for people building illegal stuff? I'm talking next generation impenetrable Silk Road, evading Chinese capital controls, war bonds,

hit me up. And this was back in 2022 when crypto was obsessed with like NFTs and just kind of like... Is that pre-FTX crash? Yeah, yeah, this is pre-FTX.

And so, yeah, I mean, I think that, I think, that we can obviously intellectually understand that like everything that happened here is terrible. It's a tragedy. It's like bad in so many ways. But it's deep down, I think if you don't recognize that like, you know, crime stories are a little bit cool and like edgy. Like that's just a very

natural impulse. And I think that's what drew a lot of people to this story. And then the whole reason that we're talking about this is this clip from Autism Capital that says, Breaking Trump confirms that if he becomes president, he will commute the sentence. of Silk Road founder Ross Ulbricht. So this was in May of last year before. So the whole reason that we wound up in this situation is because there was this massive campaign that we'll get into.

Free Ross that was driven by his mother and she'd been campaigning for him a long time. And he got married? Yeah. At one point. And she was working on the project. And then there were a lot of libertarians and a lot of Bitcoin people

that were like, get him out. And so it became this like bargaining chip with Trump as like, oh, we will back you super hard. if you give us our boy back. And so this is when, on January 21st, Ross Ulbricht has been fully pardoned. And Donald Trump says,

I just called the mother of Ross William O'Bright. He misspels his name to let her know that in honor of her in the libertarian movement. He's signing the papers. He's like, what did this guy do again? Murder for hire? Which supported me.

And they will tell you he didn't get convicted of that. Yeah, yeah. He didn't get convicted of her. Allegedly. Yeah, yeah, yeah. It was my pleasure to have signed a full and unconditional, full and unconditional pardon of her son, Ross. The scum that worked to convict him were some of the same lunatics

that were involved in the modern day weaponization of government against me. He was given two life sentences plus 40 years, ridiculous. And Sager says, I'm no fan of the Silk Road, but the way the FBI set up Ross Ulbricht for life is still one of the most insane cases I've ever read about. And so Sager is like hugely anti-drug, even anti-weed, marijuana, Yeah, cannabis. But still, this is a nuanced topic.

Yeah, yeah, yeah. And it is possible that someone can be doing something really bad, and then the government can come in and throw the book at them even harder than they should. Yeah, and so getting kind of into that, you had one of the, one of the, I forget if they were, I don't think they were FBI, but maybe Homeland Security. At one point, one of the guys had turned, like was selling secrets to Ross about all the investigations. Yep. and had made somewhere over half a million dollars selling information to Ross about the case,

even though he kind of still wanted to get him. Yep. But it became this whole game that you could tell that all the agents involved with it were so consumed in the cat and mouse game of chasing this guy. Really, he at one point got citizenship in the Dominican Republic, but it wasn't really a big factor in the story. And so the other thing is like you would imagine that this guy would take off and be in Eastern Europe or Asia somewhere. He went to Australia at one point, but he was just really visiting his sister. I think he did go to the Dominican Republic briefly.

But yeah, maybe some accounts. I think he needs to invest like 100K or under six figures to actually get that. And so let's do a little bit of a tour of the agencies involved in this. The DEA would send agents to infiltrate the marketplace by posing as large-scale narcotics traffickers. The FBI was focused on tracing the Silk Road server, which they eventually got. I think there was one in Iceland or something that they were able to get and decrypt. The IRS wanted to follow the money trail. Homeland Security specialized in customs and import control that recognized that packages from overseas were flooding the U.S. postal system.

One key figure was Gary Alford, an IRS investigator who approached the Silk Road puzzle by following Internet breadcrumbs. He started analyzing older forum posts discussing Silk Roads. launch focusing on handles like altoyd that might have been inadvertently that might have inadvertently revealed a personal connection and so this guy this irs agent ended up getting taken down because he was he he was profiting in different ways from the like from the investigation and very very corrupt but he was one of the guys that traced those old forum posts yeah found posts that an account that was called altoid created by ross olberict at gmail that then he changed to a fake email called Frosty at Frosty.

Yep. But then he had used he called the Iceland server Frosty. Yeah. So they tied that together. Not enough imagination on the username generation, bro. You got to get a password generator. Yeah.

Just go full random. Yeah, it's the Gary Alfred guy, I read a thread about him as well, about how what was interesting was that he didn't have access to like really crazy proprietary databases. And oh, the fingerprint database and the DNA database. It was really just like Google. It was just really good at just like Googling around, finding usernames here, piecing things together. And so all of these investigations were hampered by

a lack of interagency coordination rather than pooling intelligence. Each unit often withheld leads hoping they could be the ones to crack the case. This fragmented approach would simultaneously prolong the manhunt and create openings for suspicious behavior. Indeed, two law enforcement agents, Carl Force, DEA, what a great name. And Sean Bridges, Secret Service would later be indicted for misusing the case to enrich themselves. Wait, it wasn't the IRS guy then? I don't think so. I'm not sure, but.

I thought he got taken down as well. No, I think Gary Alfred was fine. But, I mean, it was, the whole thing was just a mess and chaotic. It's all just happening on the dark web and there's money flowing around and Bitcoin and everyone thinks it's anonymous. It's not. And they would have these meetings called Deconfliction meetings. Yep, yep. Or they'd bring everybody together and like some people would share information and other people would

be like, I'm not going to share that. Like, this is my edge. in the case. It's a very toxic dynamic where within those types of organizations, you rank up and get more titles and comp and paid leave and vacation by cracking cases. So everybody wanted, everybody was self-interested in wanting to crack the case themselves and get credit for it. Yeah. It's always crazy. It's like he stole like $50,000. And then like you look at the price of Bitcoin now and it's like, oh, that would have been like hundreds of billions, hundreds of millions. if they had, like, been able to hold on to it.

But, yeah, the other thing about, so, so law enforcement is, like, starting to have this, like, very, like, multi-prong, chaotic. Yeah. And it didn't take them that long to really start to get some real leads into him and start to put the pressure on, right? So at this time, they're creating fake identities. They're taking over, they're finding sellers. They're going to sellers and they're saying, we are going to put you in prison for life for selling drugs. Yep.

or we'll give you a better sentence if you give us your seller account. Yeah. So then they would flip sellers. Yep. And then they would just be in there. And they also started flipping like different moderators. Yep. And they also just started creating entirely new accounts where they would come in and be like,

yeah, I'm a narco trafficker. And then they would slowly develop a friendship. And there was this weird situation where to develop trust, they would have to give him advice that. real advice. Real advice. You should get three passports. You should get fake passports. You should be able to get to Asia. You should have a contingency plan to get here. You should split up your Bitcoin holding so that, you know, if you get taken down here, like, you know, you'll still. And then, and then on the platform level, they had created this sort of PayPal style system, right? Yep. It was very effective. Every now and then

a seller would build up so much selling history that they would, it's very easy for them to be like, anybody that sends Bitcoin to this address and sends me their address. So people would circumnavigate the platform, right? This happens with all marketplaces, right? Even on Airbnb, somebody will book an Airbnb for like two weeks and then they'll talk to the owner and they'll be like, hey, like, can I just book it through you and like save on the Airbnb fees? Yep. And so every now and then these sellers would, you know, there was one that happened on 420 where this certain seller was like, hey, like, you know, if you just pay me directly like and and they had a big selling history so everybody trusted them and they did like 250k in sales in a day and then just like gone right yeah so that would happen and then ross would be

in this position where he would have he would have all these people that were mad at him his users are pissed at him but he would be like well you circumnavigated the platform to try to save what are not crazy the other crazy things so Silk roads take rate for selling drugs during the most illegal activity is 40% lower than substack, which sells information, you know. Is that true? Subtack's not 50%. No, 40% lower. They substacks 10%. Oh, okay, 6%.

Yeah, yeah, yeah. So Socrow was 6.25%. And it's way lower than OnlyFans too. I think only fans 20%. Yeah, yeah. Yeah, yeah. So Ross was was so mission oriented that that he, he really believe, you know, the part of the crypto ideals is, you know, there's these financial institutions. They have these crazy take rate.

And so he's only getting double the take rate basically of Stripe. And Stripe is like the most regulated just like vanilla, like we sell SaaS. Yeah, he's selling drugs at scale online and doing it 30%. But he's not. Yeah, yeah. And he's not. And when and he still had issues because the kind of people he was dealing with, you know, people that are buying small quantities of drugs online are probably not the most affluent crowd for the most part. So they would have issues with it. And so he keeps dealing with this stuff. And he also has, I don't know, he has issues where, you know, something would happen to a seller. And then the cartels would be coming to him saying like, hey, like you owe us a million dollars. And then he'd be like trying to do deals with them. And so every, every marketplace founder understands that at a certain point, you have this programmatic system that you're running, right?

Where like, you know, let's say you have Airbnb. But early days, I'm sure that the Airbnb, founders would have to be talking to like one of their like listing people and being like hey like we got to work this out yeah but in ross's situation he was dealing with cartels yeah narco traffickers drug dealers and people buying drugs and so it's not like they painted a picture like he's just sitting there and like he created the software yeah yeah he's just writing code but he would be at one point he was wanting to test into product he was encouraged by someone in law enforcement who had built up this fake identity that he was friends with, hey, you should create the masters of

the Silk Road, which was a version of Silk Road specifically for global drug cartels. And so, and, and so Ross worked with his team. He's like, hey, we got to move a kilo right now. And so he was just selling drugs. Drew, just drug dealing fully. He was just fully drug dealing. It's wild. And, and then Mass, I don't think Masters of the Silk Road really took off. He also around this time, I don't know where we get to. So he, he, he, he, somebody started, people would start listing weapons. Well, yeah.

So, I have some extra. So he, he kind of finds like a almost pseudo co-founder, this ally Variety Jones, who offers him marketing tips, security advice, and even recommended that Ross consider violent tactics against potential threats. He, uh, he, uh, he starts thinking about all this violence and alleged murder for higher plots, which would become pivotal to the federal charges against him. And there's still some, uh, he starts. debate over whether these plots were real or entrapment.

But it's really insane how fast the site grows. We said it started in 2011. By late 2012 and into 2013, Silk Road was grossing hundreds of millions of dollars in transactions. Its user base extended across continents and shipments were moving through postal systems at astonishing rates. Within this environment, Ross grappled with immense stress, anxious that each day could be his last as a free man.

Still, he persisted, motivated by profits ideology, and perhaps the thrill of running a digital empire unrestrained by traditional laws. At this point, at this point he's in San Francisco. Yep. He's living in random apart. Like he's not living, he's not living, uh, the lifestyle of Pablo Escobar and nice houses and, you know, cars. He's living in San Francisco paying cash. Yeah. Under fake identity, something Terry, Terry Jones or something like that. He needs, he orders fake IDs off of the dark web. Yep. Off of his own website. So he gets a handful of fake IDs, they get caught in the mail by Homeland Security. And at one point, Homeland Security comes and knocks on his door. And he's like, oh, it's over. Like, he thought he was getting taken down.

And they were just like, hey, like, we got these fake IDs addressed to you. Yeah. It's too small of an issue to be big for us. We just want to know more about the dark. Yeah. Yeah. So he's like, no. Can you bring down for us? Like, we're not going to like do anything right. We're not going to arrest you right now. But where did you get these? And he goes on record and says, there's a site called Silk Road. He can't help himself. He goes in pitching mode. He goes in a pitching founder mode. He's like, well, actually, there's this site you can go on called the Silk Road. It's amazing. Only 6% take rate. And use this coupon code to sign up. You can add some other stuff to your car. Yeah. Guns, drugs. Do whatever. Yeah.

And so he gets on the record at that moment, he gets, that was the first time he was in the system for committing a crime. And they added base, I don't know how it works, but they added notes to like his profile that he had talked about the Silk Road. Sure. which ended up coming back to Biden later. Yeah. And so the DEA at this point is posing as major buyers. The FBI is focusing on these high-tech methods. They find a server in, it is in Iceland.

And there's a bunch of friction between the different agencies because no one really knows who's dealing with cybercrime. So Jared Dürijian. That's Dereygan. Der Yegian, who was the Homeland Security. obsessed. He, Der Yegyan had collected thousands of orders that he had personally caught. Yeah.

And he had just been categorizing them and tracing them to different seller profiles. And he's flipping the dealers to become important, to become informants. He gleaned details about Silk Road's internal workings, including how vendors ship products. Other groups try to trace the site's IP addresses using advanced hacking techniques. At one point, the FBI exploited what might have been a misconfiguration in a Silk Road CAPTCHA or login screen, gleaning. a real IP address that led them to a data center in Iceland. So the whole goal of Tor is to mask the IP address and route through a bunch of different

anyone who's running a tour server all over. So if you just have Spectrum or AT&T internet, you can spin up a tour server and then you're routing random traffic constantly. And so that anonymizes things because traffic's just going all over the place. You can't trace it all. And so anytime it goes through some random server, comes out the other side randomized. or disguised and it's very, very hard to track down. But they found the actual IP address because it seems like the CAPTCHA that was being rendered

wasn't going through Tor. It was going through like, you know, it wasn't being anonymized properly. So they figure it out. They issued the war. This is something that happens throughout the whole Silk Road journey is because he's a first time developer, first time founder. He's just making all these mistakes and oftentimes suffering financial losses, but the site's doing so well.

Yep. And growing so quickly, it doesn't, the losses are offset. But the one thing that's kind of hard to make up for is getting caught. Yeah, getting caught. And so they grab the server from Iceland, decrypted. And it's like the database. It has all the information. And then they're combing through it for clues that could lead them to who this guy is.

Yeah, it wasn't hard to flip the whatever company in Iceland had the server. Oh, no, no. You get approached by U.S. law enforcement. They're like, we believe people are committing crimes on your server. And they're just like, you just want us to copy it? Sure. No problem. Yeah.

whatever. And that goes in, and to be clear, as they uncovered, they got data on all the buyers and sellers and like presumably all the transactions that had happened in Silk Road history, right? Oh yeah. Yeah, it was encrypted, but they were able to basically crack it. Yeah. And every escrow transaction is logged in a database because it has to, even though the, even though the blockchain might be somewhat anonymized, like it's not anonymized when it's. on the database record of the Silk Road, which they got access to it. And so it's not like every person who had ever bought drugs or sold drugs on the Silk Road was suffered from that.

Yeah. But if the government had wanted to, they could have gone and knocked on all the addresses. Exactly. Obviously, people would use fake addresses and fake names and things like that. Yep. And so by 2013, enough agencies had processed enough agencies processed enough scraps of evidence that Ross Ulbrix identity was starting to come into. The focus, though they had yet to conclusively tie him to Dread Pirate Roberts, Ross, for his part,

was often oblivious to the scale of the manhunt, trusting in Tor and Bitcoin's layers of security while continuing to manage Silk Road from cafes and libraries around San Francisco. And so the corruption scandal continues Carl Mark Force, the fourth, and Sean Bridges of the Secret Service. What a great name. Mark Force. The DEA and the Secret Service is one of the most dramatic subplots here. Both men were tasked with bringing down the Silk Road, yet they exploited their positions for personal gain.

Force who operated under undercover personas like Knob on Silk Road directly interacted with DPR. Nob was basically boys with. Oh, yeah. He was like chatting constantly. Talking all day long. You know, Nob would give him, this is not, knob is a guy who'd give him advice. You should get more passports. Yeah, yeah, yeah.

Things like that. And people were like, dude, stop giving him such good advice. And he messed up because at one point he was, he was, he had a fake account and was chatting with, Red Private Roberts and was like, it was like, it's Carl or something like that. And he was like, oh, Carl, Carla, like, I go by a bunch of different names on here. That was all on record. And then was it Sean Bridges at one point that later once Sean knew he was getting caught, he, Sean was like, I will get you killed in prison. He was like threatening like, you're going to get caught. It's all beat all these records.

And so, yeah, it was basically slop on both sides. Very sloppy. And so from Ross's perspective, these agents' manipulative activities made the Silk Road environment feel ever more treacherous. Logs released in court showed Ross confiding to trusted confidants that he suspected infiltration at every turn. He also struggled with evaluating demands for hush money or hits on suspected black mailers, whether Ross seriously believed in orchestrating murder for hire or if he was

cornered into entertaining the idea is still debated. Nevertheless, chat logs indicate that Ross agreed to pay for violent solutions if it meant protecting Silk Road's anonymity and his own. Yeah, so this is, uh, when I posted this and a lot of people are like, oh, you left out murder for hire. Yeah. Obviously on this post, I was purely talking about the business, not endorsing his actions in any way. But yeah, he, he, he over the course of this 2013 era got very comfortable with the idea of paying to get people, killed. So these were people that he thought were informants, people that he thought were just generally bad actors. There's records of him ordering hits and actually negotiating the price of hits, saying, you know, being kind of frustrated with how expensive would something would be.

And the person would be like, well, it's like last minute. And, you know, it's kind of a complicated situation. And he'd be like, okay, he would be provided photo evidence of success, what he thought were successful hits. At one point, law enforcement baked one of the hits using like Campbell's soup and they were taking and using like crappy phone images to like take a picture of somebody's face like with like Campbell's soup like as if their face had been like blown off blown off or whatever. So very, very dark. Like he clearly at this point had was like living this dual life of like he's Ross and then he's dread pirate Roberts. Yeah. And he's willing to do anything to protect. protect the pirate ship.

Yeah, exactly. Silk Road. Yeah. And I think we have a post here. It's probably a few deep, Ben, but this is a screenshot from Reddit. It says, are there any merits to the claims that he hired a hitman to kill friendly chemist? Someone on Silk Road who said they would leak personal info on thousands of Silk Road users and less paid 500K.

I've never seen this discussed in our weekly Ross posts. Yeah, so this was, there was a user who hacked the server in some way. they were able to uncover quite a lot of data on like the sellers. Yep. And so Ross was worried that, I mean, if people realize that if you bought something on the platform and that your personal information could be exposed, then obviously it was going to really slow down sales. He's also facing competition at this time.

Other sites are popping up that are doing the same thing. Yep. So he really didn't want this information to get leaked. Yep. And he, you know, so you get into. And so someone replies to this and says, Ross ordered hits on 14. to five people total. The evidence of this, the evidence that he did this is clear. No one here likes talking about this because it ruins the narrative. Now, the sentence for ordering a hit in

California, if I recall correctly, is 10 years per contract. So he definitely got a harsher sentence than this, but there were other crimes he was guilty of. Not that it will save me from downvotes, but L. Chapo is significantly worse than Ross, and any system that gives Choppo less time than Ross is busted. But to ignore that Ross didn't try to have people executed for doing, is doing no favors to anyone who wants to have an honest discussion about justice. And I think that's just a very interesting thing where there, there was this like big narrative. And if we go back to this debate between Matt Palmer and Joe posting, Matt says Ross Ulbrick's work was one of the largest contributions to harm reduction in narcotics

distribution ever. The systems he created removed violence from the drug trade by anonymizing and physically dislocating transactions. He almost certainly saved thousands of lives. But Joe says, skeptical of this claim, the vast majority of drug trade violence is caused by production, namely South American cartels and opium warlords in the Golden Triangle. Silk Road made retail purchases safer, but that's small potatoes next to the back end. And it's funny because, you know, it's like, yeah, he might have removed some violence, but then he added it all back by trying to commit hits. It's like, come on.

It's just a ridiculous situation. But moving on, we're in mid-2013, and there are enough threads of investigation that they had converged that a direct confrontation with Ross Oldbricht was inevitable. Each agency. So at this time, they had figured out a number of people. So they started with profiles. We think he's his smart guy. We think he's in America.

We think he's... One of the ways they figured how was hilarious was that he he had posted something about how buying drugs on the street is dangerous. And he says something about like those people. And so they use that to justify this idea that he came from like a white middle class upbringing and was not like a street guy because he wouldn't, he would have said like when we buy stuff on the street, we get shot and killed.

Instead, he was like, they get shot and killed. And so that made it clear that he wasn't like from the rough and tumble streets. And so at one point, he was even, he was a suspect just based on his profile of being, you know, interested in technology and libertarianism and all this stuff. But they still were like, they actually, people didn't believe that he could have done it because he didn't fit the hacker, like internet hacker developer archetype because they were like, well, he didn't study computer. science. And so they had this other suspect who had run like a Bitcoin forum that was like more of the prime suspect at the time. So, so he had some nonbelievers in law enforcement that were like, yeah, it could be Ross, but like we don't think he's got that dog in it. Yeah. And so the main breakthrough came from Gary Alford, the IRS agent. And he was convinced that Ross Ulbrick made a

slip up early on. And so he scoured internet forums such as Bitcointalk.org and shrewmary. org going back to the earliest where he was posted hey anybody want to get some shrooms check out this website no he was like has anybody bought shrooms from silk road yeah and it was him selling the shrooms and so he reasoned that whoever created a site might have posted about it during its infancy before it was famous enough to attract major scrutiny and uh this is very funny this actually happened in my high school there was someone who put up a like a cyber bullying type website and the easiest way to figure out who was behind it it's still like myth in my high school but I'm convinced I know who it was because I basically traced it back to patient zero.

And the first person to say anything about it was likely the person that did it. Because how else would it go viral? Like it has to have somebody that introduces it. If a website's just up there on the internet, like no one's going to find that. But the first person to say, hey, have you seen this? That person is, you know, guilty. Guilty. And I'm very convinced of this.

So Alford, this is Gary Alford. He discovered a user named Altoyd posting in January of 20. 11 right at the start of Silk Road, mentioning Silk Road and suggesting people check it out. Further digging reveals that the same user had made a subsequent post looking to hire a tech specialist leaving the email address, Ross Ulbricht at gmail.com. This was the smoking gun Alfred had been searching for as it linked a real-life individual to Silk Road's initial marketing efforts. While other investigators were pursuing the server's physical location, Alfred's simpler

forensic analysis of publicly available posts provided a direct link to Ross. And so Ross had gone back to that forum. he knew he had kind of messed up, and so he changed the email on that account to a fake email. Yeah, but you can see the change. But then you can see the change, and then the name he used for the fake email ended up being the same name he used to name the server. So he had some, he liked the name Frosty. Yeah, and so at this point, they were definitely suspicious of Ross, but they needed to catch him red-handed, logged in as the Dread Pirate Roberts on his laptop if they hoped to secure a watertight prosecution.

And detective work in the digital age can hinge on human error. The altoyed handle and the Gmail address turned out to be one of Ross's biggest mistakes, an unintentional breadcrumb left behind before the soap road success. One thing I'll note here, this is just speculation, but his girlfriend at one point said that he was a big marijuana fan. I think he admitted to it at different points as Dread Pirate Roberts. and marijuana is a performance decreasing drug, makes people forgetful,

it makes people anxious and all these things. And so imagining him as a first time developer high all the time, trying to, you know, committing crimes at scale. And again, I respect the libertarian view that says, well, nothing he's doing was wrong, but he was committing crimes by way of our legal system. Yeah, yeah. And so imagining him just like,

hanging around San Francisco smoking pot, being a narco trafficker. It's like not a, you know, it's inevitable. There's other people that would have made it like seven years, eight years, nine years, 10 years. There's, there's narcos that like real narcos that make it 30 years. Yeah. They buy soccer teams. Yeah. Yeah. Mansions. They get presidents elected. Yeah, it's crazy. So it's actually wild to think about if he had maybe not as intensive marijuana habit or gone and went to Russia or something like that. Totally. There's a world where he would have been come

like a Bitcoin, you know, a hundred billionaire. Like one of the wealthiest was like, you know, influential people in the world and could have been out of reach in many ways.

Totally, yeah. It is so funny that he, you know, his whole philosophy was based on this idea of like drugs aren't this gateway. Nothing bad can happen if you start doing a lot of drugs. And then he is the best case study for

drugs ruining your life. Drugs completely ruined his life. Started doing drugs, then started selling drugs, then started trying to get people killed, then got massively arrested and thrown in jail for a decade. It's like, yeah, drugs were not good for you, bro. Like, calm down. Yep.

And so as the law enforcement officials homed in on Ross Ulbrook, they began careful physical surveillance. His whole laptop was heavily encrypted, and he had a single button that if he pushed it, it would lock and encrypt the hard drive. So even if you got access to it, it, you couldn't access it without a password. They had no way to guess the password.

It would take a computer millions of years to, you know, decrypt that or hundreds of years to force the decryption. Although maybe it's faster now. Who knows? But at the time, they were very worried and they were like, we have to get him while his hands are on the keyboard or hands off the keyboard, but he's still logged in. And so they were excited about the idea that he kept working in public spaces because they could stage an unexpected confrontation.

It's just the irony here of him. in another reality, he would have been in Y Combinator just absolutely crushing it on some like normal business. Yeah. And so he would have been in these coffee shops in San Francisco at the explosion of the sharing economy and mobile. Oh yeah. And he would be sitting next to people that, you know, it's not unlikely that some like very well-known founders today, like we're in the same coffee shop. Totally, totally. The craziness of doing this all like from the epicenter of tech.

I lived in San Francisco when this happened. And I would go to this library. I would go to the San Francisco Public Library. I actually worked out of there a few times. And so on October 1st, 2013, Ross Ulbrook walked into the Glen Park branch of the San Francisco Public Library, sat down and opened his laptop. Unbeknownst to him, undercover agents sat nearby. Agents orchestrated a ruse where one undercover officer posed as a feuding couple, creating a minor commotion to divert Ross's attention. as Ross turned his head.

Other agents sprang from behind, grabbing his laptop before he could even move to close the lid. The narrative of the take down is cinematic. Ross is stunned, half standing, and as FBI pushes him back into his seat and secure his arms. To visualize this, a male and a female agent start yelling and, like,

pushing each other. He looks over. And then they just take his laptop. Yeah, yeah, yeah. It's literally a woman who's an agent just sits right across the table and just is like, pull this away. And so it has it open.

Has it open. And so they immediately, what they do is they take it. They take it and they rush it to a van outside. Like imagine like you're like trying to keep your computer from going to sleep. Oh yeah, yeah. That's what they're doing. And so they're making sure they're moving the mouse. Move the mouse.

And so for the next like 10 hours, they just sit there like moving the mouse and like downloading all of the data off the computer. Yeah. Backing it. I think they backed it up like nine times. Yeah, yeah, yeah. And you could imagine because because when they, when after, we'll talk about this in a bit as they go to prosecute him. the defense's main defense is that it wasn't Ross. Yeah, yeah.

That's their main thing. They were like, somebody else, Red Pirate Roberts framed him. They put all this evidence on his laptop. But they obviously caught him. Yeah, literally red-handed, yeah. With, uh.

And it's like if there was like some sort of like screen sharing software where, yeah, okay, maybe, maybe he was just using his computer and then the real that DreadPriot Roberts was like, you know, putting evidence on there, There would be evidence of that software. But they tracked him so many times where they would be physically viewing him.

He's typing. And he logs on. Because they had moles on the Silk Road team. So they had like law enforcement was being paid to work for the Silk Road. And so they would see when he logged on, he'd be typing. And they would be recording all of this where they'd be watching him typing. And Dread Pirate Roberts is typing. And so it was kind of a tough argument to make, even though various members of law enforcement

were committing their own heinous crimes. It's sort of difficult to defend. And related to that, we have a post from Sam Parr. He did a podcast on Ross Ulbrook, the founder of Silk Road. And he says, he admitted to creating the Silk Road, but said he sold it to a new owner. And they're the ones who did the murder for hire. What do you think?

Was Ross the owner the entire time or in order to murder for hire or no? Why? And Anthony J. Descenzzo says, of course he owned and operated it the entire time. He admitted to creating the Silk Road at the beginning and was caught red hand working on the administration page at the ending. Follow his bitcoins from start to finish. And he had tried to, he told his girlfriend once that he was no longer running it. Then she walked in on him after like they were hanging out. He's on the. And then later he even had this crisis on the platform where he made a post where he's like, I'm going to change the way

I talk so that it's less obvious who I am. And then everybody was like, got, really mad at him because they were like wait like is this safe to use now did it did did they flip them or something like yeah so in the days following his arrest the global community of silk road users found themselves in disarray vendors buyers and administrators scrambled to figure out whether their personal identities had been compromised because as part of the onboarding process to work for the silk road you would have to send him a photo of your real driver's license yeah and it was kind of this like getting sworn in into the secret society there was panic over but bick Bitcoin funds held an escrow, which now seemed out of reach.

People posted frantic messages on the dark web, some lamenting the loss of their livelihood. Oh, it's terrible. You know, you build a drug business and just goes up and smoke one day. Others questioning whether the dread pirate Roberts persona might somehow live on. Law enforcement officials. Also, it's like if there was a real dread pirate Roberts who his fall guy gets busted, why wouldn't you just keep operating the site? Or like make a new thing.

Yeah, just, just, you know. But yeah, you can think about it. So at this point. Or even just move the Bitcoin. They're doing hundreds of millions. of volume a year. Yeah. And you could imagine at any one point there's, I don't know,

a very rough estimate would be $100 million of like drugs, like changing hands. And so that that was quite a lot of money that was just locked up in escrow at this time. That was clearly the Silk Road team had to be the team that would approve when, you know, the recipient would say, yes, got it. Yep. Then they would approve, you know, the payout. Yeah. And so afterwards, after the arrest, law enforcement officials lauded the arrest as a significant victory in the war on cybercrime.

Politicians who had long criticized Silk Road took credit for pushing agencies to shut it down. Mainstream media coverage exploded with headlines labeling Ross as everything from a brilliant hacker to a modern day outlaw. The sensationalism surrounding the kingpin moniker overshadowed the complexities of Ross's motivation. He was painted as a cold-blooded trafficker who profited off of addiction and despair. I think that's probably correct. His friends and family were stunned. He definitely was cold-blooded. He was able to separate the nice guy, Ross, who had a loving family from, he embodied, fully embodied the dread pirate Roberts.

Yeah, I wonder if there's anyone that gets to this level that isn't cold-blooded, though. It's kind of a prerequisite. Every, every billion dollar, every unicorn CEO that I know, even the ones that are nice, you can tell. Ice in their veins. Ice in their veins. Yeah. Meanwhile, the government's forensic teams sifted through the trove of data from Ross's computer. They discovered personal diaries, chat logs between Dreadpiret Roberts and administrators and spreadsheets,

detailing profits. The data portrayed a person intimately involved in day-to-day operations from vendor disputes to strategies and strategies for evading law enforcement. You could just call that drug argument. How to evade the government, federal, the IRS.txt. How to beat the secret service. Yeah, that was a big part of this. He was not living this high life because he had all this Bitcoin, but he couldn't really figure out a way to get it off chain. Yeah. It was clear that Ross faced a battery of charges, money laundering, computer hacking, drug trafficking, conspiracy

to commit murder. The last charge stemmed from his suspected willingness to eliminate threats to Silk Road's anonymity, whether or not actual violence occurred. Federal prosecutors used the chat transcripts to argue that Ross was willing to orchestrate killings, as the the immediate chaos settled, the case moved toward a long legal process that would captivate the public and set legal precedence about dark web marketplaces. And so the defense strategy, the legal defense strategy for Ross focused in on a few key ideas, wrongful identification. They argued that Ross wasn't the true dread pirate Roberts and that the real DPR had set him up. Reliability of evidence, they questioned whether

the laptop data could have been tampered with, especially considering the corruption charges. That's a, yeah, because of how corrupt many members, not many, but at least two, Carl Force and Sean Bridges. Carl Force. It's like, yeah, maybe there was something weird going on. Yeah, and they were encouraging him to do things that work. Like, that was a whole, like, I think it was knob who would be like, yeah, you should take this guy out.

And it's, oh, yeah, they definitely pushed him on that. That was the intractive thing. They helped him. They were involved with helping fake these sort of murders as well. Yeah. And so they wound up not convicting him of that, but it definitely set the tone of the trial and kind of get people out of this mindset of like,

he just built a website. And he was a young guy. And maybe he did this other stuff. We're not even going to push him on that, but it like threw everything else. Maybe he did a few things that didn't scale. But it also is interesting. It would have been sort of, I imagine,

untenable if he was convicted on the murder for hire stuff. Yeah. To actually get pardon. So that small detail, I think it doesn't seem like Trump knows his actual name. Yeah. But I imagine there would have been much more pushback if he had been actually convicted of. So look at this digital diary. There's an excerpt in here that says, I'm doing about $2 million a week in sales now.

They also had logs with references to specific violent acts that Ross had allegedly considered, or at least verbally agreed to pay for. Yeah, and so he had sent $750,000 of money to fund hits. So he was paying for the hits. Yeah. They weren't happening. So he pleads not guilty. And there's a public trial.

His mother becomes a tireless advocate for his cause. They argued that the government was making an example out of him, which I think is probably true since this was like the first major cybercrime. deep web case. And if they didn't throw the pick at him, there would just be a million more. And there was super, like there was Gawker articles.

Yeah, exactly. It became mainstream for sure. And so it's very embarrassing to law enforcement to have a website that anybody can go to to buy. Yep. That's, if law enforcement can log in and view it

and regular, like a high school kid can do it and some, you know, grandma can do it. Yeah, it's just such a violation of like the social contract. Yeah. in the in the in the end the case would be resolved in the public eye with ross heading to trial january of 2015 a date that would not only decide his fate but set a precedent for the future of dark web marketplaces the trial took place in new york in manhattan federal court presided over by judge katherine forest it was a high profile case and the courtroom was often packed with journalists curious onlookers and ross supporters um the atmosphere was tense the prosecution maintained a mountain of files presented a mountain of files taken from ross

So this was part of the challenge for the defense is there was like six terabytes of data that they were like using against him. And so they're like they're like, yeah, like we're going to use all of this against you. Like good luck. One other thing he benefited from is he had tried to set up a Silk Road armory. Oh yeah. But it totally flopped. Yeah, yeah. You know, like, you know, sometimes you hit PMF with one thing.

Side project. He realized that it was like much harder to like send guns in the mail. Oh, yeah. But he was trying to sell. I think you could buy explosives. on the self-row. Oh, yeah, grenades, and you can buy fully automatic weapons. Which is just like, even, I think, that gave Ross some pause around, you know, that.

Yeah, this was the libertarian thing that always got to me. It's like, he would argue that, you know, I am willing, I'm willing to consent to sell drugs, you're willing to buy drugs, therefore the government has no say in that. But there's always this question of, like, negative externalities. And what got to me was like, well, at a certain point, you were selling. like cyanide and yeah and and also it's like okay even if that's not bad between the two parties what about the the UPS guy that has to carry that like did they opt into that so if explosives are going into the mail yeah it's like oh I'm just a I'm just a FedEx driver and now I have to carry grenades

like I didn't opt into that like in your libertarian philosophy I don't consent to that libertarianism seems to work if you had a private island yes and and everybody there is there by I mean like, basically going there. But no, it's totally, when your actions are affecting other people that are participating in a society that is agreed to have certain laws. Yeah, you never think about the negative externality of, okay, you brought these drugs in and then the drug dealers on the street and commits violence because of this. Or you're funding a, like, it's always just, let's look at just the one, the one micro

interaction and not consider the knock on effects. And that's why we have like an actual society and society. contract. And so the government witnesses included law enforcement agents who had participated in Silk Road investigations. Ross's defense contended that he was set up. Friends testified to his peaceful nature pointing out how out of character it was for him to order violence or run a massive drug site. Over the course of three weeks, jurors were presented with a dizzying array of digital exhibits, as you mentioned, six terabytes or something like that, IP addresses, chat logs and email addresses. And on February 4th, 2015, after,

only a few hours of deliberation, the jury returned a guilty verdict on all seven counts, including narcotics trafficking, computer hacking, money laundering, and maintaining an ongoing criminal enterprise. Yeah, and so to be clear here, this was like three weeks. It was not the kind of trial that stretches on

for months and months or years. It was like very quick. And the deck was certainly stacked against him, right? He had all these different agencies that were, and there were so much evidence. Yeah. And when you think about other trials, it's like, well, yeah, like maybe they got some photos or something or maybe they kicked down a door and they found some with something. But it's all, there's always this question of like, oh, was it planted? But when you when you show someone six terabytes of like, okay, we got his computer, his hands were on the keyboard. He was logged in in this profile. It's a lot harder to paint the, you know, the defense case. This is why it ends up being really hard for law enforcement to pin down traditional narcos. And they get them on money laundering and things like that.

Yep. Or wire fraud. You know, this stuff that's like unrelated to. Al Capone went down for tax evasion, not drug trafficking. And it's because it's really hard, especially the, Ross was able to build a massive enterprise quickly because of the internet. But he also generated a massive amount of data related to that activity. and a narco, you know, you have like them on like a radio, you know, talking or whatever. You know, there's less of a lot of narcos, at least in Mexico, have like regular businesses too

that they're talking about. Yeah, yeah. And so not the same amount of evidence. So sentencing happens a couple months later, May 29th, 2015. Judge Catherine Forrest sentences him to two life sentences plus 40 years. And because it's a federal case, there's no possibility of parole. So he is truly screwed. And this was one of the, it was significantly harsher than typical sentences for nonviolent drug offenses, especially given his age 31 at the time.

Which is like, it's not bad young, but the whole 30, you know. He was young. He did this in his 20s, but he wasn't a teenager who was like. His prefrontal cortex was developed. Yeah. Developed. And he had a good five year run with a full cortex. Full cortex.

Full cortex mode. Yeah. And so... He was smoking pot. They should have included that. Yeah. Ross appealed the decision and this outcome underscore the U.S. justice system stance on dark web marketplaces, an uncompromising approach that signaled to future clandestine operators that they too could face the rest of their lives in prison.

And we didn't really see a replacement pop up very quickly. It was pretty... I don't know. I think it was effective. I... Well, the dart... I mean, many of these transactions, like, still occur on the dark. Like, I think, I can't tell you.

But there's not a brand name. No, so here's the thing. Any, I think historically, any time a drug has become a brand name, they get taken down, right? Sure. And so, like, that's why when you think about, okay, billions of dollars a year of cocaine gets consumed in the U.S., there's no brand that you can think of, right? And as soon as you tie a brand to it, then there's like, well, now we can start tracing this. And so it's just like.

Colombian cocaine. Yeah, like that's sort of yeah, yeah, it's sort of that that one's more obvious. Yeah. But yeah, I imagine that this way more volume happens now. It's just not these sort of brand name marketplaces, right? Silk Road was the Door Dash. It really was. Or the Uber of hits, you know, whatever you wanted. And so Carl Force and Sean Bridges, um, after the, after the Ross conviction, the scandal emerges and they get sentenced to six and a half and six years. Um, and then more when further misconduct was discovered. And so they go to jail for stealing a bunch of money and basically not doing their job at all. Meanwhile, well, it's a weird thing. They were trying to catch him, but they were also trying to make, you know, a buck in the process. Yeah. Weird thing.

And so it goes on. But let's move. So that was, let's see, 2015. In 2016 to 2016 to 2017, he, Ross Ulbricht appeals appeals, asserting evidence was tainted by law enforcement corruption, but the court upheld the guilty verdict. In May of 2017, the Second Circuit of New York reaffirms Ross's life sentence, reasoning that the massive scope of Silk Road justifies an exceptionally harsh penalty. In 2018, June 28th, the U.S. Supreme Court declines to hear Ross's final appeal, effectively exhausting his direct legal avenues for overturning the conviction. Lynn Ulbric during this time, 2015 onward, spearheads the Free Ross Movement nationwide, championing her son's cause. And Lynn is a cracked marketer.

Cracked. Absolutely. Like she went to town for her son. And so in December of 2021, the Ulbricht family auctions Ross's writings and artwork as NFTs online, raising millions through the decentralized organization Free Ross Dow to support continued legal efforts. in 2021 to 2022, a stash of over 50,000 stolen Bitcoin from the Silk Road is seized by the U.S. government, and Ross relinquishes any claim to help satisfy his $183 million restitution order. And in 2024, Pirate Wires reported on April 2nd that the U.S. government has moved roughly $2 billion in Bitcoin confiscated from Silk Road to what's reported to be a coin base wallet per coin desk. And everybody's like, are they going to market sell this?

Are they going to sell this? They're going to have that? Everyone's worried. And so in May of 2022, Congressman Thomas Massey calls publicly in the House of Representatives for Ross's commutation, viewing the life sentence as a grave injustice beyond appropriate punishment. And to be clear, to be clear, it is a bad precedent that law enforcement agencies can commit a bunch of their own crimes in the effort. in the effort to take somebody down and then still if they get like it is a terrible precedent right

we we want to live in a society ideally where nobody yeah a vibey libertarian utopia where nobody nobody's like hurting each other but then if one person commits a crime it doesn't justify another person to commit crimes in order to take that person down and so many while while we both believe that many of the things that he did were net net negative and bad it it is is, you know, the people on the other side of that that don't believe he did anything wrong, they also have a, you know, I don't necessarily think they have a good point there. They do have a point that it's a bad precedent that, you know, law enforcement should be able to commit crimes and then still, you know, deliver this like, you know, very intense sentencing.

Yeah. And so let's go through some of the more recent facts about how we got to the pardon. In May of 2024, Donald Trump promises that he will pardon Ross on day one, swaying libertarian-minded voters and crypto enthusiasts. In June of 2024, RFK Jr., running for president similarly vows in a public statement to grant Ross a pardon. In November, Chase Oliver, the Libertarian Party's presidential... RFK's spotlight, Silk Road's role in pioneering cryptocurrency, which is kind of like everybody's...

like really just trying to lean into the marketing. For sure. The Libertarian Party's presidential nominee reaffirms in Reason Magazine that a full pardon aligns with fundamental libertarian principles. I kind of debate that. And then on January 21st, 2025, President Donald Trump signs a full and unconditional pardon from the Oval Office fulfilling his campaign pledge. And so the crazy thing about this is like imagine you're Ross.

You're in jail for a decade. Yeah. Trump gets elected, signs a dock, 48 hours later, you just walk out of prison. Yeah, you're free.

And that's why I felt like this was a, I wouldn't have wanted to cover this story at all. Yeah. A month ago. Yeah. Because it was just purely dark. It's like, okay, he committed a bunch of crimes.

Now he's life in jail. Yes, he was, the execution from a business standpoint. Yeah, now he's, now he's taking it on the road show. No, and there's rumors that he's still worth, you know, presumably. Yeah.

It seems like he might have had a second or third Bitcoin wallet that was not seized. And so he might have come out with a lot of wealth. Yeah. Which would be fascinating. And I think, and I don't think. Yeah. Yeah. Yeah.

Ultimate hoddle. Ultimate hoddler. And so Ross now free speaks across the U.S. about prison reform, cyber security, and the lessons of Silk Road aiming to reshape his legacy in the libertarian movement. And he's pardoned and living under close public scrutiny, continues to champion libertarian advocacy

and prison rehabilitation efforts. So it'll be interesting to see once he actually does like the Joe Rogan style interview or the Lex Friedman like the full story. I am interested to hear his side of this. I would be surprised if he doubles down and says I wasn't the PR. That would be interesting.

Does he need to do that? No, no, no. He doesn't need to do that. So if he does do that, that would be and then he's able to prove like through some crazy thing, which I don't believe. Yep. The main thing, my main takeaway is this guy is a goaded operator.

That's true. Like clearly, like, he benefited from some luck and timing. It was like this missionary founder. But self-taught developer, first-time CEO, bootstrapped, zero to a billion in sales. Very, very profitable the entire way. Operating a remote team. He has clearly is very powerful. In some ways, he's got insane aura.

Yeah. And if he uses that for good. be very good. It could be very good. In some ways, he's playing the game on hard mode in the sense that, as you mentioned, he can't talk about people, you know, he can't talk to anyone about his business. He can't, you know, get, you know, public investors or advisors or anything. But on the other side, he is selling the most addictive and product that just has immense pull from the market because it's illegal. And so, I don't know how much credit you actually want to give him. Let's go through some more posts. We got Captain Nemo says Silk Road's historical importance is still underrated.

I thought that was interesting from 2022. This was well before the pardon. And there's another post from Captain Nemo in here, 2022 says, the ultimate Tech Anon moment was when Forbes did an interview with the Dread Pirate Roberts, the anonymous founder of Silk Road, at the height of Silk Road's ascent when multiple federal agencies were trying and failing to track him down. That's so funny that he did an anonymous interview successfully with Forbes because it's like now, you know, Beth Jaisos can't stay undoxed and like, you know, can't do an interview without getting dached. And this guy was able to to stay anonymous while appearing in Forbes. Yeah, one of our one note, a friend of the show, Lone Ranger says Donald Trump paved the way

to be to the White House as a quote unquote convicted felon, giving Ross a clear path to a presidential run in the future, Ross, who turned 40 in prison, is now old enough to run for president and could be the first libertarian president if elected. So that's a stretch to me. It's a joke, but it's one of those things where it's like, maybe. But oftentimes the most entertaining outcome is the most likely. I agree. And I, it does feel like not totally out of, like if Ross's goal is to create a libertarian utopia, the United States is awesome. It would be a great place to be a true utopia. And if that's what he believes, now he's got a few hundred million bucks. He can finance a campaign. You know, I will be campaigning against him. That's all I can say.

Yeah, yeah. We need law and order. Captain Nemo continues and says, more recently, he posted this January 21st when Ross was pardoned, hard to overstate how impactful it was to be a young, earnestly curious student when Ross Ulbricht was running Silk Road as the Dreadpire Roberts, operating a completely illegal digital bizarre openly for years, despite every effort to stop it by the most powerful nation state ever. Massive Hansonian view quake viscerally opened my eyes to technology's power to reshape existing power dynamics, and so the flow of history for better or worse. I think that's super interesting to think about how powerful the internet and this new technology was, that even though

it was only two years and he did get shut down, he went on a generational run and was able to actually do this thing that no one else is able to do. Yeah, purely in his technology. Crypto for a long time was looking for use cases outside of Bitcoin being digital gold. Yep. Meme coins being speculation and other activities. And the most obvious use case outside of digital gold that had the craziest product market fit was the Silk Road, right?

Yep. But no crypto VC could be like, oh, look, look what's possible. Yeah. Like, this is the potential of crypto because there was just like a thousands of crimes being committed a day. You know, like that's crazy. And it's funny because everybody that was, if all the people that had just been, you know, committing crimes with BTC early days of Silk Road, if they had just bought and held, they'd all be millionaires because they'd be using like 20 Bitcoin to buy like one pill. Yeah. So that was like an expensive night on the town, you know. Yeah, the stories around Bitcoin have evolved so much. Like version one was this, we've used math to solve the generals problem, and it's going to be decentralized and not controlled by the government

and perfectly anonymous. And then a lot of those stories kind of eroded, and it became just this crime money, basically. And then it somehow made it through that and just became digital gold. And then there were new narratives around, oh, well, there'll be all these, like, games and metaphors. built on top of it and then that kind of melted away and now it's prediction markets and and meme coins and digital gold and just kind of a store of value but it's interesting that at

every point in time there's been so many negative narratives about this is the worst thing ever but there's been this one enduring narrative that's stuck with it and driven value and now it's at 100k it's crazy well yeah and it's still even though what like the I don't like honest analysis of crypto technology, you could never say this isn't cool. Exactly. Like this isn't like this crazy, truly novel transformative technology. It's just that the primary case study was a nefarious use case. Yeah.

But it's still incredible that you could operate an open air drug bizarre online that anybody could go to and see and law enforcement could see and not do anything about. Like that is a wild. It was insane. wild situation. So let's go to Catherine Hahn, former GP at Andriesen Crypto now with Han Ventures. And she is quoting the, I think the founder of Chainalysis, which is a company that does research into how crypto is used nefariously. Joni Levin says, today is my proudest day at Chainalysis. We enabled U.S., UK, Germany, and South Korea to take down one of the largest child abuse material sites, law enforcement in 38 countries made 330 arrests of alleged pedophiles and rescued

23 children from abuse. And Catherine Hahn says, great work by chain aliasis and the authorities in such an important case. As I found when working on the Silk Road corrupt agent and other cases, having the digital breadcrumbs of crypto was a useful tool for law enforcement. And so she's actually saying that like the reason that she was able, I didn't know this, but I guess she worked on on those corrupt agents in the Silk Road case. And the reason they were obviously taking payment in Bitcoin. Exactly. And so they were able to trace the Bitcoin where it moved out of the Silk Road account into their account and then was eventually taken off chain.

And I think they were buying like sports cars and stuff. And so they were able to figure that out and track it down. And she was saying like if this was Fiat, that would not have been possible. And so that's actually the exact opposite narrative that people were telling in crypto, but it still worked out. I don't know if she's saying it wasn't possible, but normally. Much harder. So if a corrupt agent takes a cash payment, there's not a record of that.

Yeah, the classic, like you kicked down the door, you've gotten the drug kingpin, and there's a huge stack of cash, a huge stack of cocaine, and then a bunch of luxury watches. And you're just like, yeah, throw the watches in the back. One node. Yeah. Yeah. Oh, yeah. We got a lot of cash and a lot of cocaine.

I didn't see any gold bars there. Yeah. Throw it in the backpack. And like, you see this in like training day and like those like kind of cop movies. And there probably some element of truth and, you know, a lot of good agents. It shows, don't do that. It shows the, you know, the fact that Catherine was working on this crazy high profile case. Yeah.

Still recognize the potential of crypto and goes on to have a $2 billion fund now. She was a GP. At an historic run. Historic run. Yeah. 2021. Coin base and stuff. Yeah, fascinating.

So wild, wild story. So let's go to a promoted post. We're done with the Ross Albrecht story. I hope you enjoyed it. we're moving on. We're going to go to the timeline, but first we got a promoted post from Josh Steinman, promoting my product.

The one I worked on with Jordy. He says, co-founder just texted me confirming that Excel performs as promised. And Excel, if you don't know, is a project, a little marketing stunt that Jordy and I worked on to deliver shareholder value. So it's a nicotine product that says right on.

And this was, in many ways, the genesis of TV because we would. This was the first loud opulence podcast. Yeah, we would riff on. what nicotine meant in the context of value creation and that many of those early conversations sort of transformed into the show. So it's a cool physical. Yeah, I mean, I think basically the

first time we did anything that felt like this format was the launch video type of thing. We did a vibreel and then we did an hour sit down and we both came into it being like, oh yeah, like let's just like riff on this and talk about what we did and clip it. And we just were fully in character the whole time and Ben, our vice president was like, is Jordy an actor? Like, like, what was that? And we were, there's some amazing quotes in there. It's fantastic. That's a lost episode. We've never uploaded the full thing. So it's like over an hour. We've only released like three clips. I can't believe we recorded for now. Oh yeah. We were just on fire. We were just locked in. Podcasts high. For sure. For sure. Let's move to a very important story. Valentine's Day is coming up. I'm

there's someone in your life that you need to get something nice. I was asking, Jordi, if there's not, this is a good opportunity. Kind of an aggressive move right now to reach out to somebody maybe even thinking about. Cuffing season is here. Cuffing season. And so I was asking Jordy, you know, what is the Patech Philippe of Valentine's Day gifts? He said the Birken bag, which I think is a good answer. I did hear that Burkins are maybe not a Valentine's Day gift, more of Christmas or birthday gift and Valentine's Day is more jewelry and chocolates. So we need to find some good jewelry and chocolate options. Some controversy. We're going to do a whole deep dive on Valentine's Day gift guide

for sure. Yep. But there's an interesting story going on right now because Walmart has cloned the Birkenbag. Oh, I know. Have you heard about this? So Bethany Franklin says, we are witnessing a luxury goods rebellion and we are rewriting the rules of fashion. Last night I talked with NBC news about the viral Walmart Birkenbag is taken over social media. And they call it the working bag and it's 80 bucks and it's a clear knock on. And the Birken bag, if you're not familiar, is from LVMH, right? Or Hermes? Hermes.

Separate company. It's been family owned. And it's public now. What does it cost to get in the game? It's like 20K, 30K? Something like that? Yeah, I mean, it's the same thing as Potech and that there's trade in these crazy range. You can go up to 200K for alligator or stuff.

Yeah. You can't really just walk in off the street. Exactly. Yeah. Yeah, you have to be invited. You have a relationship with. And so it is, it is in that, you know, Otomar Paget, you know, Patac, Holy Trinity. Holy Trinity. Yeah, where, where. And they started with saddles. Oh, yeah, they did. Yeah. That's right. Yeah, so there's an amazing acquired episode on the history of Hermes and there's this family

drama and that, you know, this whole dynamic. It's kind of, it's a wild, wild story. So we should do the TB reaction to the acquired episode on Ermes. Mez. Okay. I like that. We just press play for five seconds. Pause. Yeah, it was funny because I, I was looking up like Walmart stuff and I got a ton of posts from Bernie Sanders being like, they make too much money. And I was like, well, I hate them too, but for a different reason. Me and Bernie can agree on something. We agree on one thing. They're ripping off luxury goods. They should be charging more. Yeah, I mean, but tying this back to yesterday, we were talking about what's happening in the automotive industry with the premium brands being knocked off.

The thing here is the way I can guarantee you the woman or the households that are buying Birken bags are not seeing this Walmart product and saying I'm going to get that instead. It's people that aren't buyers of Hermes that are saying I just like the way this purse looks. Maybe they like the way the fact that it looks like at a Birken. I mean, you could always buy a knockoff Birken bag at like. like Canal Street, right? On the street, the street vendors, very low quality knockoffs. Same thing with watches or anything.

The only things that are hard to buy knockoffs of are cars. They do make perfect replicas of cars. Yeah, yeah, but I mean, they still end up being like $200,000. Because it's like a super bespoke. There's only so much you can do. Yeah. But yeah, it's just funny that Walmart would do this. I feel like there would be a legal battle on the design patent or something.

But who knows? Oh, I'm sure, though. But it's also like the Birkin's been, in market for so long, can you actually defend that IP? You can have design patents. Yeah, maybe not. And that's why they can't,

they can't copy the brand. Walmart is not a move fast and break thing, like small, oh, like maybe we'll get a letter for a MRAMS. Like they clearly consider the, the legal impact of this before they launched it. And yeah,

the main thing here too is, is if you hold an actual Hermes product up to this. Oh yeah. It's not close. It's not anywhere close. And so the, some people will argue, oh, it's like it's the same exact thing. Why is Hermes charging this much?

And it's like charging that because of the brand, but also the quality. Like they've mastered the art of making these bags and they have incredible margins, but it's still a fantastic product. Yeah, like, you know, $80, like the cogs on a, on a true real Arames-Burkin bag are probably a thousand. Probably a thousand. So 10 times as much just in quality. And then you get the actual brand and all the extra lift on top of that. that was an interesting story, but stay tuned for our full Valentine's Day gift guide and our reaction to the story of Aramaz. I'm sure we'll cover it at some point. This is interesting.

I saw this going around some friends. There's a scam going on. You know, we love to highlight scams on this show. There's a new scam I've just recently heard about says David, the QR code scam. People are either sending you a package or they put it in your mailbox inside might be a small item, not of much value or no item at all. With that, there's a coupon or a piece of paper with a QR code on it for you to scan. I might say free or discounted or even a giveaway item. Once you scan the QR code, the hackers now have access to all or some of your data,

including financial information. So they're basically trying to trick you to go into like a fishing flow. So they'll be like scan this QR code and it takes you to like Amazon login.com and then it looks and then you type in your username and password and they've basically fished you. And they've used like the analog real world to get you into a fishing. flow which where that would normally be caught by a spam filter yeah yeah yeah it's not going to be filtered by real world real world action yeah yeah very i had a buddy recently that got his instagram account taken over and it was like a bunch of instagram stories of him hyping up this person

and coach maria it was like helping him invest and he's like you know i don't normally like it was written in a way it was like i don't normally talk about this kind of stuff oh he was just like text over it was like instagram story yeah so it was like text over an image and it was like i don't normally talk about this stuff but i started working with this woman named coach marri Maria. She helped me turn two grand into 50, like reach out to her. And, and of course, like, all me and all my other friends are like, awesome. Like, I got a hit up coach, all the comments. No, no, no. Finally got back in. I honestly don't think. I don't think it, I don't think it fooled anybody. But they're doing that at such an extreme scale that they're bound to get. So this is a

promoted post for two-factor authentication. Turn it on, lock your accounts down. The hackers are desperate these days. Anything for a buck? Have you heard of lumpy mail generally? No, what's that? So lumpy mail is a, so basically, like if you send someone like junk mail, they'll usually just filter it out and, and they'll just throw it out like at the mailbox. When you get your mail, you'll just throw out like, oh, this is clearly like a flyer for something. I'm just going to throw this out. Then the like physical mail retail marketers figured out that if they got a machine to hand write your name and your address and make it look like it was script

and like it was written by an actual person with a real stamp on it, it would be like, oh, maybe this is a Christmas card, I'll open it up, they open it, then they see the ad. And they engage a little bit more. Lumpy mail is a letter that has something in it, like a little chocolate. And so the envelope looks like it's lumpy. And so it's more expensive to send through the mail, but it has a higher open rate. And so you can send people lumpy mail. Conversion optimization.

And it's all about conversion optimization to just getting someone to open the physical mail, see the ad, and then potentially buy the product. And this is just the next generation of that. Yeah. The classic is, I mean, I would pay an almost ungodly amount of money to never receive physical mail from, I would even be willing to give up letters from loved ones and Christmas cards and fully switch to digital just to not get physical mail because it's just such a. Have you ever done like Earth Class mail or any of those? like P.O. Box forwarding and then it scans it all. You see it as an email. I know, but like enough people have it. Figure it out. Yeah, you have. They just can send it to your address. Yeah, yeah, yeah. It's rough. But you got to be really careful. Now is like the like, they'll just write, do not throw away. Yeah. Exactly. You're like you open it and it's like, it's like, okay, it's an ad for some insurance thing that I don't need. I feel like in here, there is a non-scam version of a very viral marketing strategy. Like if you, like if we were to do some sort of drop where we got a bunch of people's addresses and we sent them something that looked

like a package. They opened it up. There's a QR code. It goes to a vibe real announcing season two of Technology Brothers. Like that or just goes to public service. Exactly. It just goes to like our X account and just says like you want to follow it. Oh, speaking of mail, lumpy mail. We were out of town for a couple weeks, but I came home to this, a package from Furentino label. And so I can open it up here quickly. I'll read this. So if you don't know, Fierntino label is by an entrepreneur and friend of the show, John Fierentino. He says, gentlemen, it is with the deepest admiration that I present to you this handcrafted tie from Fierentino label, conceived by hand and brought to life by Fabio Calderlla, a true maestro of Italian bespoke tailoring with over a quarter of century of mastery.

This piece is a testament to the timeless elegance and unparalleled craftsmanship we hold sacred. got this wonderful letter from John himself, and then I saved it to actually open it on the show. John had given us a little bit of, you know, a subtle critique that we don't wear ties enough. And so now we don't have an excuse. This thing is fantastic. So expect to see us. Yeah, rocking that. I like that he just sent us one tie.

Well, I think you got one. Where's the second tie? Is there two other two times? I think he sent one to you too, which is like buried in your mail. No, no, no, it's our communal tie. We'll trade off. Yeah, yeah, yeah. We'll share it.

Or I can just hold one side and hold the other and it's like the fidget type thing. Exactly. No, no, no, it's great. I need to get, I need to, that's the next evolution. I got a bunch of new suits. I'm sporting the white suit today. I've got a couple more. I'll only be strutting on the show.

You look like Colonel Sanders had a liquidity event. That's what I'm going for. That's the look. But definitely the tie game is next. I like the open, the open collar look, but, you know, it's a little too casual for her podcast, for technology podcast. Yeah. Oh, that's great. Well, thank you to John Fio.

And if you do wind up purchasing a suit from Fierentino label, tell him the technology brother sent you. We've already had a few people go through. We've had a couple conversions. What was that? Okay. In the group, yeah, you want to read some Q&A stuff? Jordy, what you got? Q&A.

Let's see. Can we get an affiliate link in the description of those suits? Incredible suits today from Sky. Honestly, Thank you. We really appreciate it. AJ Sharp says honestly a groundbreaking banana bone cream. Yes.

And John says white suits for a black market. Tyler says VP, let's get a ticker tape on the bottom where the fits are from with codes. New ad skew. Yep. This is great. So this is great. This is great. This is fantastic.

We see you in the chat, guys. Ashley says, my question is what's their plan to monetize with this podcast? Brother, we are already monetizing on a scale. Never before seen. You don't become the most profitable podcast in the world with running a lot of ads. I do think that we are very much on the side of integrated brand partnerships, advertisements. And there's a very good reason for this.

I mean, first off, we don't believe in audience capture. And if so, if we were charging the viewers, pretty soon, we'd be like, what do the viewers want to see? And then we'd have to deliver exactly what they want. And so we are 100% corporate supported here. So if you want to help support the company, call a company and have them sponsor the podcast. This is the key. Yeah.

But seriously, like there's a long, long history of this debate about should you just charge like $5 or $10 a month, the strategyry model versus an ad model. we decided early on that the ad model was much better because it allows us more flexibility. We like supporting companies. We like supporting our cool companies. And also it allows you to price discriminate in a way that captures more value. So there are strategority readers that are literally billionaires who work at big tech companies. They pay the same price as the college student who's paying $10 a month. Which is cool. It's cool. It's great. But it but it doesn't allow them. And we would rather we would rather we would rather we would rather any

anybody in the entire world to be able to come and get the 12 hours of content that we put out a week, totally for free. Yeah. And then if you want to use ramp or work with other partners, if you want to go on Bezell and buy, you know, a fantastic watch, you can do that. So it's got a couple other ads. Got a couple other questions. Ashley says, asking about S&P 500 and just like public market investing in the future, is it going to be, you know, will companies require the same amount of capital as they get more and more efficient with AI? My answer there is that I think we're still in the early stages of just the hyper financialization of the world. Crypto is the hyper financialization of memes, right?

And companies, I don't think the capital markets are going anywhere. In many ways, people predict that labor will be less valuable and capital will be potentially have, be able to be more impactful because you can now, you know, we talked about this yesterday with all these rollups. These companies are being bought because they have a thousand employees and people think they can bring that down to 50 employees by replacing people of AI. So capital actually becomes more valuable. Also, when you just talk about the S&P 500, the S&P 500 is an index on American business. And obviously, like America has this narrative of like we're in turmoil. We're more divided than ever before. Things are crazy.

There's stagnation and there's so many bad things happening. But when you go around to other countries, we just did a deal. dive yesterday on Germany's economy, and they are in a much worse position than us. And China's economy has been stagnating because of COVID lockdowns and deflation and currency fluctuation. And soon tariffs. Yeah, tariffs. And there's so many other things.

So the question of like S&P 500 or not, to me, feels like America or not, and I'm still very long. Yeah. Relative to everything else. The markets are a reflection of the general value, the earnings potential of those businesses as well as their future potential and the memes. that get in the narratives that get built around these companies. And so,

thankfully, we're still going to be able to deploy capital with our absolute boys. Exactly. We got another, another just comment from CX Productions. They say, thank God the studio is back. It's been a rough few days,

but it's a rough few weeks, actually. They tried to stop us with those fires, but they couldn't. And then we had another one question. We thought we needed energy infrastructure to support new AI data centers. We actually needed data centers to justify the energy.

infrastructure. It's more of a statement potentially than a question. But clearly not Jevin's paradox billed. Yeah. Get up to speed. Yeah, I don't know. I think even with all the deep seek stuff, every hyper-scaler is still saying we still need to build data centers. We still need to have all this CAPEX and the energy requirements of the way. that, you know, on BG squared, Brad Gersner, I was listening to their episode from last week, and they were talking about how, one, the way that Open AI is projecting their chip demand just from NVIDIA, it would be like a third of all the chips that are sold this year by NVIDIA. And Abilene, Texas, where this new data center is going to be, is definitely doesn't have

the energy infrastructure today to support that kind of, you know, just energy demand. So yeah, we'll see. It's going to be exciting. But it does seem. Yeah, I think Deep Seek, just based on, you know, letting a few days breathe, it's like it changes the economics of the B2B market. It might change the economics of the consumer market to some degree. But everyone is marching forward towards gigawatt training runs and then 10 gigawatt training runs.

Yeah. Nothing's changed on that front. People want the best possible foundation model to build the other stuff on. top of and sure if you can get those optimizations let's see what happens if you run the super optimized model on the thousand times much more hardware and that just might speed things up uh cool let's go on to more timeline speaking thank you everybody before that thank you everybody for submitting uh questions in the chat if you're listening to this on the rss feed we're now live streaming yeah daily so you can add questions live or just dms post and we will get to your questions asap yeah so

Simp for Satoshi says Deep Seek went viral outside of Twitter because of the Elon like and Zero Hedge retweet of my Regera meme. This broke containment and I had a lot of Normy friends in L.A. tell me they saw this post on a bunch of TikToks. Check the app store before this tweet and after. Deep Seek went from 31 to first. And the post is Sam spent more on this than Deep Seek did to train the model that killed OpenAI. And he posted it January 24th.

Okay. I don't even want to respond to this. posts with a video. So Ben, don't, don't respond to this because I generally, I met some for Satoshi. I think he's a sharp dude, but he is the biggest hater of opening eye. And there's a lot of, there's a lot of assumptions in here. One, I think everybody sort of understands that they probably spent quite a bit more than what they're saying. Two, you can't claim credit that you're one post when everybody had viral posts this week about deep seek. I don't think you catalyzed it to go number one in the charts. It was very clear that it went number one, like in a very

inorganic way. They were just buying, you know, downloads because this started happening Sunday and it didn't really hit the mainstream until this week. So this is either humor or humor. He could be joking. I don't know. But we're putting it in the truth zone. In the truth zone. And, uh, but I do love just posting all the time. And then if something goes viral or something happens, just be like, look, I'm responsible. I'm responsible. You know, I'm the one that posted. Yeah. Oh, yeah.

I mean, he very, so Simp very clearly wants to be credited. Yeah. For taking down opening up. You know, you, you, I think on the Monday show, when Nvidia was down 15%, you said on the show that you bought. I did. And then what happened the next day? It ripped.

7% pop. Yeah. You need to take, you need to take full responsibility for that. I, when I did that market buy on Monday. Exactly. People said, whoa, there's some whales. There's some whales that are buying. This is not financial advice, we joke around here.

And I not, yeah, we never give financial advice. I did, I did post, it was so funny because like Sunday night, Sautja came out and he's like, Jevins, Jevons Paradox, all this stuff. And I posted in a sort of joking way. Just imagine being unfamiliar with Jevins Paradox on a day like today. Because it does seem like, well, like, part of it's like Wall Street is not always betting on their beliefs. they're betting on about the future. They're betting on what they think the market's going to do the next day. Yep, yeah, yeah.

And so even if you have these very intelligent analysts in Wall Street that understand Jevin's paradox, it's like energy as energy gets cheaper. We use more of it. Yeah. Things like that. They can understand that and still short the market. Totally. Because they expect it to go down and they're just trying to make money that that day, right?

They're not thinking about five years from now or they're not really, they don't really care about long-term cap-ex implication. I mean, do you remember what I said about Mark Zuckerberg in October of 22? I think I said he was goaded and he was going to go on a generational run. And then look, Joe Wisenthel says meta is up 659% since November of 2022. Look, I must be responsible for all that. Wow. Yeah, it's amazing.

Yeah. But I do. Yeah. So make a, post something and then wait for things to happen and then claim credit for. Take credit for. Take credit for sure. And then if anyone comes for you, just say, I was joking around. I would never take credit for something as big as that.

Euber's, bro. You're built different. But I do think this is a crazy stat. Almost a thousand percent in two years. Like, I mean, this was the bottom. Like, everyone was super bearish on meta at this point. And that was because TikTok, TikTok was doing well. Everyone thought they overinvested in the Metaverse.

The Metaverse was a big thing. And the AI thing hadn't happened yet. So this was pre-Chatsypti. ChatGPT dropped in like December of 22, something like that. And so... And the other thing that can't really be understated, though, was that Gen Z was spending more time on TikTok. Totally.

Yeah. And there was no real like, oh, ban TikTok, anything. So the risk was like, hey, you, yeah, you, your next generation of users prefer another app. Yep. Yep. So there were a bunch of bear theses. Ben Thompson, he gets to claim credit for this because he was writing about how it's oversold and how meta was due for a huge comeback because the core business was just so solid and was just throwing off.

The greatest ad platform of all time outside of Google and ads are the best business model. The other thing people were really worried about was app tracking transparency, the iOS update. And everyone was saying that's going to screw meta's ads and their ability to track users with cookies and stuff. That was Tim Cook trying to take down small businesses. Yes. And who did it wind up hurting? It wound up hurting small businesses and every other ad platform except for meta. Because meta has so much data that they can actually run AI across their entire data.

They were impacted, but not to the same degree. And so they don't need to know that you literally clicked on this ad, clicked on this link, came back a day later, and then converted. They can just look at all your traffic across all their network. And they're listening to everything that you say. Obviously, obviously. And so, yeah, Ben Thompson. Yeah, to help prep for our Valentine's gift guide, I just want to say,

hey, John, I'm really looking to buy some new gifts for my wife for Valentine's Day. Send me some stuff. After the show, after the show, I'll scroll and I'll get some good. I told you that my wife actually uses Instagram that way where if she's shopping for a product, she'll visit a couple of the websites, not just scroll a little bit, not really actually shop, and then just be like a king, like come to me on Instagram. And then all the retargeting ads will come and be like, I'm ready to hear your pitch, but you have to pay.

It didn't feel like it always was that way. Now they really follow you around. Well, they follow you around. But the other thing is it's so, this is why it is such a bloodbath of competition. in e-commerce right now because if you go you know Rora right like we sell water filters if you go to rora.com yeah right now rorra.com yeah uh and browse around for a second and then you go back to instagram you will be start being served a bunch of ads from our competitors granted you have a much better product much better tested it's more effective so we're confident that you'll come back to us

yeah but still it's wild that meta the company we spend more with than anyone except basically our own suppliers is like working against us in many ways. It's like you knew about Rora. Now we're going to show you all these other alternatives to it. It's a free market. Welcome to libertarianism, Jordy. Yeah, yeah. Ross would be proud. Cope harder. Let's go to David Holes. I'm not coping. We're confident. He says, I worry an unsustainable need to raise billions of dollars has broken the discourse around AGI ASI timelines. There's too many incentives to say you're on the brink of a breakthrough and too many incentives for everyone else to say they're not far behind. Interesting.

We do see a lot of this where there's a lot of the AI foundation model leaders who are saying AGI is so close, ASI is so close. Just one more training one, bro. Just one more. It's almost like, you know the, you know the scientists how they always want to build the the large Hadron Collider and there's that meme of like the scientist being like just one more collider, bro. Just one more.

It's one more collider. I just want to go a little bigger. You can trust me. And then it's like, I guess we do get cool foundational science out of it. But it's very hard to say like what has the large Hadron collider done for me lately? Yeah, yeah, yeah. I guess it discovered like quarks and stuff. But like how does that really change my life?

I don't really have like, you know, teleportation or like supersonic travel or like, you know, space exploration because of the colliders. And so we're seeing the same thing with the training runs. I mean, I am very, I want them to do it just because, yeah, you know, light some money on fire. Mesa and let's see how good is GPT 5. What can GPT6 do for me? If it's the same old and it's still messing up, okay, cut it then, but half a trillion, let's let it rip.

Trilly. Let's let it rip. Let's go to Alex Wang. He is doubling down on the deep seek hate. He says, what does deep seek R1 slash V3 mean for LLM data? Contrary to some lazy takes I've seen, deep seek R1 was trained on a ton of human generated data. In fact, the deep seek models are setting records for the disclosed amount of post-training data for open source models.

600,000 reasoning data, 200,000 non-reasoning SFT data, human preference, RLHF data set of undisclosed size, human process synthetic data for cold start data. According to Chinese AI engineers, DeepSeek actually values data annotation even more than other Chinese labs with the CEO personally labeling data for the model. That's kind of crazy. This reminds me of Carpathie, who used to spend a quarter of his time labeling at Tussie. The Deepseek V3 paper has even, even has a dedicated acknowledgement section for data annotation. Very interesting. Well, we don't know exactly what's going on there, but this is why I put out a call to action in Chinese characters. I said, if anyone on the deep seek team wants to spill any

highly confidential information to us, we'd be happy to discuss it on our podcast. So if you work for Deepseek, feel free to call in. DMs are open. Dams are open. Call in. Let us know. We'll get on We chat for you. Yeah. We'll talk it. I'll talk it over. I mean, Deep Zieg really is having an impact, even as an open source model. You can see Arvin Srinivasa from Perplexity has integrated DeepSeek R1 and FACC set to do financial research on steroids. And so they've connected crunch-based data and FACC data with R1 so you can ask it to reason

through things. And this is really cool because a lot of times I go to chat GPT and I'm like, okay, we're going to talk about, you know, Coinbase today, give me their entire funding history. and it just doesn't have access to that data. It's not reasoning through it properly. So then I have to go to Coinbase, I have to go to Crunchbase, export all the data, copy and paste that into ChatGPT, pull that. It's very cool to see perplexity integrating those extra data sources

so that you can get better results. And I think that's maybe an increasing trend. Like you saw this with the operator launch where they had dedicated agreements and integrations with a bunch of different companies to get access, Instacart, to make it work. even better instead of just trying to brute force everything with a model that might not quite be there, might kind of mess up more. It's like, hey, yeah, just give us access. A lot of these apps and platforms that you'd want to interact with have anti-bought products. Yeah, yeah, yeah.

Kind of trying to prevent that activity. And so, yeah, I mean, using deep seek has been a little bit controversial from the geopolitical angle. Obviously, Mark Andresen's firmly on the side of like open source is great. Everyone should be using R1. It's a gift to humanity. I actually, we discussed this briefly on the show, but I asked on X, how sci-fi or tinfoil hat is it to consider the idea that there might be a Manchurian candidate buried within the weights of an LLM? So basically, you get this open source LLM out there and it gives you normal answers to everything, but then at some point it switches on you and starts changing the way it responds or doing

something. And is this like five years out or 10 years out or is this doable now? And Trevor Blackwell, I believe, he was one of the founders of YC, was like, it's possible and it's happening today. And I was like, crazy. I got like one response. It was like, this is the smallest post for my account. But I was like, this guy's very high signal. Like, this is, this is crazy. So I don't, I don't think there's any evidence that they did that. I don't think it's actually what happened here, but it is something to consider in the future. We got the hat ready. Yeah, it is, it is something consider. And I think we should do the same thing. Next version of Lama, Zuck, let's embed a Manchurian candidate that, you know, send that out into the world and convinces everyone to get into luxury

watches. Brought to you by Bezal. Yes. Let's go to Neer. Neer says, why did Deepseek go viral? TLDR, class resentment, anger, and especially shot and fraud. Very little actual app usage in comparison to the above. And so Neer says, my memetic model has really failed with Deep Seek and my analysis of why I failed is so is so failing is also failing does anyone who studies this topic have any novel takes and so near has categorized the the first week of the deep seek launch showing the increase in queries for deep seek on x the posts go from 35,000 to 66,000 to 114,000, 191,000 and then explodes to almost a million on quote unquote black Monday the 27th. Like 50,000 of those were us.

And then the iOS rank, it went from 400 all the way to one. And then there's mainstream viral videos, CNBC, Forbes, CBS, CNN, et cetera. And so NIR really did a crazy deep dive here in a spreadsheet of kind of tracking how this model went so viral and what happened. And it really is a very emotional issue for a lot of people. And it spans the gamut. Like clearly Taylor Lorenz was saying, like, yeah, like go China because she just has this like funny bit that she's doing right now where she's like a

full-blown communist and like loves everything China does and let's send her over there and then but then there are other people that just love open source and and they don't care about the geopolitics at all and so they're just like I don't like that open AI doesn't open source things anymore I'm an open source purist and there's a lot of those people so they'll amplify it for that reason then there's other people who just really are sam altman haters and they're like anything that hurts sam I will amplify and I'm stoked about. Then there's other people who are just like, like, you know what?

I just want to use a cheap AI model to create my grocery lists and my recipes. And this is the best app and I was on the free plan of chat GPT. It wasn't reasoning at all. I didn't have access to 01 whatsoever. This one gives me reasoning. I'm in for that.

I love this. Thanks. It's a cool free app. And so there's all these different like... Other people, Matters. Bat, bat enthusiast. If you ask Open AI,

can you make me a recipe for a really delicious Terrible. And then also, I mean, there's even people who are like anti-NVIDIA and anti-hyperscaler, and they're like, I've been sick of, of Nvidia making all this money. I've been sick of this story about Microsoft, you know, these big tech companies. I just don't like the big tech companies. And so if this hurts them and I see their stocks down, this is awesome. And so there's all these different effects going on that were very unexpected and not priced in to NIR's model, which I think was interesting.

Lorenz was not priced in. Not priced in. Let's go to Joe Wisenthal. He says, I think one thing that's fun about AI is that it's brought back tinkering. For years, it's been impossible to play with the computer, especially smartphones. Everything felt so sealed off for most users. Maybe this helps revive a new era of tech literacy. And I completely agree with this.

I think this is so cool. It's not just the cursors and the cognitions and the devons. It's just this ability to have a new tool. I mean, I remember when I got to Silicon Valley, it was the slithes. Lomo, Co, or whatever. It was like social, mobile, and local with the trend. So like, 4Square was at the intersection of that. But there were really, really strong APIs for Google Maps, for delivering software

on the iPhone via an app store. Nothing was locked down at that point. And there were also social graphs. So very quickly you could say, okay, for my app, I want to combine Google Maps with your social graph and I'll plug into the Google Maps API and the Facebook API. And boom, I have something today. And instead of needing to build these big services, now everything's locked down walled garden. You can't really like integrate these in the same way.

And so it's been there was like that narrow window of opportunity for like. Yeah. And LLMs are so magical across all these different app experiences and model foundation models that there's this broader excitement. X has always been willing to try a new social app. Right. Something pops off. And they'll be like, oh, everybody will download it and play around with it. usually it dies fairly quickly.

But right now you have a more, the general, so many hundreds of millions of people have tried ChatGBT and been relatively impressed with it. Totally. Even if some of them are using it in relatively simple ways, like, you know, asking it questions like they would Google, that there's this broader willingness to just try new products and experiment. And this is why I think the whole like anti-chat GPT rapper meme is so harmful to the next

generation of entrepreneurs because I imagine that in a few years, like building a somewhat viral chat GPT wrapper, even if you get killed by OpenAI, even if you don't make a lot of money, is going to be a massive bull signal for entrepreneurs because you're going to have someone who, you know, yeah, they took a couple of computer science courses. They learned a little bit of Python.

They installed cursor or Devon, and they started spinning something up. And they just made a small little niche app using R1 and these free open source LLMs. They got it up. It delighted a few users and it was an end-to-end product. Good project. And then when they find that business opportunity that's maybe boring, like rolling up HOAs or whatever, it's going to be like, well, I know that this person, I know that this founder can build. And so I'm totally ready to put the new sneaker. It's the new like trading sneakers. Exactly. Exactly. Exactly. And so I am I'm very excited about the era of, of AI and what it means for the tinkerer and just having more tools. I mean, even just in 2012, Heroku was a

a huge deal because it allowed you to write basically just Python and then click, you know, you would push your Git repo to GitHub and it would automatically deploy it on the cloud. So you didn't need to install any Linux servers, even AWS, which was huge because you didn't need to rack the servers yourself. AWS was a huge increase in just developer productivity and like hacker productivity. Heroku was another step because you didn't even need to say, okay, I want to provision a server. It would just do that automatically. How big is that business today?

Heroku. Yeah. I don't know if it's been like fully, so it was bought by Amazon for I think almost a billion dollars. It was a YC company. It was one of the first major liquidity events for a YC founding team. And so they were, when I got to YC in 2012, people were like, oh man, like Heroku, that's like a sick outcome.

Like that's a great business. Like they did so well. And then I'm sure it got folded into like a million different projects because- Inspired it got to a point where half of the YC batch would be dev tools. Yeah. many ways because that one event really inspired. Yeah, you could start on Heroku. You could even, now they have even simpler products called like serverless architectures.

Are you familiar with any of this where like you don't even need to write the server? It's not, Heroku would like spin up a server and keep it on all the time. And it would kind of go dorm in every once in a while if you're on like the free tier. But generally the server was on and if you visited the web app, it was there. Serverless is just sitting there and only when the query comes in, it spins up the server immediately runs the query that you want or the code that you want and then gives you the answer. So that was like super, super cheap because you could throw something up. If you had no users, it was zero cost.

You only got charged for the queries. Whereas Heroku would give you like one free one so you could keep something small up. But if it's scaled at all, it would still be expensive because you'd be running it when you didn't have demand. There were some like automatic load balancing stuff. And then as you scale up your business, you go to, okay, now I'm using all of AWS's like bespoke tools that I'm provisioning like dedicated servers. and then maybe I'm building a $500 billion data center in Texas, you know.

Let's go to Palmer Lucky. We talked about this a little bit, but he is also beating the drum on Deepseek and the controversy around Deepseek. He says, deep seek is legitimately impressive, but the level of hysteria is an indictment of so many. The $5 million number is bogus. It is pushed by a Chinese hedge fund to slow investment in American AI startups, service their own shorts against American titans like NVIDIA and hide sanction evasion. America is a fertile bed for siops like this because our media apparatus hates our technology

companies and wants to see President Trump fail. We have so many useful idiots uncritically reporting Chinese propaganda as fact because on some level they want it to be true. They love seeing hundreds of billions of dollars wiped off the market cap of our largest companies. And so there was a little bit of back and forth in the replies. to this. Palmer kind of clarified a lot of people were saying oh well you know what are you talking about like they didn't they didn't drive down the stock price with a short it's a multi trillion dollar company like they wouldn't have been able to do that but what he was saying was that they might have just had like you know a billion dollar short position and boom they just make 150 million dollars and that's a

great trade yeah he wasn't saying that like all of the downturn was caused by short pressure no the whole I mean look at Hindenberg research they're they're taking a position releasing a report and hoping that it goes down, they're not capturing the entire loss associated with that sell-off, right? Yeah, exactly. And I think the things that make that sort of conspiracy theory or broader theory somewhat believable is it was clear that the app on a random Sunday when nobody outside of X was talking about it just happened to go to the number one app

in the entire iPhone app store. with a Chinese character developer name that just screams. Like even TikTok doesn't use Chinese characters in there. To be fair, a lot of the app chart is very gamable. And I believe that that screenshot was in the productivity category, which is also very slow moving because there aren't that many viral apps that are gaining. And the chart is based on momentum. And so if you're getting double the amount of downloads every day,

you will be number one, even if a thousand times more people downloaded chat GPT that day, the momentum is what matters more than anything else. And so no one really knows the algorithm, but it is very interesting. And honestly, it's been weird because in the paper, they say we don't count R&D dollars in that $5 million number. We don't count the servers that we had from a previous training run in this number. We don't count the data. Singapore, GPS.

So there's a lot of things that they didn't count. and they admit to this, but it's been kind of uncritically reported in the U.S. as just, oh, it's $5 million. The point is that it's still. Like going back to his Palmer's original point, there's so many people that want to see tech fail.

Like the traditional media hates tech still. Yep. There's a lot of people that want to see Trump fail. He just announced Stargate. You know, people want to position him as silly or dumb or whatever. Just give him an L. Yeah.

And so there's a lot of people that just take everything at face value. something very like memetically or like inherently viral about the fact that the number five is both in the deep seek announcement and in the Stargate announcement. So it's literally five million versus 500 billion. It's exactly 100,000. And I think if it was like four million and six million, six, six hundred billion, it like wouldn't have clicked as much in people's heads. But just the fact that it was exactly a hundred thousand dollars, uh, exactly a hundred thousand. Um, exactly a hundred thousand. times cheaper to do something like impressive and it's like what's going on this is such a big just just also disingenuous though because presumably the stargate level training runs will provide

even the theory the whole reason to do it is it will provide models that are much better than oh one and much better than deep sea and yeah I mean the original gpt4 training run was 500 million so 100 x so they were able to it seems like they were able to get 100x improvement in efficiency if the number is even close to being right. Definitely a 10x. Even if their numbers off by 10x, they're still 10x cheaper. But we don't have numbers on what 01 cost to train

because that is not the same as the R1 model. It's a reasoning model. It's a different thing. It's more about the algorithm than the training data. It's a smart marketing strategy to let everybody compete on advertising how much they're spending on training runs and then take the exact opposite approach.

It's like a fitness, trainer online being like, everybody's telling you to work out six days a week. Yep. I've got a product that lets you work out one day a week. Yeah, exactly. It's just like so contrasted. Yeah, it's similar to like the friend.com. We spent our whole seed round on a domain.

We spent $2 million on this domain. That gets a story. And then also I built a massive viral success with a $10 domain. Anyone can do it. That's also interesting and clickable. But just going with something in the, like, that meets expectations. Like you have to break expectations if you're going to go. viral. Well, let's go to some good news. Did you see the Boom Arrow news? Yes. The supersonic plane broke

the supersonic barrier, broke the speed of sound. Mock 1.1. Very good news. And Nicole Wiskoff says, my husband from the other room, babe, can you believe the founder of Boom Arrow used to be a product manager at Groupon? Unbelievable. I got to frame his LinkedIn and put it on my wall. Serious congratulations, big fans over here. And I love his, did you read his LinkedIn? in it says at Groupon he was there for two years, two months. And he says, there's nothing like working on internet coupons to make you yearn to build something you truly love. It is such a fascinating story for him.

It's been over a decade at Boom. Really rough go, hard down round, lots of regulation, tons of red tape. When I see the announcement, everyone in the YC world was going crazy because I think they just did the new round as well. And there were a lot of people that doubled down. I think Joe Gebbie it was in and a few other people. And on the face of it, I was like, this isn't that impressive. It's just a plane going Mach 1.1.

Like we did this in the 70s. And like they had to caveat the record and saying it's like the first it's the first civil plane that's domestically produced that's broken the speed of sound because the Concord was a French project. And so I was like, this doesn't feel as innovative as it could. It's not like it's like Mach 10. Like we'd never done that before. But at the same time, it is impressive when you think about just all the red tape that

exists these days in these hard tech companies. And it just shows that they actually broke through that. So it's almost more impressive as a business and regulatory story than an engineering story. To me. Yeah, yeah. I buy that. I mean, they had 50 people on the team. Yeah.

It's crazy. A group of 50 people came together and said, let's build a plane that can go, let's build the fastest possible plane we can build. It's crazy. It's amazing story. And it, I mean, if you look back, it's like we, as humanity, we gain the ability to build things. And then if you don't keep the pressure on, you lose the ability. Totally.

If we ever get to a point where we hit hour three on a podcast and we decide, oh, I'm tired. I need to go to the bathroom. Just remember. Just remember. You got to keep the pressure on. But, but this is like Elon when starting SpaceX, he wanted to buy Russian ICBMs because he was like, well, this is like this technology that that they had. that's still like it's viable and we can use it or learn a lot from it. So I think it's great.

And it's such an amazing story to me that many, many investors would have left this company for dead at the down round. And I'm sure many did and are now kicking themselves because they didn't believe in the company when it was at its roughest point. They turn around. So it's just a fantastic story. They're going to make documentaries on this. It's going to be awesome.

Jason Carmen already has. I'm pretty sure. But we need a new one. We need the full hero's journey, the bottoming out, the recap. Yeah, the really dramatic documentary that they're going to make is the Hermius versus boom story. They're in different markets because one's military and one's commercial, but there are these two founders. They're clearly racing against each other to build hypersonics and supersonic planes. And there's like a lot of similar engineering that goes into it.

And I think if you start telling that story as a horse race, it gets a lot more exciting as opposed to just the story of one company. Yeah, we're going to the middle. We're going to the Middle East later this month, or sorry, later in February. And I wish we could be going mock one. That would be amazing. Nice. Jack Raines says the Screlli versus meme coins has been a great Twitter side plot in Q1. And Martin Screlly says, ha, ha, ha, ha, barg coin and dog whiff hat tokens for R words.

What happened? It went down, it's only down like 30%. That doesn't seem that bad. I expect this stuff to go down 99%. Yeah. That's the thing. Memes have some intrinsic value. The FDV is $793 million, which is far too high for something so stupid. But people do like...

Hey, it could be the future of finance. Don't judge it until you... I mean, that's what I'm with Dogecoin. It's too bad. But it is funny that Screlli's been out there giving out lots of financial advice. Let's move on to... It actually is a... It is a really good bit for him to go heavy against just calling out things that he identifies as stupid or scammy.

He's going after. He goes after these, like, bio. He's had a couple bets where he'll post something Sunday night, short this company. And then maybe it's a butterfly effect or maybe they just, like, people figure it out. He has a huge, huge audience. So after Satina Della said, Jevin's Paradox strikes again and posted. the Wikipedia to Jevon's Paradox on a Sunday night or something because he was worried about the stock going down.

Our Karazian over at Ramp, their economist that they just hired says, we can see this in the TriRamp data. Here's data from a small market, a small sample of our base. Some thoughts, though, about whether the market reaction is justified or not. Companies purchase more tokens as the average price per token collapsed. And so you can see even just over, I guess it's just two months or one month, The average price per million tokens of LLMs dropped by, I guess, a factor of three. And the number of tokens purchased increased on ramp cards because companies are buying more.

And it makes sense. There's so many places where you want to stuff in LLM. And if it's super cheap, it makes sense on a per transaction basis. It's like, yeah, my LTV of a customer, 100 bucks. I can't afford to spend $10 on LLM inference per customer. But if it's a dollar, sure. Yeah, why not? Yeah. I got a, I got a promoted post that's in the form of a review.

Okay. Like if a podcast called All In was good and interesting. Unnecessary, unnecessary chab. We like, it's our favorite political podcast. And Jeremy Sun says, if you want to understand what startups are actually like and immediately tap into the unhinged schizo-gundo core vibe, this is the show for you. Who do you trust a couple of dudes in the, in the arena building or a guy who invested in Uber once and won't shut up about it.

This review brought to you by Symbiont AI. Symbiont helps keeps your grandma from getting scam through a free AI security app that reads, email, text, and more to identify scams in progress, stops them before they happen and cost scammers money by wasting their time. That seems super valuable. I need that. Jeremy, this is awesome. We covered a couple scams, the QR code scam.

Yeah, so it's S-I-M-B-I-A-A-A. nt.a i symbion and uh uh i need to get this for my family members uh i would love to get a little printout at the end of the month maybe actually physical piece of paper great if you can make a printer integration germany that'd be awesome do you want me to read the next one yeah we'll go back and forth uh lifelong bureaucrat great podcast although it has nothing to do with my line of work doing tps reports and publish and pushing paper in the federal bureaucracy it gives me a sense of how the other half lives and better yet the 1%. Love to listen and then drop the lingo on my GS-14 supervisors.

Don't know what that means. Get a great way to seem value add in tech native. Since no one in my government workplace understands any of these words, Bangor, Petechfili, misspelled. It's two P's and then E. Size gone. They think I must be working in the IT. The net result is they leave me alone,

and then I have more time to listen to the technology brother. podcast. Let's go. Let's go. Let's go. Let's go. Let's go. Should leave an ad for your agency.

Well, yeah, we love the, we love the government. Well, it's an ad for the government. It's an ad. Pay your taxes on time. Yep. And don't break the loss. Yep. Technology Brothers, this is from Nick Stewart.

Technology Brothers gave me the strength to quit wearing Lulu Lemon and the courage to start using nicotine again. Thank you. This review sponsored by Angelus. Attention, beloved capital allocators. Are you still working at a large tier one venture firm and letting an investment committee stop you from backing the next Lucy. Are you sick of getting Carhart merch? Wow, Savage. Do you wish you could avoid the internal politics and get more personal upside from your investing? Well,

you already know that some of your smartest colleagues have left to launch their own solo fund. What you might not know is that the smartest ones did so using Angelis full service fund offering. Angelis takes the operational complexity out of running a venture fund, helping you focus on doing what you do best, making great investments and running various parts of the federal government. Start your Sholo GP journey with Angelus today and we'll help you hit the size gong at first close. Make sure to tell them John and Jordy sent you. So anyways, I use Angelus all the time. I've invested in funds on Angelus.

One of my very close friends, Jack recently set up a new, I think it's like a $40 million fund on Angelus. And yeah, I'm an LP in it. It's an awesome experience. And yeah, Angelus is the best. I used to have to compete with them. I did not enjoy that. Now I'm just happy to be a user.

Fantastic. Thank you to Nick for the five-star review. Let's move on to Calmede 805 says, The Missing Piece Into My Chad Diet. Last year, I committed to overhauling my diet and becoming a true Chad. Existing Nutrition Apps weren't cutting it, so I took matters into my own hands. I built your body AI now available in the Apple App Store or at Your BodyAI.com,

the fastest and easiest way to track your calories and macros. But there was a missing piece, my information diet. Fortunately, I discovered technology brothers. They consistently deliver the essential tech insights and entrepreneurial inspiration that my mind craves. With both body and mind now fueled by optimal nutrients, I'm a new man. I'm down 12 pounds. I'm pushing up two plates on the bench. My wife and kids respect me.

Let's go. I couldn't be more excited for what 2025 will bring. All right. We've got to check out your body AI. Very cool. and I just love how everybody that leaves a review, like it crushes the ad. They crush the review. They're doing best when it's all integrated.

Oh, yeah. It's fantastic. It makes me, we already love ads. We love reading them, but you guys are, you guys are brilliant. So keep it up. If you haven't done it already, go leave a five-star review. And, I mean, this is untapped alpha. You can do it twice.

There's Spotify and Apple. You can create new accounts. We might have to set a limit on a per company basis, but you can do it. Do that like crazy. We will get flagged as like spammers basically. They'll be like, this is the next deep. I just, I just want to, yeah, yeah, if you, it's really funny to think about an Apple, like, podcast, like moderator.

Yeah. Yeah. Yeah. And, and, and it's like, starts removing these because it's like, it's probably against the terms. It's probably spam. But we support it 100%. Yeah. So, thank you.

So we're, we're capping it too right now. One on Apple. podcast one on Spotify mix it up um you can also you can leave an unlimited amount of ads in the YouTube comments true true yeah and we post like we're now posting five five videos a day yeah like subscribe ring the bell go on open AI make an operator bot that just spams the comment section no no just generate the text the cheap LLM and then copy paste yep it's the artisanal way of spamming or in John says he's about he's about to cross a thousand short form videos this is all it took to achieve

financial freedom mid career and never have to work for someone else ever again kind of ridiculous when you think about it so congrats to oren on hitting a thousand yeah i just thought this was cool so he started like i think really leaned into reels i've followed him on instagram i think he's also on ticot and some of these other platforms and he's just been dialing in not only building his business in public but like helping people understand these like short form platforms and really, you know, built a pretty, pretty awesome community around it. So I think that that's the amazing thing. Five years ago, it was super hard to get attention on Instagram. You could get attention, but it was super slow. It was like very hard to grow on Instagram.

There was a period of, there would be comments everywhere. I had a bot. Yeah, yeah. Yeah, so it was just like really tough. Yeah, it was spammy, tough time. And then these short form platforms now make, or it's much more democratic system where they don't like some if you're an account and you have 100,000 followers on Instagram and then you stop really posting or leaning into short form you're pissed because you're like I post and a thousand people see it yep but it makes it so it levels a playing field and you can have 100 followers and you post a good video I still will be on reels sometimes and I'll get a video that has like no views from like a new account I'm like sometimes it's super relevant other times it's total slop testing it out on you it's a free way to to change

your life. Well, we started with Elon and Starlink and SpaceX, bringing internet to your phone from the stars, and we will close on Elon Musk and SpaceX. The president has asked SpaceX to bring home the two astronauts stranded on the space station as soon as possible. We will do so. So very good. Yeah, I mean, it's absolutely wild that they've been up there for a very long time. And seemingly, Elon was very willing, excited, and able to assist in the effort and the Biden admin had snubbed him at the like EV car event that they hosted and just like ignored Tesla, ignored that he had the most dominant, you know, private space company or space, you know, force in the history of mankind. Yeah, it's crazy. And the debate over whether or not they were technically stranded was like a big thing.

Oh, well, it's just a safety thing. They could go back or they could go back anytime. They could go back. They could go back. And then it was like, oh, they actually can't go back on that Boeing pod. They got to go back on the SpaceX thing. But when will they go back? Do they do the astronauts have a decent internet connection when they're up there? Can they use like reels? Can they use reels? I don't know. I don't know how fast they do like. Can they watch like brain rot slop? Or do they have like old office reruns? They're right up there with the, they just beamed down. They got black market. I think they have pretty good internet. I don't know. But I mean, the steel man here is that like, They were supposed to go up for like two days. They got to stay up there for like nine months. Like I think going to space for nine months. Maybe it's like a mutiny thing and they're like, we don't want to come down. We like it up here.

We like our little stars now. Yeah, yeah. We like our little. It is crazy. Elon's never been to space. He's like the space guy, but he's never gone. It's so crazy too that astronauts have to keep a straight face and just like it would be very embarrassing for NASA if they were freaking out.

Oh, yeah. sending help messages out like Biden want tweeting at tweeting at Elon saying at Boeing like stop like please let us but like Biden got to just leave the White House and go on vacation and they're sitting up there being like really like we don't get to go home until somebody sends a spacecraft I mean people definitely jumped on this because it was such like a pro-Elon narrative obviously but then it was true it was like okay yeah Boeing actually couldn't do it yeah UCLA actually messed up and then it was amid all the Boeing planes and stuff. It was just such a mess last year. Yeah, rough year for Boeing. Anyway, that wraps up our show. Thanks for watching. White Soot Day. Go leave us a five-star review, put an ad in the comments.

Thank you for watching live and leaving comments. We enjoy, we love podcasting and we love live streaming. Yeah. And we're excited to do this with your support for decades to come. And we'll see you tomorrow. Have a great day. See yeah.