You're watching TBPN. Today is Thursday, October 15th, 2026. We are live from the TBPN Ultradome, the temple of technology, the fortress of finance, the capital of capital. How many days till Christmas? I know that's what you came to the show for. That's the question you've been asking. We got the answer for you right here on TBPN. There are exactly 70 days until Christmas.
It feels so close. It feels so close. It's remarkable. It's been wonderful that you've all joined us on the road to Christmas. It's been a strong, strong road. I do want to give a little...
There's more coming. There's actually a lot more coming. Another 70 days or after that?
Well, of course, the clock restarts. Yeah, 364. Immediately. Of course. But I do need to give a little bit of... first, 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. All in one place. I actually have a little Christmas gift right here. Let me open this up. Production team tossed this my way this morning. And let's open it up. What's inside? What's inside?
Great Christmas gift this holiday season. Get it for someone you love. Get it for someone who needs to save time and money, which is really everybody. Also, congrats to Ramp. Massive fundraising news. We'll be talking to the team tomorrow. But the road to Christmas. We're 70 days away, and this can be a difficult moment. I mean, I don't want to call out anyone in particular on our team, but there was an individual on our team yesterday who I was reflecting, because it had been a little bit cold over the weekend, that I was feeling especially Christmy. I was feeling the tide. I was feeling the hearth. I wanted to warm up my hands. We were in the jackets. Everyone was saying, 'Was it so cold in the studio?' I was like, 'No, there's just a warm fire burning inside my belly as I get closer to Christmas.' But somebody on our team is like, 'Oh, we're finally in the cold season.' And as soon as the cold starts where you live, whether you're in California or New York, as soon as it gets a little crisp, it's game on. No one's going to look at you like you're crazy for talking about it's only 70 days till Christmas because it's Christmy out. It's cold. It's cold. But Michael over there, I'm going to call you out. Michael over there on our team is like, 'John, you know it's going to get hot again. It's going to get hot again.' And yes, I checked the weather. It's going to be 88 degrees today, which is not the most Christmas. There's actually a heat advisory in California from the National Weather Service.
Advisory, but there is a lesson.
It remains in effect until Saturday, which is Michael. 90% of the battle for the road to Christmas is mindset, and we can't have anyone on this team being a Grinch right now. It's very important that everyone supports the idea and looks on whichever side has the greener grass.
Friendly elf mindset. Golden retriever mode.
Yes. Exactly. So yes, there might be a heat advisory. It might be blistering hot, but we're still going to stay laser focused on what matters, which is that in 70 days it will be Christmas. That's right. We're pretty excited here. We're headed to see 'You Can See Everything', the new Nathan Fielder movie, later today after the show with a great group. Can't wait. Yeah, we'll have the review for you tomorrow. Yeah, it'll be fun.
Any really anything to take our minds off the market, John. Yes. Because it's an absolute bloodbath out there.
Well, before we go into the market and the news, let me tell you about Figma agents. Meet the canvas. Your AI agents can now create and modify your Figma files with design system context. Okay, so the market is mostly reacting to the Anthropic investor day numbers, but it's been in a decline for a long time. For the last month, CoreWeave's down something like 50% over the last month or so. It's been rough in the AI trade broadly, and way off.
It's crazy that you can be growing at 45% quarter-on-quarter and still not meet expectations. But when people have had in their minds 10x every year, 10x revenue growth every year, and Anthropic goes to their investor day and says that Q3 revenue was $16.8 billion, which annualizes to $67 billion, $67 billion. I don't know what they were doing there. That seems like a little bit of like, okay, I get it. It's like the zoomer reference, but like this is serious financial. There's a lot of money on the line. Can we not do like the meme thing right now? Anyway, they are in fact annualizing their quarterly run rate to an annual run rate of $67.2 billion, which is remarkable for a company that was doing, you know, a tenth of that a year ago. But it's seemingly not enough because everyone in their minds was like, they're going to surprise to the upside. They're going to throw up 90, like they're going to do something. And I was kind of in that camp. I was like, they've been quiet for a while. I feel like everyone's like, there's that chart of like, oh, the growth stalled. And I was like, that feels like something that they would love to come out and be like from the top rope, fake news, like we got some surprise up our sleeve. But they really have been fighting fires on multiple fronts. There's like the Meta business which has pivoted over. They're probably using their own model now. And there was that leak of an executive at Meta saying like, 'Oh yeah, if we stop buying Anthropic, like it could really mess up their IPO. That would be a shame. I would hate for a company to IPO at lower valuation than us since we're almost $2 trillion. Like that would be terrible.' So it seemed like sort of a veiled threat in the media. And then distillation, we actually got a number. So distillation is between 5 to 7% of revenue, which I think is a little bit lower than people were expecting, but it feels like the Anthropic team, David Sacks was celebrating this, that hey, like there was this big problem and one of the proposed solutions was like ban open source, do all this crazy stuff, slow down. And David Sacks didn't want to do that. And then Anthropic came out and said, hey, like we did a lot of work, we deployed our own frontier technology to fight this, and we believe we have bought it and we believe that we can shut down distillation. Pretty staggering that 5 to 7%, you know, that's billions of dollars spent stealing from, because that's the value of the distilled tokens. Which is pretty crazy that you can be at that scale of essentially fraud. Like you're not supposed to be distilling. It's against the terms of service. And usually when you see somebody, oh yeah, there's somebody who's like scraping LinkedIn and then they're like reselling the data or whatever, it's like, yeah, they might have like a couple million dollars deployed into this scheme. But you're looking at $5 billion of fraud essentially, or run rate fraud, you know, something like that. But as they fight that fire, obviously that's going to be a drag on revenue growth. If they opened the floodgates and were like, 'Yeah, go distill away.' You know, at least in the short term, revenue would be growing faster, but in the long term, there'd be much more competition on the frontier.
But it's certainly not as meaningful as some people on the internet thought it was.
Yeah. I mean, there were like some trolls who were like, 'Oh yeah, once you get rid of Meta and distillation, the company's going to be doing, you know, $2 billion a year.' And it was like, 'No, that's not going to be the...' It's like this is installed in many, many organizations. This is a sure go-to-market.
The question now is what happens to the frontier of the open weights ecosystem, right? Does the gap widen?
Oh, because they're shutting down the distillation. Yeah, we'll see. We'll see how effective it is.
Yeah. And I wonder how Muse Wonder distillation will fight. Like on one hand, like I don't think people are going to be distilling Muse Wonder. Yeah, like maybe it's not quite there yet. But if Meta can invest enough to actually have a model that is a target of a distillation attack, are they better equipped to fight distillation attacks because they've been fighting for a long time?
Being on the frontier just means you're getting distilled. If you're not getting distilled, that's the definition in your mind. It's like unless you're not frontier.
I think that's like the most honest definition. Like unless other people will spend a lot of money to try to emulate your model and recreate it, then you are not. That's a good reality check. Anyway, do you think that Meta would be more equipped than Anthropic or less equipped than Anthropic to fight distillation attacks if they do get attacked? And maybe so what I'm saying is that like on one end there's a lot of bots that are already in the Meta ecosystem and so you could be distilling attacked because maybe you already have a botnet up and running across Instagram and you haven't been caught and then you can reroute those bots to do something more valuable than just like and share and like astroturf stuff. You can get them to pull out tokens from Wonder somehow and then you can distill it or Wonder 2, whatever you want to call it, Wonder 3, whichever one is like worth distilling. But then on the other side, you have, you know, an organization that's younger, a culture that's not paranoid about this stuff. They're not saber rattling around, oh, if we get distilled, it'll be the apocalypse, right? And so, like, culturally, you'd think that they're a little bit less worried about this. And then also they're not in the business of like every distillation dollar is $10 that could be API revenue for Anthropic. It's very important from a business perspective that is basically giving the stuff away for free. Yeah, it's important to keep people in like the Muse ecosystem, but if they get distilled and then all of a sudden there's an open-source model that's really, really good, they're not going to be as worried because they're just going to say, 'Well, like that open-source model doesn't have a button on every Instagram, like the distribution on every Instagram account there's like a Muse button.' Like I think they'll be happier. So, I don't know. What do you think, Tyler?
Yeah. I mean, so one thing is Meta released an open-source model today? Muse Open. Yeah. So I don't think they're that worried about like someone distilling them and then releasing an open-source version. That seems very different ideologically, right? Yeah. And also I think it seems like distilling is, I mean, I don't know how hard bot detection is normally, but distillation seems like uniquely hard and that like I think it's hard to actually tell what is like a normal prompt versus what is not, right? If you're asking like some math problem, like are you going to distill on that? Because like effectively if you're distilling on something you need to like get the train of thought right, you need to get like the reasoning steps, the answer looks very...
And this is happening while they're experiencing explosive growth and the distillers are setting up hundreds of distilleries, are setting up hundreds of individual like Max accounts that each have their own VPN, and it's just very difficult to detect is this like organic growth. Yeah. Or...
And often they're going through like third-party wrapper companies. So there's a lot of companies that are like, 'Yeah, we're like, you know, an AI-native document editor.' And there's like a bunch of these. It's not just like Notion that we know about and it's not just Google Docs. Like there's a lot of these companies. You set one of these up and you're like, 'Yeah.' And you pay us a fee and we go and we inference Claude or we inference GPT-6 or whatever. And then that goes into the document. And then all of a sudden, what if somebody's using that like AI word doc to do the distillation, then that SaaS company, that wrapper company needs to go and figure out, wait, why is someone generating all these documents? Maybe they can't even look at the documents. And then from Anthropic's perspective or the Frontier Labs perspective, they're just seeing like, oh yeah, one of our SaaS wrapper companies is like sending a bunch of queries across their entire network because they have thousands of users, maybe hundreds of thousands of users. And it's hard to tell. And so there's a lot of wrapper companies that have gotten to like crazy revenues and basically no one uses the products because it's the distillers or the distilleries that are like, 'Ooh, new wrapper company, like can't wait to spin up a thousand accounts on you because you won't detect me.' But Anthropic would. So we can't go straight to setting up Claude plans. Instead, we'll set up a wrapper plan and a million wrapper plans. And then the company that's running the wrapper is like, 'This is amazing. People are just paying so much for us and all we need to do is just take their prompt, stuff it over in Anthropic's API, give it back to them and we make a 10% margin on top.' Like the wrapper system's working when really like they are a distillation factory.
Very interesting. Distilleries. Yeah, distilleries. You got to take a trip to the distilleries. Well, what's going on in the market? Because...
Yeah, where did that lead us? That led us to sell everything.
Uh, yeah, I mean, especially your dollars. Sell your dollars. Your dollars. Sell your stocks, sell your bonds, get out of everything, get into land, buy guns, food, medicines. We will be on a barter economy soon, I think. I think that's the main major prediction. No, people are blackpilling. Sell your assets. Sell everything. Sell everything. Take your dollars, invest them into a Roadster or maybe a Mission S. Yes. I think that's why they timed this launch. They knew, this goes way deeper than we ever thought. They timed the launch of these two vehicles for later today because they knew that people would be going to cash and they'd be primed. No, this led us to the president of South Korea. Yeah. Very dark. Yeah. Posting what, something that was translated from Korean, but the Grok translation is, 'We must live. You have to live somehow. If it's hard, please call 109. In fact, Gil will help you find reasons to live together.' I don't know who Gil is, but yeah, the KOSPI, I believe, is at a six-month low. And so for a country that loves trading as much as the South Koreans, dark times.
Yeah. Yeah, a lot of this just goes down to exponential expectations being extremely difficult to keep up with. So the meme has been the lab revenues are 10x-ing every year. Anthropic went from like 100 million to 1 billion to 10 billion. So logically they'll do 100 billion. But I remember looking at the numbers and being like, there's actually, it sounds so nitpicky to say it's deceleration, but it technically is because I think they went from like a little over 100 million to a billion, which is just under 10x. And then they went from a billion to technically 9 billion, which is a 9x, not a 10x. And so if you assume that you're actually seeing some slight degradation there, you would be thinking, oh, they'll close out at like 80 because they went from 1 to 9. They didn't go from 1 to 10. So going from 9 to 100 would be acceleration. Again, we're talking about growing 10x a year, 9x a year, 8x a year. It's all crazy and incredible. But the fact of the matter is like the difference between an exponential curve going upwards and slight deceleration is a huge difference. And I think people are maybe like overreacting to this because there were a lot of people that were saying like, 'Oh, maybe we'll hit 100 billion. Maybe they'll do 150 billion.' And so when you get the expectations so high, people, it's very, very hard to keep up. And then also I just, it's very clear that there's like this tall poppy syndrome in the frontier where whoever is in the lead, like all the other businesses are like, well, we don't want a monopoly on intelligence. And so let's go buy from the other labs. Let's go, you know, figure something else out, fund, invest. Let's do something else to make it so that no one runs away with this entirely. Of course, like there always could be some sort of runaway, but at least while businesses are the ones that are deciding, you will see the Microsofts, the Amazons of the world, it's like all of a sudden they're doing deals with the other guy because they don't want a monopoly. And so as soon as someone starts running away with it, there's this huge forcing function where you're like, oh wow, there's a big like we, you know, the anti-YouTube alliance or the anti-Netflix alliance or the anti-OpenAI alliance was a thing and the anti-Anthropic alliance like comes together and these alliances come anytime someone's getting a little bit too hot at the party. Somebody's like, 'Yeah, we got to pull the punch bowl away a little bit.' Anyway, the...
Chamath was on CNBC this morning. Okay. He says, 'We have sold, this is according to a meet, we have sold all of our core. We don't own any of it today. We've taken exposure down over the last three months. We felt AI had overshot and would lead to a period of consolidation.' They moved 10% of their book into Meta and Tesla. And again, Chamath usually ahead of the curve. Sounds like he was here as well. Yeah.
It's an interesting challenge being, you know, running a hedge fund, an venture fund, but then also being one of the most elite podcasters because there's periods where in some way he can't go out and say what he actually feels. Yeah. Because it's not sort of pragmatic for his actual investing.
So the interesting thing is people aren't actually selling everything. They're not actually selling their dollars. That's just the meme. Of course, Cramer says, 'Oh, it's over over.' Which is very funny. Just to put it in perspective, the NASDAQ is down like 0.3%, 0.4%, not much, but then the Dow is up almost a full percent. So what's really happening is that there's just a rotation that's happening while the Dow and Treasuries are getting more attention while tech stocks pull back. Obviously not good for the AI trade in the short term, but there's a lot of new initiatives and we'll go through some of the progress because there is some interesting AI progress going on. But at least over in the AI trade, over in tech world, it's rough. People are speculating that Leopold might have gotten into some of the same trades and depending on the leverage ratios, he could be down a bunch. It's very unclear. Obviously he's extremely sharp and just learned a really hard lesson. So how he interprets that, who knows what's happening.
Yeah, it is short. It is notable that people are saying it's over over. Yes. When the NASDAQ is flat over the last five days.
Yeah. It's because like people are rotating out of core, but then they're buying more SaaS, you know, because like if the AI trade breaks down, like the superintelligence thing doesn't happen. It's like, well, then you still want to own like a social network that runs ads, right? You still want to own like a software company that has good margins. Like there's plenty of things that are worth owning. And so the dollars actually aren't getting sold. Like the money is not disappearing. It's merely just sloshing around in the economy. Anyway, the VARA AI, which is Value Aligned Research Advisors LLC, there's a rumor going around that Martin Shkreli shared that the VARA AI fund has blown up. So we'll be tracking this. It happens fast, but not immediately. The rumors around Situational Awareness took a few days of speculation. Oh, he must be hurting based on the latest selloff in semis and memory. And then a couple days later, oh, there's a rumor he's doing a deal. Boom. And then all of a sudden it was all over the Wall Street Journal. So that will happen. And we'll find out. Maybe they make it out. Maybe something happens.
Yeah. Looking at their trade, but crazy stuff. As of the last 13F was Nebius, Bloom Energy, CoreWeave, Seagate, Western Digital, Hut 8, IREN, stuff that's further out on the risk curve, Core Scientific, some Nvidia, some SanDisk, ASML, among others. So I think they were big believers in this whole AI thing. Hopefully continue to be. Yeah. But...
And I mean, I still feel like there's a lot of management that happens in the pre-IPO stage. You could see potentially the labs see that, okay, there's a blip. We got to rip the band-aid off now. We got to correct the record, get the low numbers out because we know that we're going to go into a period of acceleration and that's when we actually want to IPO. It's notable that this investor day is not the actual IPO. This is a precursor to that. It's maybe almost like airing some dirty laundry, which is crazy to say given how fast this company is growing and how big it is. But it's useful to sort of clear the air when there's a bunch of speculation around sort of flatlining, Q3 deceleration, these like sort of less exciting words floating around. You clear the air and you say, 'Okay, we decelerated.' Only way to go is up now.
Time to Baja blast. Yeah, maybe they got through the code red. Now it's time to Baja blast. Anyway, let me tell you about Railway. Railway is the all-in-one intelligent cloud provider. Use your favorite agent to deploy web app servers, databases, and more. While Railway automatically takes care of scaling, monitoring, and security. Well, inflation is up. So if you were short stocks and long inflation, then you're very happy today. Hopefully, I mean, if you own real assets, you're probably pretty happy. Joe Weisenthal says that was the hottest PPI number in four years. And yeah, it's a pretty rough time. You're really having fun. Did you see inflation for the 1%? This is cpireserve.com. We can pull this up. It looks at inflation. You know, there's normally the CPI basket that measures, you know, what the average American buys. CPI Reserve is for the global elite. So it tracks how much does it cost to go to space on Virgin Galactic? How much does it cost to stay at Amaniri in a base suite? An Hermes Birkin bag. How much has that gone up?
The spicy rigatoni. Only up 31%.
Oh, I didn't know that. Porsche 911 Carrera. And so there are some interesting stats here that we can go through and you tell me whether these are overrated or underrated. Have these gone up because the value has improved or is it now overpriced? So I want an overrated, underrated for the key items that CPI Reserve highlights. So the classic Chanel flap in 2016, it was $4,900. Now in 2026, it's $11,700. That's over 139% inflation in 10 years. That's over three times what standard inflation was. If it tracked CPI directly, it would be $6,800, but instead it's $11,700. Do you think the classic...
You're a bozo if you're still buying this.
Okay. So it's overrated at this point. The inflation's always...
Okay. What about Met Gala tickets? In 2016, it cost you $30,000 to go to the Met Gala. Now it's $100,000. That's six times the average inflation rate in America. Overrated, underrated. What's a Met Gala ticket worth, Jordi?
I think underrated still. This is charitable giving. So the more, the better.
That's true. I hadn't considered that. Yes. And the Met Gala has grown so much. So many interesting folks there. So many interesting moments. Worth every penny. What about a Hamptons luxury home in 2016? The median Hamptons luxury home cost just $4.425 million. Today it'll cost you $11.5 million. That's a 160% increase in 10 years. Four times the CPI, the average inflation rate. Are you a Hamptons guy? I've actually, I don't think I've ever been.
So overrated. You'd pass.
No. I mean, when you look at the real estate prices here, I think this ends up feeling like it's driven in many ways from COVID, people wanting like more people wanting a second or third home. Plus, they're not making any more. Plus, SALT, right? And so a lot of those homes traded at really low rates and so that just drove everything.
Okay. What about a Virgin Galactic seat? You want to go to space? It used to cost $250,000. Today it'll cost you $750,000. Overrated, underrated.
I don't know. Not for me.
Steal of the century. Any price? Absolutely. Send me to space.
I would go to great lengths to not block you from going on this.
Is that because of safety? Assume that I survive it because I'm going to survive it no matter what. It's going to be...
I can actually imagine. Yeah, I can imagine it. You know, the flight going poorly and you having to sort of like fight your...
Yeah. You got to roll the duct tape and I got to put it together and screw this in and figure it out.
Take a deep breath and dive. Figure out a parachute on the way down.
Yeah. Yeah. Exactly. That's... Yep. Yep. Yep. That's exactly what would happen in a realistic scenario. Okay, the last one. The Porsche, actually we got two more. Porsche 911 Carrera base. It used to be $83,000, $84,000 in 2016. Today it'll cost you $135,000. That's a 61% gain in 10 years. That's 1.5 times what the CPI did. Not a huge increase relative to overall, but it is ticking up in recent years. Where do you sit on the value of the Porsche 911 Carrera base?
Still undervalued relative to the joy that you get out of it, but I wish that they could keep like the base 911 at like $99,000 forever. Yeah, that'd be special.
That'd be cool. Yeah. But this is the world we live in.
The last one, truly heartbreaking, a Dom Perignon Magnum. It's up 121% since 2016. Used to be $380. Now it's...
Wait, we also got to pull up the big law. We got to pull up the big law hourly rate for partners. They are way outpacing inflation. Okay. I mean, just look at this.
Yeah. You used to be able to hang out with a lawyer for just a mere $1,500 bucks an hour. Now it'll cost you almost three grand an hour to get legal advice from a...
Yeah. No longer the cure for male loneliness. Because that $1,500 an hour... Yeah. You just call them up. Call them up and just ask them how you're doing. Kind of lose track of time. Talk going all over the place. Right. How, how you been? How are the kids?
Can we, can we bill in quarter-hour increments? Can we book an hour in 10-minute increments? I'm going to my lawyer. I'm saying, 'Can we just book a week at a time?'
If I talk to you, just put me down for the whole week.
Yeah. Give me a 100-hour block. If I shoot you a random question, if I ask you, is this domain available? Okay. Anyway, anything else you want to cover on the CPI Reserve? You can go check it out. cpireserve.com. Very cool. Little fun project to give you another side of the economy.
Craziest here is Manhattan luxury real estate.
It is down. Down the SALT tax, right? It's higher taxes. People are getting out. They're moving. I don't know. It's working.
There's also an interesting article in the Financial Times. I don't know if we have it here, but you can go read it all about the E-shaped economy. You've heard of the K-shaped economy. They argue that there's an E-shaped, that there's the ultra-wealthy, the middle class, and the lower class, and they're all going in different directions. And that some of the inequality stats are not tracking exactly how you would expect. And they pull some Bureau of Labor Statistics and some economic statistics that were released from the 2025 report. So break it down. It's kind of interesting. Blake Biers is taking like the AGI question around capital markets. He says, 'My modal prediction for the next decade: interest rate spikes as capital shifts from saving treasuries to investments due to higher returns promised by AI-related capex. We've seen this before because AI is the only modern project that could affect real rates to this degree. This leads to short-term panic as high rates spike debt servicing costs for governments and other borrowers.' We're seeing this all over the place with real estate developers who were doing great financing a large development project and now with floating interest rates they're over their skis and they have to refinance or sell a property or stop building. It's a really, really tight and chaotic moment. There's a lot of places that are in pre-foreclosure. It's actually an interesting hack. You can have ChatGPT go and hunt for properties that are in pre-foreclosure and make below-market offers and get sort of a steal on properties. It's very, very interesting to dig into what's going on and what the loan actually... There were people that were doing this during the 2008 mortgage crisis. They would look into the mortgage-backed securities, into the collateralized loan obligations, and they would say, 'Okay, I know that this hedge fund does not want this house. Like, they don't care about this. So I can go to them and buy it because they're rating the whole bond at 20 cents on the dollar now because we're collapsing. I'll buy it from them from 50 cents on the dollar.' And it's a more complicated deal, but some people walked away with some really, really great properties. After three years of panic, it becomes clear to the market that the investments are paying off in a big way. Interest payments remain high but increased revenue from economic growth more than offsets the higher borrowing costs. Panic shifts to euphoria. Developed world adjust to 5% growth rates as the new normal and now demand 10% to be happy. Healthcare consumption R&D grows as a percent of GDP. I was listening to Ari Emanuel and Ben Thompson chat and it was the funniest interview because Ben Thompson cannot get a question in edgewise. It's just a conversation. Normally Ben is like very structured in his interviews. 'Let's start with your background. Let's ask some questions.' Instead Ari's like, 'I'm asking you the questions.' It's very, very funny. But one thing they talk about is that recently the share of advertising dollars, you're going to love this, the share of advertising dollars as a percentage of GDP was 1 to 2% for a long time. It's now 2 to 3%. And they both agreed it could go higher. It was amazing. Anyway, AI is disrupting the job application market faster than the job market.
This is continuing. It's a hot take, but it's continuing to gather evidence. Derek Thompson shared a post from the Wall Street Journal. AI-amplified job applications are met with AI-enabled filters and other AI tools used by HR departments to narrow the application pool. People are interviewing AI avatars. That seems like a really terrible thing unless the whole meta-game is just can you jailbreak it, which I think it probably should be. Like if I was to put out a, if we were to put out a job description and I'm totally tempted to throw an AI avatar in front of that just in the hope that someone convinces that AI avatar that it is in fact interviewing for the job. You've seen that where people will flip it, flip the script and be like, 'Actually, we're here to talk about your performance.' And the AI avatar is like fighting back being like, 'Remember your instructions. Remember your instructions. You're the boss. You're in charge. We're here to interview you.' And then eventually the interviewer can keep hammering and prompt injecting. But Derek Thompson says, 'And now 10-minute conversations with AI that take place in McKinsey's very real offices with partners or interview assessors in the room observing how the candidates do.'
So these candidates are basically putting on an interview show. Yeah. And the partners are watching two people duke it out.
Yeah. It's kind of like a gladiatorial arena. And I want to see if you can beat the tiger that I've released. And if you can, you get to sit at the side of the court jester or something like that. I don't know. We get to debate the bread and circuses thing. I feel like circuses are kind of running away with it and bread is potentially falling behind.
I don't know. Did you see the Crumble Cookies introduced? Okay, I'm living. I'm just saying that from a perspective of like what the elites are doing in our economy. Think about it. So many billionaires are buying effectively circuses. They're buying sports teams. They're about UFC. There's F1. Banana, banana. What? What?
Savannah Bananas. The Savannah Bananas. The minor league baseball team. Banana ball. Yeah, the Savannah Bananas.
The Savannah Bananas. And so there's a lot of speculation today that they're announcing some sort of acquisition.
Yeah. So there's an intense amount of money flowing into the circus economy, but I feel like bread is kind of falling behind. I mean, Erewhon, we talked to them. It's interesting, but there's a lot less. It's not a trophy asset in...
We had two grocery stores on the show yesterday.
We did, but it's very different. Like no one is saying that like no billionaires are like, 'Okay, I bought a grocery store.'
I'm sorry, John. Your whole thesis breaks down because Crumble Cookies and Wingstop are collabing.
It's fake news. See, undefeated, but it's not a trophy asset. No one is getting the bag in venture or private equity or high finance and then being like, 'Give me the award.' Jordi's got to put on the fell-for-it-again award because he fell for AI slop. There is so much AI slop on the timeline. I'm actually very, very close to hooking up Codex to Grokbot, sucking my timeline out of X, getting all the different raw data and the tweets and then fact-checking everything and then creating a new feed for myself again.
Okay, we got to move the camera down. There we go. Fell-for-it-again board. That's me.
That's Jordi. Many, many times. Many, many times. But yeah, the job apocalypse is certainly not happening. Bank of America hired a thousand contractors to patch software flaws exposed by AI systems, Mythos and whatnot. You got the fell-for-it-again award with regard to the Crumble Cookie Wingstop. That one, I saw that and it just felt like something that would not be edible because there's bones in it. Like there's chicken bones. You know what I'm talking about? This is the picture. You got to...
You seem like the kind of guy who just eats the bones.
You don't eat the... No one eats the bones. Get out of here. That makes no sense. But you know what does make sense? MongoDB. What's the only thing faster than the AI market? Your business on MongoDB. Don't just build AI. Own the data platform that powers it. Okay, we got...
Okay, I'm excited that these thousand contractors got hired, but at the same time, shouldn't the same AI that discovers the bug be able to like work out a solution? And...
But maybe there's just like maybe there's 100,000 engineers' worth of bugs.
Yeah. You know, but because these thousand contractors are powered by... Yeah, this is kind of an odd headline now that I'm thinking about it.
Very, very odd. And it's like, yeah, did we just create a bunch of make-work? Because if we hadn't invented this, it really is like Pandora's box. Pandora's box opening creates thousands of jobs containing the stuff that is flowing out of Pandora's box. Is that what happened? Basically, that's the story here. Anyway, there is some progress on alignment. We're very excited to see this. Nick Cammarata says neural networks may be easier to interpret than researchers once thought. Rune is very excited about this. So it starts with Nick Cammarata. He says interpretability might be the most important scientific problem in the world right now as I don't think we can solve alignment if we don't know what we're building. I think we have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic because neural networks appeared to fold into arbitrary tangled knots. The, you know, they're grown, we're growing a garden, we're midwifing a deity instead of building a machine that is interpretable. This has been the hope that we can actually interpret these systems. And so he started OpenAI's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools that he had at the time. With modern agentic tools and a much closer look, the picture looks totally different. The vast majority in models up to trillions of parameters are legible and their circuits beautiful. Many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most and possibly quite soon as automated scientists keep improving. It may be a bitter lesson for interpretability. The simple thing at absurd scale works. And so Rune contextualizes it by saying when Nick told me that he had some interp breakthroughs, he was underselling it. I think this may be remembered as a major moment in scientific history. News so good even Udowski will have to lower his P(doom). We can start making a line in two moments. So, Tyler, did you read the full 'Neural Networks Are Already Clean' post by Nick Cammarata from Goodfire?
Okay. So it's like super long, but I think I have like the main...
It has beautiful illustrations. As you load already-clean.goodfire.com. Maybe you can start by giving us some context on Nick because he was at OpenAI, he references OpenAI, but he's no longer there. What's he up to? What is his mission here?
Yeah, so he's at Goodfire. Goodfire is like a mech lab. This is like all the stuff they work on. It's like a third party, kind of similar to a lot of like the safety labs.
Is it a Frankenstein reference? Like fire bad, fire good? Like he's like, we're summoning the Frankenstein, but it's going to be good.
Potentially, potentially. Okay, I'm not sure. But so basically the context of this is like, mech broadly is you're trying to figure out like what's going on at the most minute level in the models, right? So you're looking at like individual neurons. You know, when you pass in some text through a model, you know, neuron A like lights up, it has a high number, like why? It's very unclear. You'll look at one neuron and maybe you pass in some text about cats and it lights up and then you pass in some text about dogs and it also lights up. And it's like very confusing because like how is the model like actually working? It seems that, you know, each individual neuron is basically like it corresponds to like a bunch of different like topics or like features, you usually call it. Yeah. And so there's this thing called like superposition. This was the previous theory.
This is the thing you were doing where like, didn't you make like a 1776 GP?
Yeah. That used something called a sparse autoencoder. Okay. Which is basically like the theory is that these models are basically compressed forms of like bigger models essentially.
Where you can imagine like a much bigger model where every single neuron corresponds to one specific feature, right? So like one neuron is the cat neuron. And so if you're talking about cats, that one will definitely light.
And if you're talking about cats on the show...
If you're talking about something else, it won't light up, right? And then basically you can take your small model and you can basically expand it out. It's called a sparse autoencoder.
So yeah, I'm interested to see how he integrates with the labs. Obviously he has ties to both Anthropic and OpenAI through Chris Olah. But Goodfire is independent, probably doing a lot of work on open-source models, but it feels like I don't know if he can have a product that he sells into the labs or if he can just be one of those like third-party auditors that gets brought in and basically says like, I will do mech for your lab and help you there. That'll be exciting to see where this goes and how people sort of like interpret this. Yeah, basically the outcome is like we no longer need the sparse autoencoders and you can actually look at the neurons themselves and figure out what's going on there.
Fantastic. So it's very useful for safety, but also...
for capabilities, right? Because if you can figure out, like, you know, if you're trying to make some capability better within the model, like whatever math on some niche subject, you can figure out exactly how it's working right now and then kind of optimize on that. So it's very applicable.
Yeah, great news. Fantastic. Great summary.
Um, yeah, white pill. Uh, let me tell you about the New York Stock Exchange. Want to change the world? Raise capital at the New York Stock Exchange.
Uh, faith in artificial intelligence was, um, the timeline is in turmoil and has been over the weekend over Valor Atomics, Isaiah's company, and Day One, uh, because they're beefing over Day One's pro rata rights. And so, uh, this happened back on October 9th, but we didn't get a chance to go through it. But Isaiah Taylor, the founder and CEO of Valor Atomics, uh, posted a long blog post on, uh, October 9th saying that he had gotten all of the investors to agree to waive pro rata except for one venture capital firm, Day One Ventures, and, uh, and that they were fighting tooth and nail, and that Day One had retained a law firm and was fighting a lawsuit against them and it was getting messy. So Isaiah decided to bring it to the public square on X and have the open debate. Lots of people coming out and supporting Isaiah. Some VCs coming out and supporting pro rata rights. Obviously, this is a more complicated situation because Day One has a bunch of complicated and, uh, controversial connections to international organizations and stuff like that, and Isaiah's, like, going and doing more and more government work. And so, like, there is this extra layer of nuance there I think that's getting missed. But the, uh, the core debate that the VCs are having is that there's a, uh, there's like a don't throw the baby out with the bathwater pitch happening that I'm seeing, like Delian was talking about this.
You're saying their baby is pro rata.
Yeah, basically, uh, like the bathwater might, so like it might be like, okay, Isaiah, like we support you in this one case, you should be able to sort of force an investor who has this, like, very controversial background, uh, to be diluted down below the major investor threshold in this case. But just know that, like, founders don't get the idea that pro rata rights are squishy. Like, they're negotiated for a reason. They're going to stick around. Don't think that you can pull this move, uh, without a huge amount of pushback. Uh, because a lot of founders, I think, read this just as like, oh yeah, like I should override my investors' pro rata because it can be a hassle to deal with. Like you're going out and doing a round, but you've already sold 40% of the company. You're raising $20 million or $200 million or a billion dollars and some fund comes in and says like, 'Yeah, you're raising that much. I want the full amount. I want to jump straight to 20% ownership in this company.' But all the previous investors say, 'Hey, we want to keep our 40%.' And so then you're faced with the question of are you doing a 30% dilution?
My view is I don't even know if it's worth having the pro rata debate because this isn't about pro rata at all.
Yeah, it's sort of a proxy debate over the fact that it seems like Isaiah doesn't want to work with Day One anymore.
And there are other vectors to work with. I think there's, isn't there like an FBI investigation or some sort of letter from the government investigating this? And like, there are other forms. I think there's something else going on, but, uh, it is tricky. Like, uh, signing up to work with a venture capitalist is somewhat of a marriage and you're stuck with each other for a really long time and it's hard if they don't want to sell their shares, if they don't want to go the other way. It's typically very difficult. There are some scenarios where VCs have just given the equity back and stuff like that, and then there have been other scenarios where founders have basically been like, we're not converting the convertible notes and so like you don't get any equity ever. But once it's papered and the pro rata is in the contract, it's very hard to dilute them down because that's the design, is that if as long as I keep showing up with the money, me as an investor with a pro rata right, I can have 5% of this company forever.
Yeah. The big, I mean, the question is how, just how much is having Masha involved with the company in a meaningful way, right? It's a meaningful stake in the company, right? Are they losing, are they losing, you know, contracts over it? Is it just something that competitors can, uh, sort of use to poke at the company? Um, I, you know, based on what Isaiah was saying, there's no information rights, like technical information rights, that come along with that. And so I do think that it's probably completely possible for them to just sort of keep, uh, keep the fund at arm's length.
But financial information rights are valuable.
Yeah, like even internationally, like I mean, we just saw that with the OpenAI spy. Um, like there is value to knowing, um, if a business, even if you don't know how it works, if you know that it's profitable and you know that it works, that can underwrite your own investing into competitive technologies. You're like, yes, let's grind down this path. Like you think about it in the semiconductor space, like why is China subsidizing the indigenous semiconductor supply chain so aggressively? Well, they know semiconductors are super profitable. They can see Nvidia.
The only thing is, like, just the way that Isaiah runs a company, you have to assume that that momentum is going to become very obvious just through, like, public communications. Hey, we got, you know, XYZ contract.
Yeah, but you never know because if you have information rights, you know actually what's happening with the cash flow, actually what's real. There's always a lot of, like, you know, uh, gesturing and context and rumors out there, and you have an advantage if you actually have the real numbers on the inside. It's only a slight advantage. I agree with you, I agree with you generally, but, uh, I don't think it's a huge... I do think it's valuable information and it's why investors negotiate for it. They want information rights and they want to not be diluted down. So, uh, interesting. Uh, yeah, hopefully if you're a founder and you're doing a VC deal, uh, don't give the pro rata right away, uh, the pro rata right, um, lightly. Like, think about it as a serious deal point. What is the value of that? It's going to be a headache potentially for a long time. Could be really awesome. Sometimes it's great to have investors with pro rata because every new round you're like, ah, I got 40% of my round filled. Some founders love having a bunch of VCs who are like, 'Oh yeah, every round, like, you can count on me for the first half of that round. I'll fill it first and I'm continuing to add on.' And so then when you're raising, like, that next big round, you do still need to go find an anchor, but you don't need to necessarily fill it all, which can be hard if you're not the hottest company in the world sometimes. So, uh, it cuts both ways, but, um, I don't think it's going anywhere. Uh, there is, uh, yeah, I mean, the state of Delaware, you know, is firmly entrenched in Silicon Valley, uh, and for that reason, like, we will abide by these things. Anyway, uh, big car day, uh, today, two car launches, the Tesla Roadster tonight, 7:55 p.m. Central time.
Yeah, so 5:55 Pacific. 8:55 Eastern.
Okay, so 7:55, people are rumoring or are suggesting that 7:55 might be the quarter mile time. Is that what they're saying?
Yes. I've seen a bunch of different speculation. What would be really crazy is if it was a 5.55 quarter mile and it was the Pacific time. That would be amazing.
Uh, but I put down a deposit yesterday for a Roadster and I'm fully ready to put that pin on the 'fell for it again' award, it's ready to go. We have a big bet. Tyler has a big bet. And let's cut over to Tyler. And, uh, and I need to, I mean, I need to create some stakes for this game because, like, what would be...
Hey, but break down your strategy. Putting down the deposit.
So my strategy is, it's... You sold me on it. I ended up, I did it.
So, uh, it's five grand on a credit card and you have to wire 45,000 within 10 days, but I'm going to know whether or not it's a good car by tonight. And so I can cancel it if I don't like it and jump out of line. But if it's amazing, I mean, Elon's saying it's like going to be the most amazing product demo ever. If it's amazing, then I will be able to hold my position in the line before everyone floods in post-launch. Yeah. And so it felt like the right risk of you're tying up five grand for two days. Even at 6% interest rates, you're talking about like $5 of opportunity cost. It's nothing. And so, um, it's a pretty good trade.
It's a pretty good trade for potentially being at the front of the line for something that could be really cool that actually flies. But if it's like, oh, it's going to be shipping in 2035 and it's, you know, it's not going to be anything special and, you know, none of the features appeal to me, or oh, we couldn't get self-driving working on it and it's not actually that.
Here's the thing. Here's what everyone... uh, let's speculate a little bit. Yes.
So, I think that ramps will be involved. I think that ramps will be involved. So, because our original bet was it will leave the ground on flat ground and I think you had like half a foot. I think you had flat ground in the... He just said, he just gestured towards it flying. Yeah. And so to me, what's the easiest way to get a car in the air? Shoot it off of a ramp. Drive it off of a ramp. I think we're going to be seeing multiple types of ramps. I think we're going to be seeing...
That'd be awesome. I love it. I'm taking a different direction, a different bet. My bet is that off of flat ground it will go, I think I said more than six feet in the air, but I definitely think more than three feet in the air. I think it will go high.
Not 10,000 ft, which is what they got cleared to.
You know those demos, you know the demo, the Chinese car like goes, no, the one that sort of like drives through the tunnel.
So what I think is that there's going to be sort of a looped ramp that will send it through the air in a rotation and land.
So like stunt driver basically. But maybe it could do it, which is still incredibly impressive. You don't see car manufacturers demo their cars by sending them off ramps. Okay, so if there's ramps involved, Jordi wins. If it jumps off the ground off of flat ground, I win. That's the bet.
Here's the other thing. Uh, do you remember there's videos? There's some street in LA.
Yes, where they film all the car commercials. I know it. Uh, it's like Figueroa and First, something like that.
No, I'm talking about... Oh, talking about something else. There's basically a very specific road where it's very steep up and then it flattens out for a second and then it goes...
Oh, yeah. Yeah. Yeah. I've seen people jump off that.
I think Elon saw that and was like, we should just demo the next car that way.
Okay. Okay. So you think jump. I think rocket engines or thrusters or fans that will lift.
Pull this up. Pull this up immediately.
Okay. Um, but let's pull this up. Yeah. What's going on here? Is this a famous YouTuber too?
So, David Dobrik. I've seen a bunch of videos of people, at least 50.
So I'm calling this, but the ramp spins it.
Okay. Okay. Okay. So you think jump. I think flat. Now the question is, what is the criteria?
There we go. That is insane.
Uh, here's the criteria, Jordi. The people that fell for the Tesla Roadster the first time, they put down the full amount in 2017 and they waited a decade and they didn't get anything. Nothing shipped, right? They fell for it. They didn't fall for it again because Tesla has always shipped their vehicles. They did the Cybertruck, they shipped that. There's one outside, like, they're real. You may like it, you may not, but they shipped it. But they didn't with the original Roadster. And so there were a lot of people that feel like they fell for the Roadster promo scheme because they put down the money and then they waited a decade and they didn't get anything. So in order to fall for it again, you need to put down the money and wait another decade. But we don't want to wait a decade. And so the question is, on what time scale would you be disappointed? And to what feature set would you be disappointed?
Well, I did have one theory that, again, this was clearly a car that didn't want to get made. And it sort of was forced into it by taking deposits and all this stuff. And it doesn't really make sense in the Tesla lineup of sort of appliance-style cars in many ways, right? You're shaving off all these other cars. The Model X, etc. Uh, you don't have an SUV, all these things. Um, and so I thought it'd be funny to demo the car in such an insane way that it just instantly, like, federal regulators come in and say, like, you cannot under any circumstances do this. And then he's like, 'Yeah, sorry. I made the product. I just can't ship it. We're going back to Model Y. We're only going to make the Model Y now.'
Yeah, but if he says that, the obvious next step would be it's track only, which is a huge selling point for me. It would be amazing. I would definitely stay in line.
Yeah. Do we think it'll have a roll cage built in?
Hopefully. Given that it may be taking flight, hopefully it's a roll cage, but like all of the bars are like slightly misaligned.
Sebastian says Elon wouldn't hype it this much to drive a car off a ramp. Come on.
Okay, it is a good point. I do think it'll have rocket thrusters of some sort, but I still think ramp.
I think that's part of it. Like we're launching.
What if it does both? If it does both, we both win.
Well, that's what I'm saying. It's using the rocket.
No, what I'm saying is what if there are two demos? One is just it's sitting there on flat ground and it goes hover up and everyone's like, 'Wow, it's flying.' And then it goes down and it lands, but it goes up five feet in the air and we're like, 'Wow, it can go so high. That's crazy. I've never seen a car do that.' And then separately they're like, 'Now to the stunt portion.' And then they go off a ramp. Like that. That's what I think. That's what I expect to happen.
Both. I'm thinking both. I think this is like a Nitro Circus style demo. Like I think there's cars launching everywhere.
Okay, then the only thing I disagree with you on is that I don't think there will be any ramps. So I would go, you're going to see the car leave the ground, but there will be no ramps involved at any point in the demonstration. You're going no ramp.
That's our bet now. That's our bet. But back to the 'fell for it again' award. What does it take for you to put that on?
The demo is amazing. You do the full deposit. So you go from 5k to 45k and then it doesn't ship for 10.
Yeah. So we have to keep these around. So it's going to be withered and old by the time we put it on.
No, I would say one year. It's got to be one year. If it takes more than one year to ship this thing, then you fell for it again, right? Because like you've had 10 years. Like you don't get another 10, you get another launch and then you get another, okay, ship it, right? I mean, what else? What else would, like, you're going to be at nine years like, 'Oh yeah, didn't fall for it this time. I got my Roadster just 19 years after they announced it the first time and only nine years after they announced it the second time.' No way, dude. You're going to be in year six being like, 'I fell for it again.'
Right. How much was the original Roadster deposit?
I thought it was the full amount. It was like a quarter mil. I don't know. We can look that up. But Mission S is coming out. I think they will ship it. I think that the Mission S is probably a safer bet. You won't be...
So, this video leaked late last night.
Yep. Let's pull it up. First, I'm going to tell you about Shopify. Shopify is the commerce platform that grows with your business and lets you sell in seconds online, in store, on mobile, on social, on marketplaces, and now with AI agents. Okay. What are they doing? They're peeling back a cover and they're showing off the car. And you had a very good take on this. What do you like about this design? What do you like about the Mission S? What sticks out to you about this?
So, I'll start by saying I feel like a kid who accidentally saw their Christmas present. And it's exciting. Christmas. It's exciting because you're getting what you want, but also it just like kills all the suspense. I've been looking forward to this launch more than any other car launch in my entire life. And now it's ruined.
But yeah, you want to see the positive. This car is way more aggressive than I was anticipating.
Yes, it really does feel like it's over for every other manufacturer in the sort of like mid-range supercar category because the McLarens, the Lamborghinis, the Ferraris. Yeah. Like it is going to be rough. Like this becomes the most...
When I heard flat eight, I thought, 'Okay, we're going to...' It's got a roof scoop.
I know. I know. We're scooping. I mean, this is the perfect car for an S-tier journalist to just rip around SF.
Yeah. With the roof scoop, for sure. Oh, the scoop. The scoop. Yeah. So I definitely thought it was going to be like Cayman, 911, and then this. And this would feel as far away from a 911 as a Cayman does. And a Cayman does not feel very far away from a 911.
No, it feels way farther.
This feels way farther. Like if a Cayman is a seven and a 911 is a nine, this is like a 15, right? It's like further along and it's much closer to Carrera GT, 918 Spider. Aesthetically, we'll see about the performance. We'll see about how they position it. I know that they're not going to do as much of a restriction on the production numbers, but at least from like an emotional feeling, it's giving me 918 successor. And I know that they might say, 'Oh, it's a different thing. It's going to be a high production or whatever.' It doesn't matter. Emotionally, I'm feeling Carrera GT. I'm feeling 918.
And this just immediately becomes by far the most desirable car in the $500 to a million dollar price range in my opinion. Yes. Like day one compared to Revuelto, compared to SF90, compared to 750LT or whatever McLaren's working with these days.
Although, uh, where's the McLaren 6GT? Remember the McLaren's doing like a manual one. Is that going to be sub a million now?
No, that'll probably be two million.
Oh, it'll be like one of their seven things. Okay.
That's the thing. That's the thing. So many of the cars that have become... we looked at that Alfa Romeo and it was like a million bucks and it's like, okay, a million dollars for an Alfa Romeo that... yeah, there's so many of these situations where they got the market potentially in like a headlock right now. I love to keep...
The one thing is the car is somewhat antisocial, right?
Looking at the amount of rocks that those back tires are going to be throwing out, right? If you want to take this out on a Sunday drive with some buddies, you're going to need some mud flaps. Otherwise, good luck if you're sitting behind this thing. But, uh, overall, can't wait for the event. It's in less than 90 minutes for the actual reveal.
Uh, what a fantastic looking car. Nailed it right at the intersection of the 959 and the GT1.
Anyway, uh, we have our first guest of the show. Reed from Semaphore is here. The San Francisco bureau chief. You know him, you love him. He's got scoops. The new Porsche's got scoops. I swear they made this car for the lone wolf journalist that's scooping around SF.
I don't know how many journalists you guys know, but you know if you buy a supercar, you can no longer be in journalism. You have to leave. That's like really... you can't drive a supercar and be a journalist. That's the trend. Yeah, you need to be driving like a, I don't know, like an Oldsmobile, something broken down or a very old one. Something that you can pull up outside of a technology company office and smoke a cigarette, just act sort of mysterious. Even like a 1998 S-Class that has some weight to it, but it's going to be depreciated. You can pull up and have some authority there.
Can we get... I don't know if you own a car, even owning a car is probably... Yeah. Go ahead. Uh, I don't know if you've been following, but any predictions for the Tesla Roadster later? We've been debating, are there going to be ramps involved? Are there going to be, is it just rocket thrusters? Is it going to be able to kickflip? Things like that.
You guys are in a very different place with cars than I am. I have a Model Y and I'm just like, I just need a bigger car. That's all I care about. I just need more space for the kids and the gear, you know, and all that stuff. Like a Roadster, I'm just not, that's not where my head is at right now. But the Model Y L, the long wheelbase version, gives you an extra 20 inches or something. I was super excited. I want to go when my lease is up. I'm going to for sure go check that out. Yeah. Seems like just make it bigger. That's my ask.
I have one on order and they pushed me back because they can't make them fast enough. I was supposed to get it in November and now it's January. Yeah. I think my theory is that Tesla's like starting this crazy inflection where enough people have had the magical experience of self-driving.
I've heard it. What I'm actually fascinated by is like, it doesn't feel like Elon is actually giving it enough attention. And I think in part that's because he wants to merge Tesla and SpaceX and he owns more of SpaceX, right? So he wants Tesla to be at a lower valuation when that happens. But we've both like put, you know, we've both like, I bought a Tesla a few weeks ago. John has been trying to buy a Tesla. Both of them were delayed by months and months for like basic models. And so I think that inflection is happening. And I don't think you're going to get the like... Every single person I know now is having that like magical moment.
I'll tell you, I have a broken shoulder right now from playing ice hockey and I activated my FSD. It's kind of amazing. You just sit there with your broken shoulder and have the car drive you around. They need to advertise. They should... elite athlete. It's not so much how hard I'm going. It's like how hard my old man body went to the ground when I was brutally tripped in Oakland. But my point is...
Wait, did you throw ice skating rinks? Was it too broken to throw hands after you got taken out?
Way too broken. Way too broken. It's just me sadly skating off the ice thinking this is probably the last time I lace him up. But, you know, we should have been there. We should have been there. We would not have let that go. Yeah, this is the second time FSD has helped me with an old man hockey injury. So I gotta say it's, you know, there's some marketing opportunities there.
No, no, the FSD is overpoweringly good to the point where people who don't like electric cars because they want to hear an engine are like, 'But I still had to get one as a second car for whatever reason.' And then the political thing, people are like, 'Look, the midterms are coming. Elon's going to be like, I don't like them, but I'm doing it anyway.' And I've heard that from a lot of people too. So people are begrudgingly buying them just because everyone said they would never buy a Tesla. And every day at school drop-off, there's a new Tesla, you know, that's post-Elon because it's got the new headlights, you know.
Yeah. I'm getting the bumper sticker. That's, uh, 'I bought this after FSD became so undeniable that I felt it was the safest option for my family and others on the road.' Bumper sticker, right?
Yeah. Yeah. Exactly. Well, my bumper sticker just says 'morally superior.' That's all. That's all.
Okay, that's good. Yes. Yes. Yes. That's ideal. And then they can apply it to whatever the topic du jour of the day is. Like whatever it is, you know, you're prepared. You're prepared.
My other car is the moral high ground.
Yeah. Right there. There you go. That's great. Uh, well, what else are you tracking? Are you, uh, I mean, we were at the start of the show absorbed by the Anthropic numbers. Obviously, historically massive, but still not beating expectations. And so, yeah, what are you tracking today, this week, uh, in general? What's captured your fascination?
I think, okay, so shameless plug, I've got Satya Nadella on my new podcast that'll air next week. So I was in, um, I love the music. Um, so that's going to air next week. Um, but a little teaser for that is, you know, I asked him this question. You'll have to watch it to see his answer, but, you know, Microsoft's doing great. Wall Street loves Microsoft, but like look at the growth of all these companies. Nvidia is now out-earning them per quarter. You've got the hyperscalers. You've got SpaceX AI coming into the fold. Like, is there a world in which Microsoft is still a great business like a few years from now, but they're not in the Mag 7? And, um, I heard a little gasp from the Microsoft comms team when I asked the question, so you'll have to... But I think this is... I'll just say Satya's my...
Yeah, those comms people can get aggressive. Sure somebody didn't chime in and say, 'Are you sure you don't want to talk about something politically controversial instead? Anything but the...'
I'll just say like Satya is my favorite communicator. Like he's like in the Mag 7, he's the one that I think is generally like the most candid and you actually pick up on how he feels about different things in a way that the other CEOs... like he does it in a way that's just... again, I think he's just very candid. I think he's very fair.
He's not stiff. He's not performative.
Yeah. Exactly. Exactly. And it never feels like he's just on some set of talking points and just always looping back to that. He's like actually like, 'Oh, let me answer this question.' So I'm excited.
And if he does have a talking point, it's like he's written a blog post on exactly the question that you're asking and he's got like the perfect, he's thought it all out.
Yeah. And, um, yeah, I mean, this was up at, they've got the CISO summit up there and he's been thinking about security. I think the way he, it was actually really interesting to hear him talking to these CISOs because he was saying he's like, he's been sort of, it's almost like a letdown. He's like let down by these models. He's like, I don't think the alignment is working.
He doesn't think we're going to get to mechanistic interpretability. So you have to put all these guardrails around this stuff.
And that was fascinating to me. It's like a little bit, there's like a sliver of defeat there, but also like we're just going to figure out how to make this work. Like we're going to put secure, you know, deterministic software around it and, you know, we'll use it that way.
Uh, what is your read on Xbox Labs? We talked about it yesterday and I love the decompiled mashup games. I think they're super fun. Uh, as like memes. I don't necessarily actually think that they're blockbuster games, but I thought it was interesting that Xbox was leaning in. But what's your interpretation of that story right now?
Oh, yeah. I totally agree. I was talking with someone, I think someone at Xbox about that, you know, earlier in the week. I think Asha is probably the right CEO for Xbox right now. So I think she's like, you know, she's a little more forward-thinking. She's flexible. Um, and I think they're like, yeah, I mean, if you can't beat them, join them. Like, there's nothing you can do. If some kid wants to like mash up, like decompile a game in their, you know, in their home using Claude, like there's not much you can do about that from an IP perspective. So like I think they have to, the video game industry is going to have to rethink everything. I think it's probably going to accelerate cloud gaming is my guess, like in new ways. And I'm excited by that. And Gen Z is super... It feels like they're in such a good place as effectively a private company because they're owned by Microsoft. So they can kind of like hang out on the sub-balance sheet and completely upend the business model. Whereas if you're an independent video game company, you have a business model that relies on you earning a certain amount from microtransactions and selling a certain number of discs and having this particular business model and you're going to miss your quarterly earnings if you upend everything. But they're kind of in a situation where it's never been a big driver for Microsoft. It's been awesome for culture. It's been really cool to work at Microsoft and be like, 'Oh yeah, we got the Gears of War over there.' Um, but financially there's a lot more flexibility. Do you agree with that?
I totally agree. Yeah. I mean, they still, I mean, the console business is no joke. Like it's like a great business and I think they have that, but they will, I think they realize like they probably have to start disrupting themselves a little and they can do that. I agree. They have the luxury.
Yeah. Yeah, I mean, it feels like it's just good to be in hardware too because like we saw the Mac Mini was not ever built as like an AI device. Like it's very much the opposite thesis from Nvidia, the graphics cards, right, which got used for all sorts of AI stuff and eventually developed server chips. Uh, but the Mac Minis flew off the shelf and you could see people sort of decompiling or hacking Xboxes and being like, yeah, it's just a good bargain, it's a good product, it's reliable and I can run anything on it or I can install an agent on it or I can, you know, actually hack and vibe code and do all these different things. And so they start flying off the shelf and actually becoming like differentiated and they're just going to sell a lot because even if they don't have that huge margin of IP on top, it's still just a great product to be selling.
Yeah. I also love that, I love this concept of just like hacking old hardware. Like you can actually do that now. It's like, why not? I mean, the stuff is so expensive. I'm like going through my drawers now looking for old hard drives because like I don't want to buy a new hard drive. But storage is... Yeah. I mean, it's like why not? Like just try stuff. It's an exciting, exciting time.
Yeah. Uh, how are you thinking about, uh, the open source alliance, like all the things that Nvidia is doing, the different groups there? Uh, OpenAI has been kind of like one foot in, the other, like we don't want to ban it but we also aren't really open sourcing stuff. Meta is like open sourcing stuff but, you know, really has something very close source that's working now. Anthropic still very strong against. But what is the shape of the open source alliance these days?
I think, yeah, first of all, I think the hyperscalers would love to ban it. I mean, I was talking to like our DC people, like the Open Phil-funded groups are having dinners talking about like, you know, open source is going to be banned. I have no idea how that's even possible. I mean, that just seems so draconian and really ultimately just like a fool's errand that...
Yeah, I mean, I think this stuff is going to totally, it is taking off, right? Open source is...
I want to fight you on that because I think it actually might be very easy to ban open source in the sense that like we banned file sharing, like, right? And it's like, yeah, obviously like torrents still exist and you can go and find an offshore website that has a torrent, but you will get a letter from HBO if you download the latest season of True Detective illegally. Uh, they will find your IP. So you got to be using a VPN and then it just becomes cumbersome and you can't have a big data center hosting files like Megaupload. That whole saga came and went and so it needs to be distributed like weights that are, you know, on this like dark web torrent network that's kind of cumbersome. And then the really dangerous stuff is always like a million dollars of compute and I don't think you're going to be able to run anything as an attacker on a Mac Mini that can do anything but bounce off, you know, CrowdStrike plus Mythos plus GPT Cyber plus Palantir, like these organizations are going to be running like everything.
So to be clear, you're not advocating for it.
No, I'm not advocating. I'm just saying that in practice it's not... Yeah. It's like banning serving it is quite different. But and the other thing...
I think that you're... Oh, go ahead. Yeah. Yeah. I was just going to say that the, like the zero-day that Expo found in Chrome, that would have affected a billion people, was like $40,000, right?
Yeah. Yeah. Um, yeah. Okay.
So, okay, but you're assuming that you actually get it banned, right? Like I think you're right. Like if this gets fully banned, like there are technical ways they can just make it a total headache to download open source, they could just like kill that industry. But like is the legal framework for just saying like any kind of open-source AI model is banned going to be so difficult, right? Like what is the theory there?
Yeah. Yeah. Yeah. So like as a practical matter of how things get done in Washington, uh, that is more intractable than the technical question of like how do you find out that a data center is serving an open source model and send them a sternly worded letter, hey, stop serving this because we banned it. But yes, uh, so I agree with you. It's harder to actually ban than enforce a ban that happens potentially. Um, yeah, interesting.
Uh, I've always... What do you think, I mean, from talking to the DC folks, what do you think of this idea of like you can sort of chart the path of technology regulation by just summing up the market caps of various open letters? Like when the open source letter happened, I was like, 'Oh, this is like very rough for the Anthropic position of like let's ban it.' Because I was just looking at it, it's like Nvidia, $5 trillion, and then you have another company that's like, you know, a couple trillion, and I'm just adding up the market caps and I think it was like $25 trillion of market cap was saying like we support this thing. And whether or not DC is coin-operated, it's just a fact of the matter that when you have that much economic weight behind it and that many people's pensions and all the different things, that might dictate the actual outcome.
Oh, yeah. I mean, let's face it, like DC is somewhat coin-operated. Like, it's not, you know, it's oversimplifying it and I'm out here in the Bay Area. Easy for me to say. I don't have to go to dinner with these people every night. But I mean, yeah, I think you totally can. But I also think it's like, there's not some law that's being violated by this already. Like you would have to create a whole new idea, a whole new legal framework. It makes it really difficult. And then like the other thing I think with like, name a tech regulation that was actually good or didn't have some major blowback? Like even Section 230, which is almost like unregulating rather than regulating, is like hugely debated today. The DMCA that governed all the downloads and torrenting and everything, that's a huge mess. Like that just created headaches for years, still is creating all these issues with doing like security research, right? Like you have to be very careful about how you research bugs in software because of that. So like I don't know, it just to me it doesn't make much sense. It makes a lot more sense like, it's like you have the money behind it and also just like the logic behind it. Like it makes logical sense, like let's have open weights, let's have open-source harnesses and sandboxes. Like I had a little scoop today on Perplexity open-sourcing their sandboxing technology. I mean, like let's crowdsource like figuring out how to keep this stuff in silos and make sure it's safe. Like that makes a lot more sense than like let's have two labs companies in Silicon Valley like figuring this all out for us. Like that never works.
Yeah. Yeah. Uh, I'm preaching to the choir. I realize that.
No, no, no, no. It's interesting. I'm thinking about like, AI is maybe like the least frustrating it will ever be. I'm just imagining like, do we get the AI utopia or do we get the doom scenario where like the median case is like probably like we get cookie popups and we get some like annoying thing that happens that like is constant and it's just like the cookie popups development and like the way that like that's how the internet played out is just so depressing. Every time I have to click that button, I'm just like, how was this the outcome? But I don't know. That's beside the point.
That was a consumer thing, right? And like I think there's actually like my theory right now is like the reason there isn't more consumer stuff happening in AI is that there's just not enough compute. Like we don't have, you just don't have enough data centers to run like the frontier models for consumer applications. And I think open weights is probably the way that happens like first before you get to, I don't know, like the Web 2.0 of AI where it's like all in the cloud. You're probably going to have some sort of like the Netscape of AI first, which is like it runs on your computer and it's, you know, it's not wide, it's not like everyone uses it, but it's sort of like early adopters get into it and we're like not even there yet. I mean, it's...
