I see a large IPM on the horizon. You're surrounded by journalists. Hold your position.
>> Overnight success. The place 3000. Double. Right.
There's misinformation. Clearing order inbound. Let's just roll.
>> We are surrounded by general. Hold your position. >> Get up. Trust the expert. We are experts. Founder Mel 5.
>> I see multiple journalists on the horizon. Stand by. UAV online.
>> Blaze. >> Double blaze. Triple blaze. Double kill. Fight is please team deathmatch. We are experts. Triple blades. That's just wrong. Right.
>> Hardy clearing order inbound. We are surrounded by journalists. Hold your position.
>> Strike one. >> Strike two. Activate golden retriever mode. Triple blades.
>> Marky clearing order inbound. UAV online. Very. I see multiple journalists on the horizon. Stand by.
>> You're watching TV. Today is Thursday, March 26th. >> We are live from the TVP and Ultradome, the temple of technology, the full finance, the capital of capital.
>> Let me tell you about ramp.com. Time is money. Save both. Easy to use corporate cards, bill pay, accounting, and a whole lot more. Let me also pull up the linear lineup because we have a monster show. Jord's back in the TVP Ultradome. I know you missed him. Uh we have Eric Goldman coming on to help break down the meta YouTube lawsuit. We're going to...
>> By the way, yesterday was the first time we've been live that I was not on the air and so I was watching on the third show.
>> How'd I do? Uh here we can drop the linear line. Linear of course is the system for modern software development. 70% of enterprise workspaces on linear are using agents. Uh give me a review. How'd I do?
>> Uh I thought you did fantastic. I was texting you in real time. I probably sent you like 20 something messages of like ask ask about this, ask him about that. It was uh yeah, somebody in the chat said I I was free soloing TV and it was great. But we really were down to our last few uh few associates here. Tyler was out, Jordy was out. Uh who else? Yeah, Nick was out. We we had we were on uh what do they call it? A shoestring or something like that. I don't know. But uh it worked out. It was a fun show. Uh it's not the same show without the soundboard. I was going to do this whole bit about like, oh yeah, I got it dialed. You can just we don't need you on the show anymore. And then of course the fans would be like, "No, bring back Jordy." And I'd be like, "Yeah, okay. He's back." Uh, but I can't even joke about it because we need the soundboard. I only got...
>> It really is musical. It really is. Imagine if you were I was waiting for you to just be spamming that.
>> Just sitting there having a serious interview.
>> Apparently, we have some audio issues.
>> Okay, we will bear with us. We're working on it. But I can go through I can go through a bit of our guest lineup today. We have Nema some highlights...
>> friend of mine uh he is a founder of salt and stone they uh recently uh exited the business as of this week a massive you know multi-hundred million dollar deal very excited to um get to understand the business better have John McNeel coming on he's an author to talk about Elon Musk's five-step algorithm for scaling Tesla...
>> uh then we have Ki from linear on how linear is evolving in the AI age. And then we have a bunch of different founders including Mikey from Sunno uh coming on.
>> Then we got back to back Zachs. Zack Cananter, Zack Pereé going back to back Zachs. That's what I like to see. Uh well, the big news of the day of the week, this has been going on all week, is that ARM, the intellectual property developer that creates uh intellectual property designs for CPUs, is now getting into the chip game and they got a big ARM pump in the stock market. So, um the company is up uh 15% over the last few days on the news that they will sell their own chips. This is new for ARM. ARM is a very old company. Uh fascinating history. I actually made a 25-minute YouTube video all about the history of the company uh back in 2023. Um but we'll recap a little bit of it today. Uh but uh they normally just license out their intellectual property and that is a phenomenal business. 97% gross margins. 97% gross margins. Yeah. Let's give it up. Let's give it up.
>> That's amazing. Um and you know it's a big business. Four billion in revenue last year. Nearly 800 million of net income. Uh they're this new move they're expecting to ramp revenue uh to $15 billion by 2031. So they're expanding the market significantly. Now margins will be different. Uh but the market cap for ARM is now around $166 billion. So it's a big company. Uh it trades at a very high high...
>> four billion last year currently trading at 165 billion...
>> but very very high gross margins. Uh so this is a big shift shift in uh strategy. ARM's not an AI loser by any means, but it hasn't gotten the attention that other GPU makers have received like Nvidia. Uh CPUs are far from dead, though. In fact, we are in we are currently in what seems like a little bit of a CPU crunch. Intel can't make CPUs fast enough. Nvidia is starting to sell their Grace CPU that goes with their Hopper. So you get the H100 and you pair it with the Grace CPU. You get the GPU and the CPU all on one system. Well, now you can buy the CPU by itself if you're CPU constraint. So you don't want to just be GPU rich and CPU poor. You got to be rich in both camps. And a lot of companies are jumping in to fill this gap. So um agents and a lot of this is because of agents. Agents need CPUs. You need to fill the GPUs constantly with new data and tasks. Also, all of the agents use CPUs to make web queries, search the web, run Python, spin up web servers, interact with anything. Uh we've just seen, you know, we talked to some folks at semi analysis about this like uptime is going down for non-GPU accelerated workloads. Like you go to a SaaS product that is basically just a web server that's running on a CPU somewhere in a data center and you're like, ah, this thing isn't loading today or like the there's downtime. And a lot of that's just because we're writing more software. We're using more software as a society, as a country. And we need more CPUs as well as GPUs, even though GPUs are like the the hot thing to talk about. So ARM has a very interesting backstory. Um, as I mentioned, you can go watch my 20-minute 25-minute YouTube video about it to go all the way back and this all the precursor companies. Uh, it's a really long story, but you can think about it basically as a joint venture between three groups. Uh Apple, Acorn Computer, and VLSI. Uh and so they needed to design a low power CPU for mobile devices. Before phones, you know what was hot, Jordy?
PDAs. And have you ever seen a PDA, Tyler?
>> You've never seen a PDA?
>> Okay. So, before the iPhone, you know, everyone had flip phones, but before...
>> It's not a pager. You don't know what this is either. Wow. So there's something called a Palm Pilot and basically it was it was like a like a you know iPhone size screen and it has stylus and you could write on it and you could make notes and PDA stood for personal digital assistant and I feel like that brand is just itching to come back in 2026. Like that's what so many people are...
>> everyone is working on a PDA.
>> Yeah. Personal super intelligence assistance. There's a whole bunch of things we're building PDAs folks. Uh they just live in the cloud and it's not a physical device but these PDAs back in the 90s were physical devices. This was post-pager pre-smartphone and it was the it was the device that you carried to take notes and do things that you do today in apps you would do on your PDA but it wasn't always on internet connected none of that it was uh it was it was a Palm Pilot and this was like a fantastic business for a while but there were a lot of problems with this because for the first time you needed a CPU that could live within a plastic like shell basically these were like you know plastic enclosures needed to run out of battery. It needed to be need needed to be able to do some things, not crazy compute. But the CPU industry in the '90s was very focused on mainframes, servers, desktops. There were a few laptops popping up, but there was not the mobile phone revolution that's happening now where you have Apple silicon chips and those are of course based on ARM architecture. But that's where all this came from. They they people said, "Okay, we need a lower power chip that can actually run off a battery, not overheat and melt the plastic, do all of this, and it's can be somebody can be can power a device that you carry with you as a personal digital assistant, a PDA." And so...
>> Chicken in the chat says, "PayPal started on PDAs."
>> That's right. That's right. So, PayPal started originally the idea was uh these PDAs had they didn't have like tap to pay or anything like that, RFID. They had uh basically the same device that you'd see on a TV remote. So IR uh and it would flash a light that could be seen by another sensor and if you flash the light at a certain rate, you can send a message. You can basically it's like advanced what's what's that SOS thing. Morse code. It's like a more advanced version of Morse code. And so you could send a specific packet of information from PDA to PDA. And this was the original idea for PayPal. That's correct. That's a great great piece of lore tech lore. So um the so you know you ARM starts to you know build these uh later there's uh Robin Saxby who was the CEO of ARM at the time uh he wanted ARM to become the global standard for CPUs. And so in the 90s there were lots of different CPU makers, lots of different architectures. There's this whole CISC versus RISC debate. There's the x86 architecture that in uh that Intel pioneers. Um and uh slight differences in architectures can limit interoperability. You see this with what's going on with Mac where a whole bunch of applications have been have needed to be rewritten for Apple silicon because previously you would have an Intel Mac before we got to the M1, M2, M3, M4, M5 chips. Those are Apple silicon. Those are ARM based. And sometimes you'll go to a website and they'll be like, "Oh, do you want to download this for your Mac? Like what do you have? Do you have an Intel Mac or do you have an Apple silicon Mac?" Well, it's important because when you write the software that runs on the Mac, you need to use specific instruction sets. Now, there are ways to abstract that and run on either, but there are lots of pieces of software that interact with the CPU at a low enough level that they need to be aware of the instruction set. So, ARM sets out to be the global standard for CPUs and they create the ISA, the instruction set architecture. Uh, and that ultimately let Apple design their own chips. Uh but within the architectural guidelines set forth by ARM. So Apple pays a license for to ARM for every chip Apple sells. It's a very small license fee because Apple does a lot of design work. They do the manufacturing TSMC fabs it and like there's a million other pieces of the value chain. But for this one little slice they have to pay ARM and ARM just takes that and says thanks. Like cool like you used our intellectual property successfully.
>> You know who else says thanks?
>> That's true. Because they own 90% roughly 90% of the company. It was a full buyout in 2016. They bought the entire company for...
>> um something around 25 30 billion US USD...
>> bought the entire company. They still own 90%...
>> uh today. So their SoftBank's holdings just in that one company...
>> are somewhere in the range of 140 billion.
>> Great. What I was looking at...
>> And of course they massively 100 billion...
>> Well and they've massively levered up uh against that position right they've used...
>> they've um they've raised uh debt against uh their holdings in ARM but certainly Masa's somewhere out in the world seeing it go up 15% just happy...
>> smashing a gong somewhere.
>> I hope so I hope so. Um and although everyone knows uh Apple my my my clickbait for my YouTube video was like Did you know the iPhone is secretly British? I think it's funny because it uses ARM deep inside, but uh ARM licenses to a ton of different tech companies. So, uh Amazon uses uh ARM for the Graviton chips. Uh a lot of Android phone makers have ARM-based chips. Uh tons of other tech companies license ARM's technology to build chips. But now ARM is going to be making the chips themselves. And they're going to be working with Meta Platforms and OpenAI. Uh this means a shift in the economics of the business. So 97% gross margins for just those licensing ISA contracts. Um this will be closer to 50%. But it of but it is should be offset by the huge gains in market size and revenue potential. So the company has one of the highest multiples in semiconductors roughly 90 times forward earnings. Uh so there's a lot to live up to. Uh but there's a lot of value coming from AI agents and that have access to plentiful CPU resources. Uh the industry dynamics are also particularly interesting. Ben Thompson pointed this out in Stratechery. Um, so Nvidia sells an ARM-based CPU that Grace CPU. Uh, and so they're Nvidia is sort of competing with ARM. Jensen showed up and gave like a remote talk at this ARM event where they announced this chip. Um, and so they're competing, but they're also in some ways working together because they're challenging x86 options from Intel and AMD that are still really popular. And so if Intel's constrained, it's like they're like if Nvidia sells a bunch of chips for, you know, AI workloads, well then that actually makes ARM more likely to sell their chips as well because whatever software is built for the Nvidia ARM-based chips will probably run on the ARM ARM-based chips. So there's some sort of like integration there. And the x86 moat is not as strong as the CUDA moat, but there's still this like dynamic of of Nvidia and ARM are going up against Intel and AMD in the same way that uh that different GPU makers are going up against the CUDA moat. So there's like this all these different uh interactions there. Um but it'll all be interesting to uh to to follow. Uh let's let's move on. The big news uh that drove this was of course Meta. Engineering at Meta said, "Today we're announcing a new partnership with ARM to collaborate on the development of multiple generations of purpose-built CPUs to support compute and AI infrastructure." ARM called it the ARM AGI CPU. What a great name. We love it. Let me tell you about Labelbox. RL environments, voice, robotics, data, eval, and expert human data. Labelbox is the data factory behind the world's leading AI teams. And let me also tell you about Sentry. Sentry shows developers what's broken and helps them fix it fast. That's why 150,000 organizations use it to keep their apps warm. So, uh the the Wall Street Journal has another good write up of this story uh talking about how this the timing is good. It is there is a boom in CPUs right now, but uh the stock is already priced very highly and uh everything has to go um perfectly because it's it trades at one of the highest multiples of all the semiconductor companies. So, um, what else is going on in the timeline?
>> Well, let's talk about this data center moratorium bill.
>> Uh, I wrote in the newsletter today, yesterday, Senator Bernie Sanders and NAOC introduced a new bill, the AI data center moratorium act of 2026 that if enacted would require all current and planned data centers to halt construction/production. It even it would even block upgrading existing data centers. So if you have an asset and you want to make changes to it in theory as as this bill is written today, it would be blocked. Uh they um uh they sort of like define data centers based on power demands uh cooling capabilities like how much power you can get to each individual rack. Uh so they have been fairly specific but trying to be like...
>> because I was I was thinking I was fine with this. I was thinking like I don't need any more AI data centers at this point. We can freeze those. I just want to build GI data centers and ASI data centers. And so as long as I can just build tons of those, like it should be fine. But uh it is interesting that they they seem to have figured out the semantic loopholes that might happen if it's defined.
>> Yeah. The bill would halt all new new data center construction and upgrades until more legislation is put in place to guarantee the following.
>> Uh and these will be tough to guarantee. Uh, so from Sanders' site, they want AI to be safe and effective, preventing executives in the AI industry from releasing harmful products into the world that threaten the health and well-being of working families, our privacy and civil rights, and the future of humanity. The economic gains of AI and robotics will benefit workers, not just the wealthy owners of big tech. And AI does not increase electricity or utility prices, harm communities, or destroy the environment.
>> this stuff seems good.
>> Yeah. All all generally good, but...
>> no one wants unsafe and ineffective AI.
>> Well, yeah. And and and the bigger problem is is anytime you're creating I don't think we've created a technology ever that didn't have...
>> some negative impact crash. Um I'm sure this will be rewritten and debated and obviously it has a long way to go before...
>> becoming law but the this set of requirements seems completely impossible...
>> uh to to to actually achieve.
>> Well, it depends on how you def how you define it. I mean, uh, AI does not increase electricity or utility prices like the ratepayer protection pledge and behind the meter stuff could...
>> I'm more talking about...
>> I'm more talking about uh number one...
>> uh number one safe and effective I I mean it's it's it's all in how you define that. Like some of the uh some of the uh parental controls are a good example of like how to take that that overarching thesis and then boil it down into something tractable. And when I hear that, I think like, oh, like I don't know how we are defining safe, I don't know how we're defining effective, like this feels like this could be like some sort of very vague thing where like if the if one particular administration likes this company, they just approve it or not or whatever. Um but then when we actually talk to lawmakers and you hear something like um...
>> like, oh yeah, we're going to, you know, require parental controls. So if your if your offspring has a you know an AI account, you can say, "Hey, like they are this age. Don't show them anything that's inappropriate for that age. That seems good."
>> Yeah. So I don't know. Overall at least this this first bullet point preventing executives in the AI industry from releasing harmful products in in the world.
>> Yeah. That feels like they you could end up having something like an FDA that's like every product you create needs to go through years of of studies in order in order for the government and and it's like hey I just wanted to create like a slightly more AI native version of this SaaS tool...
>> I want to make SaaS...
>> I just want to make SaaS...
>> at the same time like the like the the bull case I mean I I I think we're I think the FDA model would really slow things down based on how long the FDA takes to approve things. At the same time, uh what is the definition of harmful here? Is it net harmful products? Because that is the goal of the FDA. They release drugs all the time that have side effects. You take this, it cures your cancer, but it's going to make you throw up or it's going to make you lose your hair. And people are like, "Yeah, I'll take that. I'll take that trade." And so, you know, if you if you went through the government, you said, "Okay, yeah, I'm going to give you this tool that can like write code, but sometimes it's going to hallucinate and like you might get a code, you might get some code that doesn't pass tests." I'd be like, "Yeah, okay." Like, it speeds me up. On on average, I'm I'm in like that. That's fine. Um, and having some of those disclosures. And it's the same thing with knowledge retrieval. Like, I I I do a deep research report. I get something that's 99% of the way there. Maybe there's something in there that's like, "Oh, that's actually like misattributed or that number is I I know that that number is from this report online that was wrong and they and the model doesn't and so I need to fact check it." Like I still see that as like net beneficial, but there are of course like like flaws in every system and so you know again it's like where how does this get defined over time that's like important. Uh, one thing I noticed from the announcement was that they are using AI leaders' own statements against them...
>> and it's easy to see how uh, this would resonate with their constituents.
>> This is really powerful.
>> Uh, so they on Sanders' website he included this quote. Okay. In December, Elon Musk who leads xAI said he had quote a lot of AI nightmares and would quote certainly slow down AI and robotics if he could. It's so interesting because Elon doesn't talk about that with like the roll out of electric cars or the roll out of of space travel. He's not saying like, oh yeah, like you know, 2030 is too soon to get to Mars. Like we need to we need to slow down on the on the race to the moon. Like let's really figure out the the space suits first. Yeah. You know, he's like, let's just go.
>> Yeah. Uh, and then another one. In January, Demis, the head of Google's DeepMind, said he would support an AI pause if he knew other countries and companies also paused development. In February, Dario, the head of Anthropic, said he was absolutely in favor of trying to slow down AI development if other countries also slowed...
>> That was Davos, I believe, both January, February.
>> I wrote, but the problem of course is that there is zero movement on getting other countries to slow down. I can imagine some companies it would be like...
>> there's already comments in the chat about like yeah let China win.
>> Yeah. Um I I I'm not going to name the countries that would be down to slow down but uh I think we all know that that China even if they agreed to something like this wouldn't just automatically...
>> do anything about it. So uh but the problem of course...
>> are there any countries that are like yeah we we definitely should we're ready to slow down like we're France.
>> Well I think if you're way behind if you're way behind there's kind of a benefit incentive. Elon Elon had said, I think in 2023, that he supported like a six-month pause. Yeah.
>> And at that at that moment, that would have been awesome. Yeah. Because you, hey, if I could just have six months to like get my...
>> I feel like I feel like if you pulled people...
>> in the south of France or or the Amalfi Coast, like those folks would say, "We should just slow down generally like AI, but also just slow down our lives, enjoy a glass of wine, hit the bar."
>> Yeah. Or even like a summer break.
>> Just a summer break like four weeks.
>> Yeah. Or even like during the workday like taking a break, taking a nap, just taking just slowing down generally. I think I think there's a lot of people that are just in favor of that.
>> But okay, Jordy, like if if you're saying um like people who are behind should want to like slow down, like doesn't that mean that China should be more in favor of slowing down?
>> That is interesting. That is interesting. Well, I'm just saying I can imagine in the same way that like for as many chips as we give them, advanced GPUs as we give them, you can assume they're still going to put an immense amount of pressure to kind of stimulate...
>> uh the local semiconductor industry. In the same way here, I'm sure they would love for the United States to just pause all new data center construction. And I think it's possible that they would like generally say like yeah like this seems good but then what would they actually do? They would just use that as an opportunity to yeah...
>> So uh my question is like where does China actually stand on this? I'd be very interested to know maybe you could look it up. Um like has the Chinese Communist Party actually put out any statements about whether they want to accelerate or pause AI development because they're I I think that they might refrain from taking that stance because it would discourage local like indigenous uh development like if the government is coming out and saying like we want to slow down then a lot of entrepreneurs are going to be like okay I'll go back to e-commerce or I'll go back to manufacturing like I I'm not going to work on this because the government doesn't want me. And and so I feel like there's this tug and like even though I agree with you like it would it would be in their advantage to say, "Hey, we want to slow down. Everyone should slow down. We're pro slowdown." Um if they actually said that, it would have an immediate slowdown effect on the local uh AI progress.
>> Uh yeah, like it would slow down. Yeah. I mean...
>> yeah. Because even even if they even if they if they just take that stance because it's an authoritarian country like there's like this like like the the Bernie Sanders comment stands in opposition to other politicians who are saying like no actually things are going well and here's how we're going to you know like advance energy and to build more and so but if you don't have that and it comes down as like a dictate uh or you know like this is the stance from the government it's much harder for local entrepreneurs and local AI labs to like push back against that because it feels like they're all of a sudden in opposition to the government. I would imagine I don't know. I'm very interested to hear how China's actually reacting.
>> Yeah. Anyway, so so just to kind of finish my thought, I all all these quotes...
>> like must go extremely hard...
>> if uh if you're not kind of acknowledging the full picture, which is that the AI leaders are saying, "Yeah, if you get other countries to agree to slow down, we'd be open to it."
>> But that is like the big elephant in the room. They don't mention any even conversations or dialogue with other countries around slowing down and I don't think there's been any. So anyways, the act has a long way to go and it seems like the odds of it getting into law are low but not zero. Safe to say that as written the requirements in the bill would be an incredible gift to America's adversaries and catastrophic for overall AI progress.
>> Yeah. Uh could be popular though. There are already some candidates who have uh you know made pausing AI or data center like like pausing data center buildouts uh sort of like their main thing and uh it does seem to be getting traction I don't know how much of that's just on social media we'll see how it shows up in the midterms but Sagar and Jetty was tracking this for a long time like it is going to be a uh like a political stump speech that hits like people are going to be like yeah I don't like this stuff...
>> yeah the the the question becomes if anything like this were to become law. What what are the...
>> what are the effects of that? Right. Space, right? The space data center people are are saying like space data centers don't seem so silly now, right? Uh taking that angle. Totally.
>> Although I'm sure they would also be like you can't put them up there either. We're going to we're going to try to...
>> Yeah. I mean there's also there's also just like the globalization process that happened uh based around environmentalism in like the 90s WTO ascension for China. Like the reason that a lot of the mining happens in China is because like we said like we don't want that here, right? It's dirty, it's gross, there's chemicals, there's pollution. And so like out of sight, out of mind. Let's push it abroad. And we could do that again with data centers. We could just be like they're all in Canada or Mexico or they're all in some, you know, Australia or New Zealand or like, you know, there could be a receptive country out there that just says like we would love we we are an ally now and we'd love to get all these data centers and then you have to ask the question like what does that look like in 30 years?
>> Data centers generate a a they they don't create a lot of jobs locally. They create a a meaningful amount of work during the development process. They continue to generate massive amounts of tax re local tax revenue.
>> Yeah. I actually don't know how you tax...
>> the data center that blocked I think in New Jersey was going to be generating like tens of millions of dollars of local...
>> taxes. Seems pretty good.
>> Real estate or I mean yeah like one one medium ground here is like just tax that and then have and then use the tax revenue to the environmental concerns. I think I think everybody should want to make sure that if we're investing hundreds of billions of dollars into these things that that we're not destroying our uh lovely...
>> But uh and and the the energy costs again real concern but we're making we're making progress there. So...
>> well let me tell you about CrowdStrike. Your business is AI. Their business is securing it. CrowdStrike secures AI and stops breaches. And let me also tell you about the New York Stock Exchange. Want to change the world? Raise capital at the New York Stock Exchange.
>> Uh, do we want to talk about uh this?
>> Oh, this this minor headline on the front page of the Wall Street Journal.
>> It says Meta YouTube found addictive harmful. It's like one of the hardest hitting headlines I've ever seen on the front of the Wall Street Journal for two companies that are usually relegated to the business and finance section. They made it to the front page because uh they were found guilty or you know uh by California jurors. Uh California jurors say the tech companies designed their apps to cause injury to kids. Very very bad. Uh but Brandon Guerell had a take and a write up uh and some explanation of what's actually going on here. And then we are of course having Eric Goldman from Santa Clara University come on and break it down for us in more detail uh because there is a lot of debate about this uh and it's been very interesting uh watching it play out. The total uh damages are uh I think 3 million each company roughly 6 million total and I had a friend who's a lawyer who texted me and sent me the number and was like hey I'm predicting that it'll be like three 3M and I didn't read the M and I was like okay three billion like what will that do? This is not that big of a deal and then it was 3 million and I was like that's extremely low compared to like the numbers that we normally see from big tech companies but this has much broader implications because this is precedent setting and there will be a flood of other...
>> like I did spend all of my free cash flow on data centers but if you give me another two seconds I will have the I will have the cash flow to cover this so just give me like two seconds...
>> yes but it is it is a very important case even though this particular ruling is not changing the the the cash flow structure of these businesses um because it has a lot of ramifications and there's a lot more plaintiffs that are in the queue. So, uh this is what Brandon Guell wrote in our newsletter today which you can sign up for at tbpn.com. Yesterday, a Los Angeles jury found both Meta and YouTube. It's fantastic. I want a lot of soundboard for this. Both Meta and YouTube liable for a 20-year-old woman's mental health crisis in a bellwether trial that treated platforms as quote defective products and potentially marks the end to the absolute immunity nature of section 230. In the case, the plaintiff's lawyer Mark Lanier argued that Meta and YouTube built quote digital casinos that use neurobiological techniques similar to those employed by slot machines. Fun fact about ARM throwback. The first chip that the company, the precursor company ever built was a chip that went into a slot machine. They call them fruit machines in in England.
>> Fruit machines. They call them fruit machines because they have like the cherries and the strawberries and the bananas and you line them up. Got it. And and they needed chips to run those. So the iPhone traces its lineage and and Apple silicon and all all this tra...
>> This goes much deeper than that.
>> It all goes back to literally gambling. Very fascinating. Anyway, um, so the jury found that specific features of Meta and YouTube are designed to be addictive. Infinite scroll creates an environment where there are no natural stopping points. Algorithmic recommendation feeds users feeds users highly engaging content. Autoplay removes users' agency in choosing whether or not to watch the next video. Notifications pull users back in by exploiting their need for validation. Instagram beauty filters contributed to the plaintiff's body dysmorphia. Features like the like button exploit users' biological need for societal approval. This is what the lawyer argued. The plaintiff's lawyer argued.
>> Shake Shack exploited my biological need for food.
>> Yes, there is this question. There was uh Taylor Lorenson had some great takes here. She said uh So Taylor has come out in the last like 48 hours. I would say is like the number one defender of big technology.
>> she does have one of the top technology podcasts in the world.
>> She she had a she had she had a take that was like uh so does Spotify did Spotify addict you to music by playing like good songs for you on demand? And and...
>> this DJ this this AI DJ is simply too good...
>> basic. I mean that's kind of the argument. They didn't obviously go after Spotify. They went after Meta and YouTube. But uh yeah, there's a question about like, you know, what UI features, what are dark patterns and how do we regulate those and uh it's very interesting, but uh the decision has been made and these companies will have to pay $6 million to the plaintiff in total damages. Uh this is obviously a trivial amount of money for these companies, but the bellwether nature of the outcome has significant implications for social media. There are over 10,000 individual personal personal injury cases, almost 800 school district claims, and 40 state level cases pending nationwide that are similar to KGM versus Meta and YouTube. More broadly, the social media industry's reliance on section 230, which has up to now shielded them from liability for user generated content, uh may no longer be enough to protect them from litigation like this. Now, to be clear, this case is not it's not attacking section 230 because it's not making the argument that someone uploaded a video to YouTube that said, "You should be sad. You should be afraid. It's over." Uh, and that made someone sad. That is the content that is user-generated. That is protected under section 230. You might be able to go after the individual creator of that video. If I make a video that says like Jordy Hayes sucks and should be sad and then you get sad, you might be able to sue me, I think. I don't know. But that that's that's a little bit more defensible because it's like me on you. But you won't be able to sue YouTube for for platforming that content, for putting that up in in just general cases. Um, but this is different because they went after the like button, the infinite scroll, the recommendation feeds, the features that are built by the platforms themselves. So, uh, if the decision makes it through appeals, and this might go all the way to the Supreme Court, we'll see. Platforms may be forced to redesign their user experiences and algorithms, put up age verification, uh, even deprecate infinite scroll. Uh, obviously, changes like this would have an effect on both these platforms' ad-based revenue models. Uh, Meta and Google plan to appeal the decision, and it's not hard to imagine this one making it to the Supreme Court. Uh Taylor's also been uh sort of, you know, sounding the alarm bells around um around age verification. It sounds very good. I'm I feel like I'm pro-age verification based on like I have kids and I don't want them seeing adult content. Uh but she was worried about the the the KYC and and needing to be tracked privacy tracking anywhere and and that is a good that is a good argument to be made. Um well let me tell you about Graphite code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. Uh and let me also tell you about ElevenLabs. Build intelligent real-time conversational agents...
>> inbound.
>> Reimagine human technology interaction with ElevenLabs. Uh we have to start apologizing to the schizophrenic community. There is a surveillance drone reportedly flown by infiltrator elements and disguised as a natural bird such as an eagle that has been spotted in Iran. This goes back to Taylor Lorenson because I believe she worked with the folks behind the the viral stunt Birds Aren't Real. That was sort of a commentary on the conspiratorial nature of the internet. And in that stunt uh they make the argument that birds need to be recharged and they're all spying on you and no birds are real. Of course, uh that is very satirical and funny. Uh but apparently there someone made a drone, some, you know, organization made a drone that looks like a bird so it can sneak behind enemy lines and spy during the the conflict. Uh which is uh remarkable. Um but uh...
>> do you think there will ever be video games that are effectively...
>> you're just remote piloting something in the real world?
>> Do you think there would be demand for...
>> Yes. So I have heard of this years and years ago that there was something along the along those lines uh that would allow you to hunt remotely. So you go to a website and you control a weapon that can hunt an animal which is crazy. I don't know if it's crazy.
>> do they close the loop like help you actually get meat? Yes, they so after you down the animal, they will go and ship it to you...
>> so you can mount it on your wall.
>> Dang. A lot of people...
>> a lot of people are not going to like that.
>> I don't think Joe Rogan would approve. He He's a bow hunter, you know.
>> Yeah. He wants He wants the connectivity with...
>> I think so. I think so.
>> It's destroying just destroying the...
>> Now I'm curious. Is it Is it like a Is it Is it like a Is it a gun mounted to an ATV? Not an ATV, but uh sort of like a a two-axis swivel. Yeah, like a sniper rifle on a on a turret with a webcam.
>> Pretty crazy. I don't know if this is real. I heard about this like years and years ago. Uh anyway, uh the reactions to the Meta and YouTube trial continue. Ariel Grier says, or Gibner says, "This is disgusting and I can't wait for the appeals. The precedent set by YouTube being liable for screen time addiction is kind of scary. Treating algorithms like a defective product opens the door to endless lawsuits over addictive tech. What's next? Books, video games, junk food. I mean, video games we got to do something about. Those things are too too fun. Too fun. Truly, we got to make them. We got...
>> Except Except you you ascended. You beat your addiction.
>> Yeah, I definitely did.
>> You don't You don't game anymore. You actually did. I mean, it might be one of those things like you're still addictive. Addicted. You just...
>> might have been playing a game earlier this week. Got sneaky little 45 minutes.
>> Yeah, it was good. Player Obscura.
>> I'm gonna go I'm gonna go look.
>> Agree. You gotta pull me out. I'm addicted.
>> No. I'm gonna go look back in our text and see like when did it take you 45 minutes to respond to something.
>> I think it was like Tuesday night.
>> Yeah. I'm going to be coming after Eric Jorgenson. Oh, yeah. For the Book of El...
>> uh Eric did a fantastic job.
>> And uh I think it's highly addictive. We gotta we can we sue this lawyer because this is one of the most interesting cases and I I actually can't pull myself away from it and I feel like I'm just going to spend hours and hours uh reading about this lawsuit. Yeah. And and uh I I don't know. It's depressing. It's making me sad thinking about these uh these tech companies having to pay fines and it just it's just ruining my mood entirely. Uh, and it was like, you can tell that the that the the the way the lawsuit was designed was designed to take an emotional toll on me, suck me in, make me read uh the all of the all the transcripts from the court, all the all the reporting, all the research, and uh eventually to, you know, talk to Eric Goldman from Santa Clara University on my show about this. Like, I just can't pull myself away from this. So...
>> Yeah. I I I think I think should he be held accountable for what he's doing? I don't know. We'll see. But there might be losses.
>> Meta of course trading down massively today almost 9% off of this. And you know it there's some real concerns, right? If they there's new legal risk, there's thousands of other kind of lawsuits floating out there. You might get more copycat uh lawsuits, class actions. And then the real question is like do you do they have to make any product level changes? Does that end up impacting...
>> time spent in the app which will impact the advertising business.
>> Um such an such a weird one because certainly certainly I don't think back throughout...
>> NASDAQ is also down 2% today. So there's a general selloff but yeah...
>> yeah general selloff but like you know Meta Meta down dramatically...
>> hundreds of billions of dollars. Um, yeah, this is a weird one on a personal level because like on one hand, my kids, I'm going to keep away the infinite scroll machine as long as I possibly can. Yeah. Right. Um, and I will, you know, I don't I don't think back on the time that I've spent on social media and think, you know, uh, wow, I'm so glad I I put in those those long hours and I really...
>> Yeah, it just doesn't seem that addictive to me. I can pull myself away like anytime.
>> It's not a big deal. It's just not a big deal. Mr. Beast...
>> It just sounds like It sounds...
>> This place picks our honeymoon. Okay. You just got to get more.
>> I ask you a question. If you follow me, I'll double your bank account on Instagram. Okay.
>> I mean, this is good kindness.
>> This is the route you would have to take if you want to go to...
>> This is my subscriber...
>> and this is my subscriber. This is such such a such a good final mystery box.
>> it's fine. We're going to get back to the show in a little bit. It's not a big deal.
>> Not to surprise that college classroom.
>> He grabbed 100 people. Okay. He He took He took people that are ages one all the way to 100 and he's having them taste test lunch. I got to watch this.
>> The world's largest LED.
>> Okay. It's the largest LED floor.
>> No, this is just my normal feed. Who has the best...
>> I am scrolling, Mr. Beast.
>> It's so It's so funny because it it it um I I think I think whatever that was was completely inverted from, you know, a meditation, you know, soundtrack, right? It's just like it's so insane. Anyway, let me tell you about MongoDB. What's the only thing faster than the AI market? Your business on MongoDB. Don't just power AI. Don't just build AI. Own the data platform that powers it. And let me also tell you about Cisco.
>> Critical infrastructure for the AI era. Unlock seamless real-time experiences. New value with Cisco.
>> Well, no. The So, so the funny thing the funny thing is how addictive are the apps themselves? Can they argue that the apps themselves...
>> I've seen some people's channels not very addictive.
>> many in many in many ways like our our content, right? We talk about niche subjects in technology and business. There's a lot of content on YouTube that is...
>> far far far more addictive.
>> yeah. No, I I I've I've generally had a good experience on social media.
>> You know what else is addicting?
What?
Oh, you want to skip to high-speed trains? You don't want to do uh...
We can come back to that.
Okay, we can come back to that one. Uh, what's going on with this? So, in China, high-speed trains have become a sightseeing attraction. This is a good... I would definitely go to this. I would 100% just go see the high-speed train rip through the countryside. That looks amazing. I love it.
Also, this is the most...
Incredible.
Video of China that I've ever seen. What is... What is this mountain range at sunset?
I think people are posting up there even if the train isn't there. It's like one of the most beautiful...
You think that secretly everyone there is like, 'Boo, the train ruined my beautiful view,' because like they could just be sitting there watching the amazing sunset and land.
That guy's leaning forward as...
I think they're pumped up. I think they're cheering. I think they're cheering. Uh, that's great. Um, where else do you want to go?
Uh, we can go back to this article.
Okay. Okay. I... because I need to know where everyone stands. I need to know where do you stand on shoplifting? Are you pro or anti?
You're anti-shop. What about you?
There we go. Uh, I had one moment as a kid when uh I think I took a piece of candy and my parents immediately...
I'm gonna put you under citizen's arrest right now.
Yeah, it's time for citizen's arrest. No, it took me back and it... and I was like the most deeply embarrassing...
Moment that's like a core memory of telling them to learn that.
And I... and I think it was... it was at an age that I didn't really know that it was...
Wrong, but still.
Well, that's not what's going on actually. So, the Wall Street Journal uh wrote a piece about in free expression. Emma Camp wrote this uh op-ed and... and there's a new generation of the... campinator. Uh, there's a new generation of shoplifters who aren't children who have... who should have learned their lesson a long time ago, but are actually uh fairly affluent and are just sort of sneaking out luxury goods. I think that's a piece of it. Uh, but it was... but she makes a very compelling argument for why this is a very bad thing. We need to make skydiving more accessible.
Was... was a dollar for the thrill.
Yeah.
I think a lot of people might say, 'I'm just going to go skydiving today. I don't... I'm not going to steal anything from...'
Well, I mean Ben on our... on our team is working on uh a new weather report of thrilling things to...
If we could make jet skis more accessible. Universal basic jet ski.
I think a lot of people would say, 'I don't have time to steal. Yeah, I'm getting ready for the race.'
Yeah, my cortisol is super low. I don't need to spike it by... by sneaking something out.
So, at least once a week, uh, as I begin my commute from home, uh, from a Midtown Manhattan subway station, uh, I some... I see someone in a suit and tie dodge the fare. Well, it isn't always a suit. Sometimes it's a chunky gray J Crew sweater and ballet flats. Each time I see the fare evader slip through an... an open emergency exit or casually hop over the turnstile, a wave of outrage bubbles within me, says Emma Camp in the Wall Street Journal, free expression. And I consider shouting out after the shameless thief. You should do it. I... I am in favor of... of yelling at fare dodgers in this case. Uh, but before I can work up the courage, he's melted into the crowd. It's tempting to give fare hoppers and shoplifters the benefit of the doubt and assume they are all Jean Valjeans. Do you get that reference?
Les Mis. You got to see Les Mis. Jean Valjean is the pinnacle of morality. He steals a loaf of bread for his family because they're starving. Goes to jail, gets revenge. It's a great story. You will love it. You should see it on the... in a theater because it is, of course, a live stage production many times. But there's also a bunch... there's also a great movie with uh uh Liam Neeson. The Liam Neeson one is the one you want to see.
The Liam Neeson Les Mis is amazing.
I think I have that right. Uh, but... but that's far from the truth. According to one 2025 survey, and this was what was interesting about this article, uh, the likelihood that someone would purposefully take an item at self-checkout without scanning actually increased with income. So the rich get richer. The rich are stealing from the rich. Uh, a viral article in Curbed, a site published by the New York Times mag... or by New York Magazine, uh, paints a vivid picture of this kind of thief, a member of a certain subset of the city's wealthy-ish, as writer Nora Delight... uh, puts it, uh, for whom a little shoplifting on your grocery run has become as... about as mundane as jaywalking. This isn't exactly surprising. A report from the Council on Criminal Justice released earlier this year found that while reported shop... shoplifting decreased during the pandemic, rates have returned to roughly pre...
Wait, did they just decrease because people were ordering groceries online?
I think so. I think so. If you're not in the store, you can't steal anything. Uh, you got to... you got to hack the system or something. Um, it becomes a cyber crime to steal from DoorDash. Um, but rates have since returned roughly to pre-pandemic levels. Uh, this... uh, this New York Magazine profile in uh Curbed profiles several... several shoplifters who were caught stealing from Whole Foods. None were acting out of necessity and most stole what can act... what can accurately be described as luxury goods: $30 eye cream, strip steak, fancy organic chocolate. None expressed guilt and no one... and one explicitly justified her actions as being in a kind of... a kind of artist's subsidy. Petty theft is a vice for a certain kind of loser. Go off, Camp. I love this. Uh, many of those profiled in this... uh, in this article are... uh, are creative types: photographers, graphic designers, sculptors, and musicians. Presumably, these people are highly educated but down... downwardly mobile. They steal because they feel entitled to the kind of life where they can thoughtlessly drop $50 on French cheese and sushi rolls while paying Manhattan rent. They can't... uh, that they can't indulge in such luxuries feels to them like a moral outrage, one that can be rectified in some small ways by taking what's owed to them. When shoplifters justify their actions online, they make themselves out to be Robin Hoods. They claim it's good to steal from large and therefore evil corporations. This is a teenager's leftism, one in which fighting the man means ripping him off. It's a post-hoc justification for a baser impulse: the belief that everything you want should be free. This motivation is made particularly clear when people justify fare evasion, given that public transportation is inherently not a big private enterprise. It's not a mega corp. It's... it's uh, you know, taxpayer dollars that you're stealing from. Even when we grant... uh, even when we grant the assumption that large companies are inherently evil, which for the record I don't, says Emma Camp in the Wall Street Journal, that doesn't justify stealing from them. If 'thou shalt not steal' doesn't work for you, consider another argument. Petty theft damages the institutions and businesses in a community, making them materially worse for your neighbors. You may think you aren't hurting anyone. Actually, you're hurting everyone. Businesses hike prices to make up for shoplifting losses. Have you heard of like loss? It's called like loss rates or something, like every retail store just assumes that they're going to get 20% or 5% or some...
Yeah, I don't think it's...
It could be high. It could be high in some stores, like significant amount of merchandise just like gets... loss prevention is like the term that they focus on, but it's like a constant problem for these businesses. Uh, and they have thin margins. People don't realize... I think people think like grocery stores because like you go and check out, it's 100 bucks or something. It's like they're making a ton of money, but like the margins are very, very thin on those. It's not... it's not... it's not ARM. They're not licensing IP to Apple. They may even close. Public transportation crumbles without sufficient revenue from riders. As journalist Kelsey Piper put it in a post on X, 'When you shoplift, you directly and unambiguously impoverish your community. That bar of chocolate you swiped isn't nothing. That little pot of eye cream you slipped into your purse is more than a rounding error on a corporate balance sheet. What you're doing ultimately creates a worse, less functional, less trustworthy society, and even shoplifters need a place to get groceries.'
Uh, yeah, the comment section is agreeing with Emma.
They said, 'Well said, and I'll extend it but beyond just shoplifting to include things like buying something with the intent of returning it after using it for a few days. I remember someone buying a grill at Lowe's with the intent of using it for the weekend and then returning it, which they did.'
Yeah. Just didn't... stakes didn't really come out that well.
I don't think it was me. I think it was the grill.
I'm not... I'm not a big returns guy. I... Costco does have a huge uh return policy that some...
Yeah. It's so... It's so interesting. The um, the Coogan or the Hays mind cannot comprehend the return.
Like it's... I... I've maybe returned...
In the last 10 years. It's just not an impulse that I have. It's...
Because you're deeply in touch with everyone.
Um...
So funny. Anyway, okay. Let... let me tell you about Turbo Puffer, serverless vector and full-text search, built from first principles on object storage. Fast, 10x cheaper, and extremely scalable. And let me also tell you about Gemini 3.1 Pro with a more capable baseline. It's great for super complex tasks like visualizing difficult concepts, synthesizing data into a single view, and bringing creative projects to life. Uh, and thank you. The clapping in the studio is at an all-time high now that we're back to full strength in the TBPN Ultradome. So...
That's right. Uh, Wilmanitus is giving some more context uh uh around anti-AI backlash. He says, 'In case you're skeptical that anti-AI backlash will become a central issue in the next election cycle, Sanders just introduced a bill to halt all new data center construction and require the federal government to review and approve AI products before release.' Interestingly, this idea of a... of an AI FDA, the first time I heard about this concept was on the All-In podcast, I believe. Uh, Chamath Palihapitiya was making the point that an AI FDA was something worth considering. I think he's evolved his position, but uh it... it was something that uh, you know, the first time you learn about like the possibility of a fast takeoff, the possibility of AI doom, it seems really rational to say, well, let's review everything beforehand. Uh, but then there's natural effects and then there's limitations to the models and we can't even do a puzzle on ARC-AGI. So maybe there's a little bit more work to be done. Maybe there's a few more data centers to build before we actually get good things out of these models. But um...
Wrote a piece that uh I think is relevant to the data center moratorium, which is 'AI has the worst sales pitch I've ever seen.'
He... his quote, 'Our product will make you economically useless and possibly kill you,' is not... says is not a value proposition. And AI leaders need to change their public messaging. And fast. We don't need to read through all this, but...
I do need to read this quote. He opens this quote: 'Hi, do you have a moment? I'm from the cursed microwave company. Our product is much better than a traditional microwave. Not only can it automatically and perfectly cook all your food, it also microwaves your whole body so you and your family are paralyzed and unable to work ever again. Don't worry, though, because when everyone has a cursed microwave, our society will probably implement universal basic income, and you and your children can just go on welfare. Oh, by the way, we estimate that there's a 2 to 25% chance that our microwaves will put so much radiation... will put out so much radiation that they destroy the entire human race.' If a door-to-door salesman came and gave me this pitch, I would gently see him out the door and then quickly call the FBI. But this is only a modestly exaggerated version that... the pitch... of the pitch that the big AI labs, OpenAI, Anthropic, are making to the world about their technology. And I completely agree with this. Uh, it's very... it's very important that... that AI lab leaders uh, you know, focus on the actual tangible benefits of the... of the technology right now. Like I think there's this... you... you keep... uh, you keep identifying this disconnect between like... like people are afraid of AI and there's so much fear-based marketing, but then like... like every single person, like a billion people, are just using ChatGPT to like get answers. And... and every person you talk to, no matter how they... they can tell... they can tell you for 20 minutes, lay out a very coher... like cogent thesis of AI doom, what might happen, Terminators. They can take you on this very reasonable scenario and then five minutes later the topic changes and they'll tell you how they use ChatGPT to like, you know, plan their vacation or something like that. And... and this disconnect just like cannot last.
Yeah, I wrote on January 9th, uh, the phrase 'techlash' was originally coined by Adrian Wooldridge in The Economist in 2013. He correctly predicted that the big developments of 2014 will be the growing peasants' revolt against the sovereigns of cyberspace. The Silicon elite will cease to be regarded as geeks who happen to be filthy rich and become filthy rich people who happen to be geeks. Over the coming years, the world experienced the Cambridge Analytica scandal, fear around the mental health impacts of social media on young people, and growing concerns around monopoly power in our digital world. The internet and the platforms that dominated began to test the core foundations of our country and... and the free world: privacy, democracy, censorship, and more. The second techlash has begun. This time because the average American believes that technology, specifically AI, is now a threat to their very way of life. Uh, starting at the bottom of Maslow's hierarchy of needs, Americans have heard the data centers use a lot of water. They heard their power bill is probably going to go up because of it, or already have. They've seen the movie Terminator and have imagined how that might go from science fiction to reality.
Watch Terminator. You still haven't seen it, right?
Uh, no, I have seen Terminator.
You have? Have you seen Terminator 2?
Uh, probably not. You're killing me, John. It's enough to ask me to finish one movie. Now you're like, 'Oh, you got to see two.'
Two is... two is the best one, I think. Uh, but the end of Terminator, it's a good ending. The humans win. We beat the robots.
It's a good outcome. You got to push back on that.
Uh, no, I was... I was going to add. So earlier we were talking about like China and like how is China thinking about some stuff? I found some...
Okay, so um, Xi Jinping is like... if you basically compare like Chinese government to US government, Chinese government is probably like much more concerned about safety than the US is.
Um, like there's always things like... so Xi Jinping, we don't want to... we don't want other people to be able to do surveillance as well as we have.
Yeah. There's always like the... the take which is like...
You know, if you're super concerned about safety, well, what does that mean? You need these... the... this oversight over everything, right? Which is like if you're China, you want to be... that's like the thing you're going to do.
Um, but basically he was like... uh, he shared this. It was from Time: 'A rare Politburo study session on AI, warning of unprecedented risks.'
Um, it's listed... AI safety is listed alongside pandemics and cyber attacks in their national emergency re... response plan.
Um, he's also said... uh, he stressed the need for monitoring, early risk warning, and emergency response.
Right. So it's like sure, that's like about safety, but also, yeah, you know, monitoring, right?
No, no, no. It's kind of... I wonder... I wonder how this interfaces with the Bernie Sanders position. Like will there be some sort of uh like open discussion? Uh, you know, Bernie's going around doing a lot of different conversations. That would certainly be like the next step if he wants to get through the uh, you know, the geopolitical pushback.
Even when you talk to a lot of like the Pause AI people, very few of them are... are like ever really say the US needs to pause regardless of what China does. It's always like we need to have international...
Uh, you know, agreement on... because otherwise like it obviously doesn't work.
Yeah. If everyone could just give their GPUs to Tyler...
He will watch after them.
So this story will continue to evolve.
That's good info. Thank you.
Uh, Rune hit 300K. He says time to dock on TBNN.
Okay, let's do it. 300. Let's go. Congratulations.
I think we hit the gong at 252.
Every... every time, every time there's a Rune milestone, we must celebrate it to the fullest. Really quickly, let me tell you about Cognition. They're the makers of Devin, the AI software engineer. Crush your backlog with your personal AI engineering team. And let me also tell you about Console. Console builds AI agents that automate 70% of IT, HR, and finance support, giving employees instant resolution to access requests and password resets. Uh, continue.
Let's pull up one of the greatest videos created in the last year.
Spor says, 'RIP Sora, you gave us the greatest AI video of all time.' Yes. Let's play it.
This was a great video. I did enjoy this one.
Yeah, you got that. It's midnight. Give me that.
No, sir. No more fiddle tonight. Got it. GO HOME.
HEY, WHAT DID I SAY ABOUT THE NOISE? IT'S THE MIDDLE OF THE NIGHT. NO MORE STINT ON THE PORCH.
The door. I'm taking it inside. Sora model. I don't know what the... what was in the training data, but it nailed this like doorbell footage.
Like the graininess really leans into the flaws and it... it hides the flaws in the model. When you try and do something that's photoreal, it looks a little waxy still, but this is just perfect. And it's too... it's too low res to detect like, oh, the finger's perfect or whatever.
That's a good instrument. That's fun.
She really... She really hits the cat with that door.
Yeah. Well, the digi do...
Here's the thing. Here's the thing.
Creativity. You had to string all these together. This was not one shot. This is like...
This is like, you know, 10 prompts.
I've done it. I... I... I've gone to chat and said like, think of a funny prompt and it doesn't come up with something as novel as this. Otherwise, I would be the creator of that. Hey, that's our gong.
That's literally our gong.
We made it into the training data.
Volume down. Pause off the neighbor's cool down before 3 a.m. Okay, that's good. I love it. Uh, yeah, the Sora... the Sora news is interesting. I mean, obviously good to... as a reflection on like, okay, era of focus. We heard that news from like a week ago. Uh, this is like a very clear step and like refocusing. Uh, there is a question about like the value of models like Sora. Jumped ahead of Meta Vibes very quickly. Meta Vibes launched like a week before and when Sora came out, the consensus was like, 'Oh, wow. This is like definitely a step above what you can do with Vibes,' which was much more like an animated picture from Midjourney. It had cool... cool aesthetics. Um, but it was... it was way less like... it was... it was very constrained in terms of what you could do. With Sora, you could... you... you could automatically insert different cuts, um, generate something that looked a lot more like a native reel, even if there were sort of rough edges. But uh, the question with every model is, um, you know, you spend money to train it and then it has ongoing inference costs. The cost of a video model, it feels immense because I went back and looked at some of the first Sora videos that I generated. Uh, I'm thinking of me in the orange and white suit.
Yes. One of the greatest AI videos. We'll have to play it at some point.
But um, that one felt like it had about a thousand times as much compute allocated to it than the last one that I generated because I generated Sora like a week before this announcement. And it did... it does feel like they had... they had shifted the GPU workloads to other... to maybe Codex because that was the product that was taking off. What do you think? There were some rumors where they were spending something like 10... $15 million a day on compute for just... for Sora, which is like, you know, like three and a half billion a year or something... something in this range.
So it's like so much compute just on inference and it's like, yeah, what's the actual value of...
Yeah. Yeah. And... and if it's not monetizing with crazy ad rates immediately or crazy API costs, like... like you just... you're just in GPT-2 territory where Greg Brockman came on and told us that, you know, I don't know, they spent probably millions of dollars training GPT-2 and they were like, we had to pay people to use it. Like every dollar of inference was... was negative margin. We were losing money on every... on every token generated. And then GPT-3 was in a similar boat. It was like almost there, but they weren't really making any money for it. And then the GPT-4 training run happens. And that, I mean, looking back, that has to be an incredibly profitable uh model because I think it maybe cost $100 million to train and then the inference cost... like they distilled the model pretty quickly, uh, got to a place where I think that was very, very ROI positive based on the subscriptions during that time before they rolled out 5. And then, you know, you're always facing this battle of like, you know, a little bit of a Red Queen's race, training the new model, you know, depreciating the model across whatever the... the inference margin you get during... during the period where you have the best model. Um...
Mitchell Green went on Invest Like the Best and Patrick... Patrick pulled out one of... he said Lead Edge is probably most famous for this letter. Mitchell calls it the hierarchy of BS. It's his way of distilling what he learned from cold calling 10,000 companies and I will go through it. If companies have good cash profits, they report those. If companies don't have good cash profits, they report on adjusted profits. If companies don't have good adjusted profits, they report on gross profits. If companies don't have good gross profits, they report on revenue. If companies don't have good revenue, they report it on adjusted revenue indicators like GMV. If companies don't have good GMV, they report on monthly active users. If companies don't have good monthly active users, they report on subscribers. If companies don't have good subscribers, they report on downloads. If companies don't have good downloads, they report on page views. If companies don't have good page views, they report on what they were voted the best place to work in XYZ City.
It's very funny because the only news that we've shared from TBPN is that we're like the 42nd most innovative company recently, but we're not a public company.
If you want to trade a public company, go over to public.com. Investing for those who take it seriously. Stocks, options, bonds, crypto, treasuries, and more with great customer service. I want more of your take on this. People that... that... that flex how many lines of code they're writing.
Oh yeah. Like if your product doesn't have any... you... if your product doesn't have any users or revenue, you just report on lines of code.
Lines of code... lines of code per year run rate is a crazy one I saw recently.
That's an insane stat. But we have the solution. It will be released soon. Uh, do you have a general update? Yeah. What's the timeline looking like?
Did you... when you went on... when you went on the road, did you stop walking in...
Work. The uh project is... will... can be done today, tomorrow. Really? Does it have sound effects yet?
It does not yet have... in a day or two.
Just one more fundamental mechanics. I was fine-tuning tomorrow 1207 chat remind.
I talked to Tyler and he said he's... he's generating 25,000 lines of code an hour in advancing of this project. So I mean, he's doing... he's clearly putting out a lot of code and that will translate to a great product eventually.
The best entrepreneurs I know are generating millions of lines of code daily just to keep their about page up and running.
And how many... about page rough. Well, let me tell you about Restream. One live stream, 30 plus destinations. If you want to multistream, go to restream.com. Just do it. Uh, Gwynne Shotwell is on the cover of Time magazine. SpaceX is racing to build the most powerful rockets yet with the goal of returning humans to the moon. Gwynne Shotwell is leading the charge alongside Elon Musk. This is a good deep dive. You should go check it out. Um, my... the first YouTube video I ever made that went viral or like broke out because when you have...
Was 'Why is no one talking about Gwynne Shotwell?'
Literally. Literally. That was not exactly the title.
But it was... it was that... that was the whole premise. That was the whole premise. And... and uh normally I think I had a thousand subscribers and so getting like a thousand views was like, 'Okay, this is great.' And if you can get more views than subscribers, that's like a banger.
The title was 'The Only Reason SpaceX Works.'
That might have been retitled a few times, but um, you get the idea.
Anyways, no surprise that Gwynne Shotwell...
Running a company that is shooting things...
Rockets into the atmosphere.
Well, uh, FT has some reporting on the SpaceX IPO.
Um, the bankers involved are toying with the idea of allowing existing shareholders to sell out of their positions on day one, the people said.
That would do away with rules to prevent insiders cashing out or trading shares, which are typically imposed for 180 days after lockup. I would think rolling lockups.
A third person said SpaceX is considering a so-called staggered lockup that would allow initial investors to slowly sell down their stakes over the course of several months.
'The deal is,' quote, 'too big of an IPO to just unlock.'
Yeah. 'Musk's influence over the size and timing of the listing has been highly unusual,' said one investor close to a banker on the deal. 'It's kind of scary how Elon is setting the price. It's not being done formally.'
Oh, interesting. It's not... it's not a full auction format. He says, 'This is the price and you can either buy or not and hopefully you buy.' Um, yeah, interesting. The... the staggered lockup, it feels like it makes a lot of sense. You could just see, okay, this person's held these shares for 20 years. This person held it for 10 years. They get a 180-day lockup, a 90-day lockup, a, you know, 270-day lockup, whatever you want. I don't know.
Just imagine how much money the lawyers are going to make from unwinding all the different entities and the SPVs and the SPVs and the forward contracts with the SPVs and the forward. It's going to be...
Full employment for lawyers for sure. Um, xAI co-founder Kroce...
Is departing the company, which is, I think, officially the last co-founder...
Remaining outside of Elon. So...
Well, it'll be interesting to see what the next... uh, what the next run, the next version of Grok does. Is there... is there a new 4.2 came out? Is that right? Yeah. I mean, that was a while ago.
That was a while ago. So, we're expecting like a 5 or 4.3 soon.
Yeah. I don't know if anyone expected super soon.
Right. Because they're kind of revamping the whole thing.
Yeah. I wonder what they'll try to do with ARC-AGI V3 because they were the fourth lab that tested and they were the only lab that got 0%. Although, I don't know how much worse that is than getting 0.25%. Uh, but certainly, uh, it'll be interesting. Jordan Peterson had an idea for Elon Musk. He said, 'If Tesla makes a car with three rows of seats, each with its own pair of doors, so nobody has to climb over everybody else to get in their seat, they will create a baby boom the likes of which we haven't seen in... in 80 years.' And Elon Musk says, 'Noted. Would you buy a car with three sets of doors? It would look extremely odd. It wouldn't be the very iconic, but that's sort of Tesla's thing.' And the Cybertruck stands out on the road.
Lord. Okay. Well, uh, yeah. So, so the Y L...
But it has two sets of doors. Four doors. Six-door car.
Someone has to have done this at some point. A six-door.
Having a kid jump in and just crawl to the back.
Was... when back in my day it was up when you climbed in that door. Yeah. It was uphill to the third row. Yep.
Both. And then... and then uphill on the way out.
Yeah. But it made you who you are... into the man I am today.
Yeah. And so you don't want... you don't want easy street extra doors back. No way.
Your kids are going to be climbing. They're going to be climbing all over the seats.
Um...
Yeah. It's interesting. It... it feels like Tesla's entire strategy from a merchandising standpoint has just been to make like more and as much as possible just make...
The... the... the lowest number of cars.
And so he's getting all this feedback but...
They already have like a six-seater coming out.
Okay. Someone... someone shared an image.
Yeah. This is the AMSTAF...
The AMSTAF... the world's oldest and largest manufacturer of custom multi-passenger vehicles and it's a Suburban that was converted and it comes with six or eight-door models so you can have rows and rows. I don't know if you can scroll down and find it. It's from Talos Dreams in the replies there. Um, but I don't know. Uh, anyway, uh, where do you want to go?
Meek Mill says, 'I'm getting 250 for a tweet.'
Oh, interesting. Uh, I... I was fascinated because he's gone on like this big tear of like promoting AI.
Specifically, uh, Claude.
What is he... what is Nikita Bier going to say about this if he's getting...
Paid? Like is he basically just saying like, I'm... I'm doing undisclosed paid promotions?
Uh, because he's certainly getting more than...
Likes. Yeah. I don't know. He got about 250 retweets on this. 295. So I don't know. It's possible. Uh, let me tell you about Okta. Okta helps you assign every AI agent a trusted identity. So you get the power of AI without the risk. Secure every agent. Secure any agent with Okta. And let me also tell you about Lambda. Lambda is the superintelligence cloud building AI supercomputers for training and inference that scale from one GPU to hundreds of thousands. And I believe we...
Last... last thing before we go into our first guest. We have to pull up this post from Peaceand. He says, 'I've seen some fire fits in China, but I've never seen a jacket this hard in my life.' And the guy, the back of this guy's jacket just says, 'Mr. Enjoy the Money.'
Mm.
Mem. Mr. Enjoy the Money. This is...
Apparently, that's a... that's a legit brand. That is very tough. I like it.
Mr. This might be the new Meta. This might be the puffer jacket with the... the stitching on the back is very nice. Uh, there's a lot of fun examples of other... other swag in the replies. Well, without further ado, let's bring in Eric Goldman, who's at Santa Clara University. He's the associate dean for research and a professor there. Eric, how are you doing?
What's going on?
I'm all right, thanks. How are you?
I'm good. Thanks so much for joining. Uh, please, since it is the first time on the show, I'd love a little bit of an introduction and then of course I want to go into the case today.
Uh, sure. Uh, I'm Eric Goldman. I'm a professor of law and associate dean for research at Santa Clara University School of Law. I've been doing internet law for over 30 years, teaching, researching in the area. And my area of specialization includes how users talk to each other online.
Okay. And when did you start tracking the Meta YouTube uh lawsuit? How long has this been going on? And then I want to hear about your expectations for the case, your interpretations of the case, how things played out.
Uh, so I started tracking the cases when they were filed. Uh, I can set up electronic alerts to notify me of new filings in the cases.
Um, and I had to give that up after a while. These parties were just beating each other up in court um with an enormous volume of very punny filings. So um, I stopped tracking at that level um a while ago. Uh, it's hard to track the state court cases. There isn't an electronic tracker for that. So it's been a little bit harder to get the information in that case. Um, and uh, this has been going on for a couple years and many millions of dollars have been spent by both sides in these cases.
So uh, take us through what actually happened, how we should be interpreting the... the... the verdict and uh... and then we'll go into where we go from here.
Uh, so we need to step back for a moment and let me lay the... uh, describe the landscape and then we'll see how this case fits into it.
Um, in California state court in Los Angeles, there have been hundreds of uh plaintiffs who filed lawsuits against the social media services. These cases have been joined together in a proceeding but they're not a class action. So there's hundreds of parallel lawsuits all sitting next to each other. A similar lawsuit is also pending in federal court in Northern California. It's called a multidistrict litigation or an MDL and there are thousands of plaintiffs in that case as well. Similar format. They're sitting next to each other. It's not a class action. That case also includes public entities like school districts um and uh uh Native American tribes. So, there's other entities in that as well. And then there have been individual filings across the country by state attorneys general and others uh in the lawsuits. Um, so there's basically just a litigation mania taking place across the United States.
Yeah.
And what does it... what does it actually take to qualify a plaintiff? Do they just look at their screen time and say like, 'Ah, you used social media for five hours last week. You're addicted.' And then they kind of figure out how to... how to kind of, you know, turn suit.
Uh, there is no screening for who gets named as a plaintiff. They just raise their hand and say, 'I think I qualify.' Um, and they've been joining these lawsuits. Um, the lawyers of course are helping them and each trying to... each lawyer is trying to add to their portfolio of plaintiffs that they uh have that they're representing. Um, so... but whether they're actually going to get any recognition from the court or money, they'll have to make their case eventually. So that's part of why the... the trial was held that... that led to the verdict that came out this week. What they did in the California state court case is they said, 'We're going to pick three plaintiffs and we're going to do what we call bellwether trials. We're going to let the parties make their case to the uh jury and we're going to see what the jury thinks.' The goal was to do three, not just one, with the idea that plaintiffs have so many different stories, the victims are each in their own unique situation that we wanted to get multiple data representations from the pool of plaintiffs to start to be able to estimate the overall value of the case. So, the first one came out and the jury agreed with the plaintiffs, did not agree with the defendants, and awarded $6 million of damages. Now, this is just the first data point of what's expected to be three just in the state court case. There's lots of activity taking place throughout the country.
So, how... how should we think about the actual ruling? Uh, the Wall Street Journal has a headline: 'Meta and YouTube Found Addictive and Harmful.' Uh, what... how... how are you interpreting the nature of what the plaintiff attorney proved?
At the core, the plaintiff's lawyers had to make a number of novel arguments both to the judge to get to the trial and then in front of the jury, they had to convince the jury to recognize the harms that the victim suffered and attribute them back to social media uh services. And the jury agreed with all of that. And so the way to... I think to interpret this uh uh verdict is that we've now heard from some ordinary Americans. I don't want to say, you know, metaphorically picked off the street. Um, and asked them how do they feel about the accountability of social media services for the harm that their... of their users and the uh jury said we think that the social media services should be responsible and we should assign some pretty significant dollar values to that responsibility.
Is... is that uh... is... is $6 million seen as a uh... a significant dollar value in the courts because these companies are so big? Uh, but again, this could just be the beginning of it by potentially thousands.
Yeah. So... so where does this go in terms of uh the... the... the damages as they snowball? Will there be a class action, some sort of master settlement uh that... that goes way bigger?
Just about the numbers for a moment. Um, $6 million is uh far less than what the plaintiffs requested. Um, it's far more than I think the defendants had hoped. Um, and as you were pointing out, if you multiply it times 3,000 potential plaintiffs that are already in the queue, um, that translates into something on the order of like $20 billion. Again, maybe numbers that Meta and Facebook and Google have, but TikTok and Snap, if they're also going to be coming along for the ride on this, maybe they don't have that. Maybe that becomes uneconomic.
So the numbers grow I think really rapidly because not only is it the people who've already filed, it's all the other potential users who will say just like um uh like you were suggesting earlier who will say, you know, I've been addicted for five hours on my site. There could be tens of thousands of more plaintiffs coming through. The numbers on this could get extraordinary. Now remember, this is only one data point. The other two bellwether cases could establish no liability or they could assign much larger damages. Yeah. And so we'll see if that $6 million is even the right number for the parties to start estimating the settlement.
Sure. So uh, will those other bellwether cases uh for sure resolve before this gets appealed? Do you anticipate that this ruling will get appealed all the way to the Supreme Court? How do you think about the interaction between the bellwether cases and then other legal proceedings?
So my... I believe that the appeals are going to be lodged now, but I would expect that by the time that the appellate court is ready to hear the appeals, the other two cases will have been resolved and ideally they'll be all handled together. Whoever doesn't get the results they want in the other two bellwether trials would go ahead and appeal those as well. All these cases are going to be appealed until there's some kind of definitive resolution. Um, and I would expect the appeals to go up uh all the way as high as the court will take them. The US Supreme Court has discretionary review, so they may not take it, but I expect the cases to be appealed to them unless there's a settlement. And you did ask about that, and let me just close the thought. The difficulty with a settlement is because they are not class actions, the defendants cannot settle against all the potential plaintiffs. They can only settle against the ones who've already expressed uh... who've already signed up to the case. Um, and so one of the tricky aspects of this case is how could they price the settlement to pick up all these other potential plaintiffs who are waiting in the wings.
Are there any other countries that have gone through this process of uh litigating social media addiction? It feels uh like uncharted territory. I... when I see things like the effect of the like button or endless scroll or feeds, uh, there's been a conversation about the effect of those technologies for a long time, but uh, this is the first time I've seen them actually been, you know, in the court of law.
That's a great question and I don't actually know if there's been any trials on these questions in other countries and other countries don't always have a jury system like we do. So, I do think we are in some new ground here. Yeah. Um, having said that, other countries often can regulate the specific mechanics of online publishers in ways that the First Amendment may not protect... permit here. So the uh other countries might very well have already dictated that certain things cannot be done that are part of the litigation now.
Yeah. So, do you think that there's a hope for sort of new federal regulation alongside what is going on with Section 230 that sort of sets the rules of the road at a national level that then if you're in compliance to... you can't be held guilty of, you know, this... these all these lawsuits?
Uh, it's possible that we could get some kind of federal legislation that provides some kind of security or safe harbor for the defendants. Yeah. Um, that would also preempt any of the state laws that would conflict with that. I don't see that anywhere in the Overton window, you know, the window of what's actually being debated today. Yeah. Um, but maybe if the services are feeling enough pain, they would be willing to um... uh, to pursue that. The reality is today that states throughout the country are passing laws that um govern the same issues that are at issue in the litigation. Um, and so unless there's some federal preemption laws, the... the defendants, the social media services are going to have to navigate this ever-growing stack of state legislation.
Do you think we get to a point where when you open a social media app, it has like an addiction warning like you might see on a cigarette... on a cigarette box? Is that... is that where this is headed? Or if this goes far enough, are we... are the platforms running the risk of, you know, needing to remove the like button because it... it uh hurts people's mental well-being or... or uh eliminating the... the infinite scroll uh or... or app-level usage limits? Like what are... what are kind of like the extreme sort of downside scenarios for the platform?
Um, uh, so there have been some states that have enacted uh mandatory warning labels like you're describing. Uh, those are going to be subject to constitutional challenge. It's not clear to me that they're actually constitutionally permitted. Uh, for example, Texas's law has uh I believe been enjoined on that very point. Um, so the warning label approach I don't think is likely to be the final resolution in any... of any of the disputes here. Um, I think that... that's going to be an effort that probably isn't going to either solve a problem or be constitutional.
Um, it is already the case that states have passed laws that have done things like tried to ban auto scrolling or um...
Infinite scrolling, or autoplay, or the like button, or like counts, or whatever. States are going to get to that level of granularity, and they're already doing so. Many of the things that are already part of social media features today are regulated by state laws that are on the books, but possibly being in the process of being challenged. And I think the bigger question isn't, will we lose our like button, or will we lose infinite scrolling? I think we would consider that, for those of us who are aficionados of social media who think that there's some value to them, we would consider that to be maybe the best possible outcome. I think the more serious outcome is what social media services drop out of the industry entirely, what social media services add barriers to user populations having access to social media, and how much we see the circumscription of social media conversations entirely. In other words, I don't think we should assume that social media will exist in its current form in the future. The question is whether we will even have social media in the future. I think that's the state of the cases and the legislation.
Yeah. Yeah. Yeah. That's fascinating. What other industries have been through this type of pivotal moment, and where should we be drawing analogies, or where should we avoid drawing analogies? People have said social media is the new cigarettes. Jordi was making a point about the warning labels on cigarettes. It feels like that might not be the actual best analogy. I was talking to Tyler on our team about the risks associated with driving a motorcycle, and I was making the point that there's an instruction manual that you buy, that you get with a motorcycle, that puts a lot of the risk on you if you choose to transport yourself.
This really progresses like unprecedented sets. It's like, okay, fast food companies shouldn't be able to use bright colors in their logo.
There's a lot. Yeah. There's a lot of different...
Musicians shouldn't be able to play hooks that are just too catchy.
Catchy, maybe. I don't know. Yeah. But where do you think the best analogies are?
I'm not a fan of analogies between online and offline activities. And in particular, I don't like the analogy to tobacco. And let me make clear why. And I think that'll start to enlighten where the risks and opportunities are. Tobacco, we would generally say there are no health benefits to consuming tobacco. Nobody is healthier because of that. It's just either neutral or negative. But social media consumption has many social benefits. There are many communities that are thriving online and who are, you know, I think justifiably concerned that their benefits are at risk in this litigation without them even having a say in it. So unlike tobacco, social media is a mix of potentially very helpful things and things that could potentially be harmful for some of its consumers. And those kinds of mixed-use technologies that have both benefits and potential detriments are really, I think, a better analogy than tobacco. And I don't know that we have great circumstances where we can point to where mass tort litigation has targeted and potentially removed from the industry these mixed-use scenarios.
And in the end, the things that people are being addicted to, or that are allegedly causing harm, are people talking to each other. Yeah. And that kind of speech-related theme is unlike almost any physical product that we might draw an analogy to. Mhm.
Yeah, it's a fascinating case. We really appreciate you taking the time to come chat with us about it. I hope you have a fantastic rest of your day.
Yeah, come back on as there's more news.
Yeah, we'd love to talk to you again in the future.
