A (3:43)

In fact, they are quickly. They did. No, really.

B (3:46)

They were accused of working together. Oh, and, and that would have been illegal and they would have been colluding. They are bitter rivals. Kind of going back 80 years, they've never worked together. They do not work together. They in fact are fighting over this market completely. And so what actually wound up happening was that the first, you know, the first drug that kicked us all off was Novo Nordisk. They patented Semaglutide, which was marketed as Ozempic and then Wagovi.

A (4:15)

And to my point earlier, referencing the Canadian market, this happened around five months ago. Novo Nordisk actually lost Canadian patent protection on Ozempic and Wegovy in Canada because a few years ago they didn't pay the very nominal maintenance fee on their patent.

B (4:33)

It's crazy.

A (4:34)

Which will ultimately cost them billions.

B (4:36)

Well, don't do that. Instead, get on ramp. Time is money save. Both easy use corporate cards, bill payments, accounting, and a whole lot more all in one place. So back to GLP1s. No one can patent a biological mechanism. You can only patent a particular chemical, a particular structure. You can't just, you can't patent the idea of inhibiting, mimicking GLP1, which is what all these drugs do. They mimic the GLP1 hormone and that's what's important. So Novo patented semaglutide. Lilly had tirzepatide already in the works. Again, they were working on, on diabetes, not on weight loss, something else. But, but it was, it was already kind of predicted to be a big market. So they are both. They both have products. And then as soon as. As soon as Novo is able to pivot to weight loss, Lilly's able to run the same playbook very, very quickly. And accidental discoveries and quick rebrandings are nothing new in pharma. Viagra was initially created by Pfizer as a heart medication.

A (5:39)

Yeah.

B (5:39)

They gave it to both men and women, I believe, and. But male subjects obviously had a particular side effect, and that unlocked an entirely new market. And then that discovery came out of left field completely. So they had a true. Pfizer basically had a true monopoly for five years before Cialis came to the market. This is obviously different because both companies were working on the underlying structure. Adderall also had a really odd path. Amphetamines have been around for over 100 years. They were used by soldiers in World War II. So obviously you can't patent. 1970s, they were sold as weight loss drugs. And so there was a drug.

A (6:17)

Is it an appetite suppressant?

B (6:19)

Yes, yes. So it was called Obitrol. And Ovitrol was sold as a weight loss pill. In the 90s, a company bought Ovitrol, renamed it Adderall Meth.

A (6:30)

Yeah, it's a great weight loss drug. Try it today.

B (6:33)

It's actually so crazy, I imagine that it didn't really take off as a weight loss drug because it had crazy side effects. I don't actually know. I would imagine it was effective, but probably just very risky to use. But in the 90s, a company purchased Ovitrol, rebranded it Adderall, and then marketed it for adhd. And they didn't have the patent on the chemical. Remember, you can't patent the chemical amphetamine because it's been around for so long. So what they were able to do is develop a brand. And Adderall is a product that the brand name, somehow it just completely broke through in a massive way. And it's just synonymous with Kleenex adhd. Yes, yes. It's actually broken through in such a way that it's synonymous with adhd. It's synonymous with, like, study medication, even. And so it's been wildly successful. They actually did wind up getting a patent later on for Adderall xr. So Adderall XR is a bunch of beads in a capsule and they were able to patent the delivery.

A (7:31)

Yeah, exactly.

B (7:32)

Form factor. Exactly. And so even when you don't have the patent on the underlying tech, you can still patent things around it. Patent the biosimilars There's a bunch of different things you can do.

A (7:44)

And so shame that the Gila Monster never is not really getting any upside, not getting the bag, any of these weight loss drugs because that really is the origin. Right.

B (7:55)

They really should do that. They should do like 1% for the Gila monster and just 1% off the top. They should donate it to like Gila Monster.

A (8:02)

Maybe a foundation of sorts.

B (8:04)

Yeah, yeah, a foundation.

A (8:06)

A non profit.

B (8:06)

A non profit that just makes the Gila Monster live in absolute luxury.

A (8:10)

That's right.

B (8:11)

As a thank you. So Eli Lilly will be fighting this endless battle with every other.

A (8:16)

And to be clear, for those that don't know, it was Novo Nordisk that figured out what was happening in the Gila Monster. Which Gila monster?

B (8:26)

I don't remember.

A (8:27)

Yeah, it is, it's the Gila Monster because there's two.

B (8:29)

Yeah, there's the Komodo dragon too. And I often get those two creatures confused. But yeah, it's the saliva in the, in the Gila Monster, I believe that, that, that uses the same pathway, the GLP1 hormone. So, so basically over the next couple of years we're going to see this interesting dynamic where Eli Lilly, Novo and basically every pharmaceutical company like wants they have to get it on the action because it's such a big market. Obesity is a massive disease. It really is that that trillion dollar opportunity. It's been that for a long time. There's you know, hundreds of millions of obese people. Everyone who has that, you know, doesn't want that. There's. The market is so, so broad. And so what's interesting is that the patents haven't resulted in a perfect monopoly where it's not this winner take all situation where Novo is just going to run away with. Eli Lilly is just going to run away with it. What's actually happening is that they, it's a duopoly right now, but demand is so high that they still have high margins. So they're in like a price war. But the price war, and that's even.

A (9:39)

With the competitive dynamics of the compounders, companies like him, which are creating a ton of different GLP1 products, you know, just in their own.

B (9:48)

Yeah.

A (9:49)

In their own facilities.

B (9:51)

And so the price has fallen 3x. And normally, you know, if the price of your good is falling 3x, you're charging 3. If you're, if you're selling a $50,000 car and then a couple years later you're selling it for 15k, like that's.

A (10:02)

Not going to cost something. Like in the same range to produce. And so you're just getting massive margin compression but the growth is so insane that it hasn't really mattered.

B (10:14)

Yeah, exactly. So the market is growing so, so much margins have fallen from 80% to like 60% but the mass adoption just completely offsets that. And so they're making way, way more dollars in total. And so the next race now you know, Eli Lilly's a trillion dollar company. What do they have to do next? They have to go a pill. They're, they're going to try and make these weight loss products in pill form. And there's a whole new battle duking it out between Eli Lilly and Novo Nordisk again and some other players. But it's, it's a, it's a fascinating time and there's a lot of, they have their own like little barnacle economy. There's a lot of analogies to the, to the AI race, trillion dollar companies and high margins and all this. But at the end of the day, what matters in the GLP1 market is just the actual supply. Like you have to go and build a massive facility and once you get the facility up, people will buy it because just the demand is so high that if you can make it, you will be able to sell it and you'll be able to sell it at a good margin. Anyway, let me tell you about Restream 1 livestream 30 plus destinations. If you want to multi stream go to restream.com There was some interesting news in the SEO world. Adobe is buying SEMrush for $1.9 billion. Should we do size gong for Semrung?

A (11:34)

I think so. I think, I think so. Yeah. Fantastic outcome. It was notable to me because obviously we've been tracking the generative engine optimization space and there wasn't really a great comp for a lot of those. I remember we were talking about it and it's not like you hear a lot of SEO companies getting acquired in that range. And so anyways, the geo market is still highly, highly, highly competitive. Obviously we work with Profound, which has been the clear leader in the market to date. But I would expect this to drive even more companies to come in and compete and ask any VC they're getting pitched companies in this category multiple times a week still. So people are just piling in and that's interesting.

B (12:28)

I wouldn't think of Semrush as a geo.

A (12:31)

No, it's not. It's an SEO. It was the SEO leader. That being said, I would expect them to also want to play In Geo, it's not like they're going to ignore.

B (12:43)

I would almost be thinking about Semrush having their own build versus buy conversation. It's sort of unclear what it would take them, how much data they have, how they can jump into the geo market. Obviously a lot of companies have been able to get something up and running.

A (13:00)

Very, very quickly and they do already have a tool, it's called Semrush, a visibility. But again, there's so much demand for this kind of new product because every single company is spending a lot of time thinking about how they're appearing in different LLM queries. So still up for grabs.

B (13:21)

Reportedly SEMrush did $250 million in revenue back in 2022, probably grown a little bit since then. Founded in 2008, a true overnight success at almost 20 years to get on to the next thing with Adobe. Speaking of Adobe, I need to push their tools to the limit at this point. They were so behind in terms of generative AI and their models. They were training their own models. But eventually I saw a demo where nanobanana was in Photoshop and I feel like that is really, potentially a really, really big unlimited unlock just for workflow. We've seen so many different examples of this in AI where the models are incredible, but having the harness and having the wrapper, whether that's a product that was introduced five years ago, 10 years ago, 50 years ago, like they're all getting better.

A (14:23)

Adobe works with fall so they can integrate pretty much any model in the world and inference it. So I expect them to. It is funny that whatever it was two years ago, every single player felt like they needed their own models and now I think the narrative is really flipped.

B (14:43)

Well, speaking of fall, build and deploy AI video and image models with fall trusted by millions to power generated media at scale. And also Gemini 3 Pro, Google's most intelligent model yet. State of the art reasoning, next level vibe coding, Deep moly, deep multimodal understanding. We've been having a ton of fun with Gemini 3 Pro and Nanobanana Pro. We have generated an image of the Aura farm the world trying to understand a barnyard. A barnyard sort of an AI market map, but through the lens of a barnyard because that's really what is going on in AI these days. Let's see. I think we need to zoom this out just a little bit to get it to line up, but we can show and we can take you on a tour of our farm based market map to explain what's happening in the state of AI in November of 2025. Tyler, do you want to take us through it?

C (15:44)

Sure. Yeah. I mean, so we like to use a lot of animal idioms on this show. A lot of these you might recognize. But we've kind of expanded out to try to cover all the major players and, you know, so. Yes. So this was made with nano. Nanopro. Very extremely good model. But you will notice as you get more and more complex, it gets a little bit sloppy. There's some slop in some places.

A (16:09)

Okay.

C (16:09)

But broadly, I think it did a very good job. Had a lot of fun making this.

A (16:12)

Yeah. Before nanobananapro was released, we would have needed to hire somebody that maybe illustrates children's books. And if they were an expert, maybe they could have whipped something together in a day. This took you far longer than that, prompting over and over and over, but I'm very pleased with the outcome.

C (16:33)

Yeah. Okay. So I guess let's just kind of go through each of the animals. Maybe let's just start with some of the more obvious ones, the ones we've talked about a lot on the show. So let's start over here with the piggies. So here we see the pigs at the slop trough.

A (16:50)

Who's there?

C (16:51)

And so this is probably. This is just labeled as meta, but you can imagine a lot of people here. Right. You could see Sora here.

B (16:57)

Sure.

C (16:58)

Bill Peebles. Bill Peebles could be, you know, eating some slop.

B (17:03)

Yes. Meta vibes.

C (17:04)

Yep.

B (17:05)

That makes sense.

C (17:06)

I mean, just like broadly, AI in general. Some people think basically all AI is slop.

B (17:11)

They do, they do. There's been a lot of criticism.

C (17:14)

Yeah. But I think that one's fairly clear.

B (17:16)

You don't want to be at the. You don't want to be the pig of the slop trough. Although sometimes, you know, maybe the profitability of being a slop farmer is underrated. Obviously, America has a lot of pigs that live at the trough. They eat from the trough. They. Eventually, the pigs go to slaughter. They become bacon. They are sold into the economy. You can make a good living as a farmer who maintains a.

A (17:45)

And then there's the Jeremy Geffon take, which is we hate slop now because we know that in a few years, it will no longer be slop out. Don't count out the piggies yet.

C (17:54)

This is the sloppiest slop will ever be.

A (17:56)

That's true.

B (17:57)

That's true. Okay, who else is on this market?

C (17:59)

So then let's move up. Let's see, we have the henhouse here. Yes. And who's in the hen house is the fox.

B (18:05)

The fox.

C (18:05)

Right. And so this you can say is kind of Oracle in St. Mautland, right? Oracle is the hen house.

B (18:11)

Yes.

C (18:11)

They let the fox in. They're letting the fox in a little bit.

B (18:14)

Potentially.

C (18:14)

Yeah. We'll see what ends up happening.

B (18:17)

But again, now that I'm looking at this image, the fox isn't actually in the hen house yet, but he's kind of circling it. He's circling the hen house. And there's a question of how far into the hen house has the fox gone?

A (18:30)

And I think it's notable that the chickens don't seem too disturbed yet.

B (18:33)

Yes, yes. Yeah, yeah, they're getting along because the fox could just be browsing, could just be stopping by. Could be friendly, friendly fox.

A (18:40)

Friendly fox, yeah.

B (18:42)

But of course it is. If you have a lot of cash flow as a big company, if you have the ability to borrow billions or hundreds of billions in the debt market, you have to protect your hen house because foxes might come by and might want to use you to co sign a loan.

A (18:59)

Yep.

C (19:00)

Okay. So then we see the cash cow. The cash cow, this is Nvidia.

A (19:04)

The only profits in AI for a long time.

B (19:07)

I don't know if anyone has really unseated them. They're certainly the most profitable by far. I think for a long time they were making more than 100% of the profits across everything. Everyone else was losing money.

C (19:20)

Yeah. Now you're starting to see maybe some cracks with TPUs, but even then, I mean, they're still.

B (19:25)

It's still a cash cow. We just saw. They just. Yeah, they just beat earnings. I mean, we have, we have an article here from the journal. Nvidia results fail to quell AI worries. Not enough people are milking the cash cow.

A (19:36)

I think I gave you the perfect.

B (19:37)

Or maybe, maybe the casualties.

A (19:38)

I gave you the perfect quarter.

B (19:39)

I know. And you still sold off 60 billion or something in revenue. Yeah. Maybe the cash cow has been overly milked and is out of milk at.

C (19:48)

This point, but I think it's still producing.

B (19:50)

I think it does seem like it's still producing. That looks like a healthy cow to me. That looks like the crown on the cow in particular. Definitely helps out there.

A (19:59)

All right, moving over, what do we have going on here?

C (20:01)

So here we have the bull in the China shop.

B (20:03)

The bull.

C (20:03)

And that's Elon. I think there's a couple ways you could read into this. So, Elon, one is just kind of on the. Maybe the data center infrastructure side. Where he's kind of the bull. He's. Imagine that you're a contractor and you're building data centers for these companies and then you see Elon come in and he does it in eight months. What took you in a year and a half?

B (20:21)

Yep.

C (20:22)

He's kind of messing up your world, right?

B (20:23)

Totally, totally.

C (20:24)

He's like, oh, man, this bull, he's messing up my business. I need to be working way harder, way faster.

B (20:28)

Because he's also shipping features so fast. Before people could even have a discussion over adult content. It was like, boom, Annie, here it is, Valentine. So fast.

C (20:39)

This is the other way you could say is the bull. Because there's kind of this maybe unspoken, unspoken rule between the model companies where it's like, okay, we don't want to go too hard into the, the companion space. There were companies doing that, but they weren't. It wasn't the opening eyes, it wasn't the anthropics. Wasn't that the Geminis?

B (20:55)

Yeah.

C (20:55)

And then now you see, now you see a frontier model actually moving into their.

B (20:59)

Also, you know, Grok is of course the maximally truth seeking LLM. Just yesterday I was asking Grok who is the strongest CEO in tech and it told me that Elon Musk is the strongest by far. And it actually compared him to a bull. It said he was stronger than a bull and that he could beat a bull in a fight if he went head to head.

A (21:17)

I believe it.

B (21:18)

With his bare hands.

C (21:19)

Absolutely right.

B (21:20)

Yes, absolutely right. What else is going on here?

A (21:25)

I bet Claude would agree. If you really pressed Claude on it, they'd say, yeah, you're absolutely right.

B (21:30)

You're absolutely right. Who else is going?

C (21:32)

So let's move down here. We have the lipstick on a pig and I think you could say, is Apple intelligence?

D (21:40)

Yes.

C (21:41)

Right. I mean, although I feel like this is actually almost flipped a little bit. Right. Because the pig is usually like the ugly thing and then you add the makeup to make it look nicer.

A (21:48)

Yeah, Well, I think that's what they did here, Right. They took a bad model and they dressed it up and they said it was great and they said you should buy a new iPhone because of it. And the model was bad. And there's no amount of marketing that is like change public perception once the product hit the market.

B (22:05)

You could also run the pig is the privacy focused iPhone ecosystem that is pretty difficult to just run in and throw AI on top of. Apple has been building a brand around. We don't access your data, we don't store your Data. We would never train on your data. And that put them in an awkward position where they couldn't just snap their fingers and deliver a great AI experience. Whereas Google has been, hey, you know, we'll give you Gmail for free. But we'll also, you know, probably train on wherever you go on the web and try and understand and route ads to you appropriately. Meadow is in a similar position. And so those two companies were a little bit more equipped to just, on day one, go and go and deliver AI features. Apple had to put lipstick on their pig, which is their privacy.

C (22:54)

It was more of like an afterthought. It was more of a kind of shallow integration and then hopefully, we'll see with marketing.

B (23:00)

The marketing for Apple Intelligence felt like they were. It felt like a lot of makeup on top of something. It was not a lot of. It was not a foundational rewrite of iOS in any way.

C (23:09)

Not a lot of substance.

B (23:10)

It didn't feel like an entirely new thing. It just felt like little lipstick all over the place.

C (23:14)

Yeah. And then before we go over there, let's go up here. We have the rooster.

B (23:19)

Yeah, that's Jordy's cockatoodle do. This is Jordy calling the top every morning.

C (23:23)

Every morning.

B (23:23)

Every morning he wakes up and comes on the show. 11:00am Sharpie calls the top. He says, this is. It's actually over now. It's so over. It's never been over.

A (23:31)

He call the top every day. Eventually you'll be.

B (23:32)

Eventually you'll be correct. That's true. Eventually you'll be correct.

C (23:36)

Okay, what else do we have? So then we have the dark horse.

B (23:38)

The dark horse.

C (23:39)

And this is ssi.

B (23:39)

Okay.

C (23:41)

We're still yet to see anything really.

A (23:43)

From ssi Elliot just charging around in the background?

B (23:46)

Yes.

A (23:46)

Yeah.

C (23:47)

Kind of this figure still hanging out there. It's kind of. I mean, like. So the big kind of departures from OpenAI, it was Miramorati and Ilya, you could say.

B (23:56)

Yes.

C (23:56)

And Miriam Roddy's company, Think Machines. I mean, she's raising up rounds. They have blog posts. Yeah, they do actually have a product. They do have a product.

B (24:03)

Let's hear it for. Why did you tell me earlier? Why'd you bury the lead? They have blog posts.

C (24:14)

They're good blog posts. I like them.

B (24:15)

So in a couple years, we could be seeing vibreals, Is that what you're saying? We could be seeing lots of videos.

C (24:19)

So they do have a product. Right? It's like RL fine tuning.

A (24:21)

Watch your head. There's a hook.

B (24:23)

The horse is really, really about to hit you. Okay.

C (24:27)

But I mean, they're not really dark horse, like, you can. They're a bright horse. I don't know if that's a phrase, but, like, you can see what they're doing.

B (24:34)

A Clydesdale or a beautiful snail.

A (24:36)

They're a unicorn. 50 times over.

B (24:38)

50 times over.

C (24:39)

Yeah.

B (24:39)

We didn't put a unicorn on here. That's an animal. Maybe for the next.

C (24:43)

After a certain point, I couldn't add any more. It would get too sloppy.

B (24:48)

Okay, what else do we have? What else is going on in the Aura farm?

C (24:50)

So here we see another horse.

B (24:52)

Yes.

C (24:52)

And there's two ways you could look into this horse. One is this phrase, look a gift horse in the mouth.

B (24:57)

Yes.

C (24:58)

Right. Do you want to explain this, John?

B (24:59)

Yes. So do you know this phrase? You don't know this phrase? Looking at gift horse in the mouth. Okay. So looking a gift horse in the mouth is if someone gives you just.

A (25:07)

Like a term that was thrown around in the 80s.

B (25:09)

Yeah, yeah. This is back in my day. Back in my day. So obviously, just giving someone a horse. A horse is a valuable asset today, but also hundreds of. Of years ago. It's always been a valuable tool on the. On the farm. And the way that you assess the quality of a horse, one way to assess whether it's been taken care of, whether it's healthy, is to look in its mouth.

A (25:33)

Always look a gift horse in the mouth.

B (25:35)

No, never look to get. Never look a gift horse.

A (25:37)

Oh, because it would be offensive.

B (25:38)

It's offensive if I show up after.

A (25:40)

You get the gift later. Imagine after they leave.

B (25:43)

Imagine if I got you a GT3RS and you're like, popping the hood and.

A (25:46)

You'Re like, oh, you track this?

B (25:48)

Yeah, yeah, yeah. You're feeling the tire. Oh, I don't know about this. It's not that good.

A (25:53)

Oh, it only has 2,000 miles on it. But they're around the Nurburgring.

B (25:57)

Exactly, exactly. That would be looking at GT3RS in the tread. You don't look in the tread because you know. Yes. If I gave you a horse, just be happy that I gave you a horse. You know, don't look. Don't look in the mouth. Don't assess it. And so what's happening with that?

C (26:11)

Yeah, it's like you're not expressing gratitude. So I think that this. These are just the public market investors. Right. We've been given that gift of why.

B (26:19)

Are they selling their stocks right now?

C (26:20)

Why are they selling giving?

B (26:22)

They need to be buying this beautiful.

C (26:24)

Gift and they Just don't want it. Yeah, they're selling their videos down. Yeah, they beat earnings. Guys.

B (26:29)

Why are you not buying the dip? They need to be buying more. They need to be levering up. They need to be going further into the public markets. Every dollar.

A (26:36)

They have not financial advice at all.

B (26:39)

But instead they are looking a gift horse in the mouth saying, oh, can OpenAI afford all the different $1.4 trillion.

A (26:46)

Is Andy Jassy the guy for the.

B (26:47)

Job is the Google. You know, is Google set up to actually take advantage of AI? They're digging in a little bit too much. They should just be happy that they've been given AI the next mega trend.

C (26:58)

Exactly. So also the horse could be a workhorse.

B (27:01)

A workhorse.

A (27:02)

Right.

C (27:02)

And so I think you could say this is Amazon.

B (27:04)

Yeah, it's kind of cooking hard to.

C (27:06)

Define Amazon's AI strategy. Part of it is they're building data centers for anthropic. But they're definitely not getting overly ambitious. They're not like, you know, getting over their skis, but they're just doing the hard work. They're building the data centers. They're serving. Yeah, serving models.

B (27:21)

Maybe not fast enough. They might not be building the data centers fast enough. They might not be super aggressive. They're not show jumping. They're just dragging the plow. But they are consistently dragging the plow. All reliable Amazons cooking along. They're staying out of the slop trough. They're not doing a deal. Amazon has not done a deal yet with OpenAI for the agentic commerce thing. Maybe that changes. But for now they're just plodding along. Who else we got here?

C (27:49)

Let's go to the black Sheep.

B (27:51)

The black Sheep.

C (27:52)

So black sheep. There's also a lot of people that could fit in this.

B (27:54)

Yes, of course.

C (27:55)

But I think maybe Carpathy is one of them.

B (27:57)

Yeah, right. Andre Karpathy.

C (27:58)

Some of these contrarian tanks. Contrarian takes about maybe just a decade of agents. Not being.

B (28:03)

To be clear, we're not calling him a sheep. He's not.

C (28:06)

It should be more of a black wolf.

B (28:07)

Black wolf. But the idea is that he's standing out. A lot of people in his class of ultra respected technologists, some of the most experienced. He's really been at Tesla. He's seen the AI wave. He's been working in it for decades. He's really worked on this.

A (28:25)

He's printed all of it.

B (28:26)

Yeah. He's done so much.

A (28:27)

Not sitting with the rest of them saying he broke, but he broke rank.

B (28:31)

Even though he's worked at OpenAI, even though he's worked at Tesla. He broke rank. He went on the Dwarkesh Patel show and said, you know what? I think we're more in a decade of agents, I think AGI might not be right around the corner. He took a contrarian stance at a very controversial time. But it was not. Yeah.

A (28:50)

And he popped the bubble with it.

B (28:51)

He basically popped the bubble with it. We'll see the bubble.

A (28:54)

Of course, the elephant in the room. The elephant in the room was in my view really more of the 1.4 trillion. Right. That was the big question. That was what everybody wanted to talk about or at least understand better. Frad asked the question and it's been downhill ever since.

B (29:12)

Yes. And it is. Yeah. The elephant in the room is in every conversation, every financing round at this point. Exactly. Is this, is this deal predicated on a continuation of exponential growth in investment in everything? Like how much more, how risky is this relative to the amount of froth in the market?

A (29:34)

Never ask a woman her age, a man, his salary, or a lab founder how they're going to spend 1.4 trillion.

B (29:44)

Yeah. Or how they're going to pay for it.

A (29:46)

Yeah.

B (29:47)

What else we got on here?

C (29:48)

So then in the top right, we have the bird's eye view.

B (29:50)

Yes.

A (29:51)

So situational awareness.

C (29:52)

Yeah, this is situational awareness. This is Leopold.

B (29:54)

It's the perfect name.

C (29:55)

Right.

B (29:55)

He has the situational awareness.

C (29:56)

He's seeing everything. He's kind of the master of the board.

B (29:59)

Yes.

C (29:59)

He's taking his bets and he's been doing pretty well off them.

B (30:02)

Yes, yes, yes.

C (30:04)

Yeah, that one's pretty self explanatory.

A (30:05)

Yeah.

C (30:05)

Okay, so then we go down, we have lion's share.

B (30:08)

The lion's share.

C (30:09)

This is Satya. Satya taken great position.

A (30:12)

I'll take.

B (30:12)

I don't know why.

A (30:14)

I'll take. I'll take a third of the company and you can give me 250 billion and I'll take 20% off the top.

C (30:22)

And I'll take all the IP as well.

B (30:24)

Yeah, everything.

C (30:25)

I'll also take your chip and I'll make it my chip.

A (30:28)

Your chip is my chip.

B (30:29)

The entire contract is not your chip.

C (30:32)

As mech.

B (30:33)

The entire OpenAI Microsoft deal is basically could just be summarized in one line. I get the lion's share. I get the lion's share. And that's what he did. It's fantastic.

C (30:41)

Incredible. Incredible dealmaker. So let's go down here and we.

A (30:44)

See and we can zoom out a little bit.

C (30:47)

Yep. What do we have? Monkey business. These are the podcasters.

B (30:50)

That's us. We're having fun. We got props, we got all sorts of stuff. Sound effects, we need some monkey sound effects on there. We have a bunch of animal themed sound effects we need to have more fun with anyway.

C (31:05)

And then maybe let's go into the pond. We have the sitting duck.

B (31:08)

Yes, the sitting duck is incredibly cute. So Reddit. Why was Reddit a sitting duck?

C (31:14)

Yeah, so I think they brought it.

A (31:16)

It's because whether or not you have a permit, you're getting like, the hunters are going to get you.

B (31:23)

I think Alexis Ohanian sort of laid this out, saying that when Reddit did the deal to give the data to OpenAI, they didn't realize how valuable that data was going to be. And that relationship has grown, grown, grown. And so they were kind of the sitting duck just sitting there. They didn't really kind of got caught off guard. Maybe could have trained their own model on their data internally. Maybe could have, you know, maybe had something that was more, more of a valuable resource where they, they would have had more leverage if they'd waited a little bit.

A (31:55)

But it is crazy that CoreWeave and Reddit are sitting at roughly the same market cap now.

B (32:00)

Very interesting. Yeah. Which one is more valuable over the next couple years? I don't know.

C (32:06)

Yeah, I mean, I feel like Reddit has mostly been, you know, they've kind of already gone out to the slaughter in a sense. Like every model company has trained.

B (32:14)

You think so? But look at the market cap. It just keeps going up. Like Reddit's been. I mean, Reddit was sold for $5 million or something when it launched.

A (32:23)

Like, it's been, it was 10, it was 10, it was ten to Conde Nast.

B (32:27)

Ten to Conde Nast. That's a crazy, crazy low valuation. And then eventually spun out. And then for a long time it was in the hundreds of millions and no one was thinking this is a tens of billions. For a long time, like people, the narrative was not, oh, yeah, this will be orders of magnitude more than snap. No way, no way. And then it just sort of finally came together.

C (32:45)

So then who's the headless chicken? Yeah, let's go down a little bit. And we have the headless chicken. This is perplexity, perplexity. So I think perplexity. You know, they have a lot of different strategies. Some of them kind of seem opposed to each other. Right. Maybe they're doing a browser, maybe they're doing a Bloomberg terminal.

A (33:02)

Maybe they're rebuilding Yahoo Finance. Maybe they're building their Own the browser, to their credit. But I do see more positive reviews of their browser than any of the.

B (33:14)

Other AI browsers from independent folks or from.

A (33:17)

Yeah, independent, independent people.

B (33:19)

We've had a lot of investors come on the show and say, like, I love the browser, but it's hard to take that seriously.

A (33:24)

Every investor that's invested in a browser says their browser is the best browser. I don't listen to what they say, but I've just repeatedly seen people saying, like, the Perplexity browser is great. So whether or not that means a win. But the headless chicken thing, going over and giving $400 million to Snapchat kind of, you know, I don't know. I think a lot of people. Good for Snapchat.

B (33:50)

Maybe there's just a lot of stuff going on. It's the Snapchat deal, the trying to buy TikTok, trying to buy Chrome, launching a venture fund, Bloomberg Terminal. It's like. Like, are you competing with all of these? Really? You're gonna be TikTok and Chrome and Bloomberg and the next thing and see the next great company? It is just a lot. It's a lot to process.

C (34:13)

Yeah. Like, headless chicken is kind of like a fun thing to watch. And then people, you know, love to talk about perplexity. They love to say, like, oh, we're all short perplexity. That's true.

B (34:21)

Yeah. It's become consensus. Yeah. But let them cook. We'll see what happens.

C (34:25)

Yeah. So then if we go down a little bit, we have the snake in the grass.

A (34:28)

Have you guys ever witnessed headless chicken?

B (34:32)

No, I haven't.

A (34:34)

Growing up in the country, I have.

B (34:35)

You grew up in the country?

A (34:37)

What are you talking about?

B (34:38)

I thought you grew up with the East Bay rationalists.

A (34:40)

I was born in Berkeley, but I grew up in wine country.

B (34:42)

Oh, okay.

A (34:43)

And I have, fortunately, unfortunately, watched my father take one of the chickens. You know, every now and again, a chicken needs to move on to the next chapter.

B (34:57)

And as a child, you sound like you're firing the chicken. Very amicable split. I mean, we're wishing the chick in the fries.

A (35:06)

I mean, yes, it happens.

B (35:08)

It's the next chapter.

A (35:09)

I will say that it is a real thing. They run for, you know, at least 20 seconds or so, and I'll never forget it.

B (35:21)

Well, before we move on, let me tell you about cognition, the makers of Devon, the AI software engineer. Crush your backlog with your personal AI engineering team. Okay, let's continue.

C (35:30)

Okay, so next we have the snake in the grass.

B (35:32)

Snake in the grass.

C (35:33)

This is the Chinese open source models.

B (35:34)

Right.

C (35:34)

These are kind of lurking. I still don't really see that much coverage of Chinese open source models out there.

A (35:41)

Nobody wants to talk about it.

B (35:42)

Well, no one has bags so they can't pump them.

A (35:44)

Yeah, it's almost an, it's an elephant in the room to some degree.

B (35:49)

So this is where I sort of disagree with you because Snake in the grass feels like it's going to attack elephant in the room means we got to address it. It's like it's totally possible that the entire open source LLM ecosystem is just like, yeah, there's like a $10 billion business there. It's sort of like you know, a stalking horse. Another, another animal based analogy. But it's sort of a stocking horse for like hey Gemini and, and anthropic. Like if you guys don't lower your prices we will go to the open source option. And that's sort of what Linux, how Linux works. It's like that you can kind of bid the closed source models against the open source models. But for all like the doomer takes around like oh, Chinese open source, like one day they're just gonna snap their fingers and all the Manchurian candidates are gonna activate in America.

A (36:39)

I think everyone's gonna be like, I.

B (36:40)

Didn'T realize I was using quad and now I believe that nothing happened on Tian Square.

A (36:45)

I don't believe that's more of a threat to the, it's a very real threat to the business models of some of the closed source models.

B (36:54)

Maybe. I don't know.

A (36:58)

Again, it's certainly a pressure. Kimmy is hot on everyone's tails.

C (37:05)

I mean Brian Chesky he was talking about.

A (37:07)

Yeah, Brian Chesky. The other thing is a soft power thing. Ask a Chinese open source model about Tiananmen Square.

B (37:14)

I just don't think that's going to be that big of a deal. Like I don't, I don't, I don't, I don't see. Those are two very different things. Like one is like Brian Chasky inside of Open, inside of Airbnb becomes a.

A (37:25)

Big deal if they proliferate all over the world. And it's. And, and I'm not saying it's the end of the world.

B (37:31)

No, no, no. What I'm saying is that with TikTok, if you're a business owner like Airbnb and you're like Brian Chesky at Airbnb and you're deciding between Kimi, you know, some open source model, like deeper in the stack, like you're not you're not surfacing something that has, like, cultural relevancy or like, or like some sort of, like, you know, where American values come through. Like, you're using it for, like, fraud detection, or you're using it to, like, transform every Airbnb listing into, like, grammar check. Every listing that comes through, like, spell check it all. You're doing something that's very commoditized. And it's not. It's not this weird thing where you're going to inject the, the, like, the. The values of the Chinese people. I just see that as two separate things now. Now, if you had a whole bunch.

A (38:17)

Of people just think that LLM be the new, are effectively globally going to be history books. And if you have some control.

B (38:24)

So that's true as the application layer.

A (38:26)

At the application.

B (38:26)

I think at the application layer that matters. So I would be maybe more worried if it was like, oh, it turns out that OpenAI is going to just use Deep Seq now under the hood.

A (38:36)

Sure. I'm even talking globally. A bunch of different application layer companies use these models because they're cheap and good, and then they have a very specific point of view that's not necessarily aligned to the West.

B (38:50)

I just think that point of view is irrelevant in many, many automation use cases, for sure.

A (38:58)

I'm not worried about Airbnb using Chinese models for, for like, business process automation.

B (39:04)

Yeah.

C (39:04)

I also think you can. I think you can probably kind of fine tune a lot of that stuff out totally. Because, like, the, you know, back to Perplexity.

B (39:12)

They did it. Did they do Deepsea 1776 or something like that?

C (39:15)

Yeah, I also did that.

A (39:16)

You did that?

C (39:16)

I did that actually before then.

B (39:18)

So it's whatever.

C (39:18)

But it's whatever.

B (39:20)

Okay, now we see it's personal. That's why you put perplexity there, because you got an ax to grind. I did it before they did, but.

C (39:28)

Like, the Jianan Square stuff is probably just a fine tune that they made it more Chinese.

B (39:32)

Yeah, yeah.

C (39:33)

So you can just undo that.

B (39:34)

Yeah, exactly.

A (39:34)

I agree.

C (39:35)

One would assume anyway, but yes.

B (39:37)

Okay, so we agree it's a snake in the grass, but unclear how poisonous the snake is. Unclear the size of that snake.

C (39:43)

Yeah.

B (39:43)

Anyway, let's move on.

C (39:45)

Let's move to.

B (39:45)

Who's the early bird that got the worm?

C (39:47)

The early bird that got the worm. There's no sign here that's kind of an artifact of the nanobananna Pro, but early bird that caught the worm. This is Josh Kushner.

B (39:55)

I think so.

C (39:55)

Right.

B (39:56)

I like that.

C (39:56)

He got the worm. The worm was OpenAI?

B (39:58)

Yes. He was very early investor.

A (40:00)

Early with size, early with size.

B (40:02)

Doubled down many times. He got the worm. He's the early bird.

C (40:06)

Incredible.

A (40:06)

Yeah.

B (40:07)

Okay. Why is anthropic Donkey work.

C (40:11)

Donkey work. What does donkey work mean? 100% sure that this is like a real phrase. Apparently. I was asking how Tyler coinage. I asked GPT 4.5 and I asked Gemini 3 Pro.

B (40:22)

Okay.

C (40:23)

And they both said it was a real thing.

B (40:24)

Okay. It's boring or laborious. Part of a job. It's drudgery. Donkey work.

C (40:28)

Yeah. So I think there's a lot of ways you could do anthropic.

B (40:31)

Yeah. Anthropic is all over the place.

C (40:33)

So in this case, I think it's kind of that anthropic has gone very hard on coding and API and enterprise.

B (40:40)

Okay. The laborious work.

C (40:42)

It's kind of laborious.

B (40:42)

Yeah. Yeah. They're not doing, like, the really hot, sexy.

A (40:45)

Like, they're not using a browser.

C (40:47)

They're not solving science.

B (40:48)

Yeah. They're not doing consumer science. Yeah. They're doing the donkey work.

A (40:51)

This is good.

B (40:52)

Okay.

C (40:52)

I think it's some of the more laborious stuff, but there's a lot of ways you could do it.

A (40:55)

Daria's sitting there being like, who's gonna take all the jobs?

B (40:59)

Who's entering college?

C (41:03)

But you could do anthropic as the scapegoat. Right. David Sacks really does not like anthropic.

B (41:10)

That's true. They're kind of the scapegoat a little bit. They never get invited to the White House. Right.

C (41:15)

They're the lame duck.

B (41:17)

Lame duck. Is that a no? Lame duck is when the president is the second term.

C (41:22)

Ugly duckling.

B (41:23)

Ugly duckling a little bit.

C (41:24)

That's the one.

B (41:24)

I'm thinking there's a little bit of Chicken Little. The sky has fallen.

C (41:27)

Scaredy cats.

B (41:27)

Scaredy cats.

C (41:28)

Scaredy cats. They're worried about safety.

B (41:30)

There's a few of these, but I like donkey work. That feels the most accurate for what they're doing. They're going after the entry level white collar workers. They're doing the donkey work. Google artists love this image.

C (41:42)

I can see this is the last one. Google is the fat cat.

B (41:45)

Google is the fat cat.

C (41:46)

Before we had Google. Snail's pace. I think this would have been true maybe a year ago.

B (41:50)

Fat cat's so much funnier. Look at that cat. It's so fat. I love it.

A (41:55)

All the cash flow.

C (41:56)

All the TPUs.

B (41:57)

All the TPUs. They're GPU rich. They're rich in every sense. They have a ton of researchers they're releasing.

C (42:03)

I mean, now they have the best image model, the best text model, depending on what benchmark you look at.

B (42:09)

They're on top of the world.

C (42:10)

Yeah.

A (42:11)

Okay. Slightly concerned. We have a sign for golden goose, but there's no goose.

C (42:15)

Was removed.

A (42:16)

But who stole the golden goose?

B (42:19)

Yeah, who killed the golden goose?

A (42:21)

Or is there just no gold? If there's no golden goose on the. In the AI barnyard. Is that. Is that. Is that bearish?

C (42:27)

I think one take, maybe the golden goose.

B (42:30)

Wait, wait, wait. So first establish what the golden goose is. The golden goose is a goose that lays a golden egg. One every day. And in the parable in the story, the farmer kills the golden goose to get all of the gold inside, but it is revealed that by killing the goose, you no longer get the passive income from the golden eggs.

A (42:53)

So it's Jack from the beanstalk. Studied this.

B (42:56)

It's being a little bit too greedy. But we couldn't land on who the golden goose was.

C (43:00)

Maybe one take is the golden goose. They were the neoclass. And then the market is hurting, hurting. Is taking the goose away. Core weave. Not in the past what month. They've been selling off a little bit.

B (43:14)

Core weave continues to lay golden eggs, but the market got overheated and so that sort of killed the gains or something.

C (43:22)

And the gold is the GPUs. Subsidized GPUs.

D (43:25)

I don't know.

C (43:25)

In the data center.

B (43:26)

It's not quite.

A (43:27)

There's something with the bitcoin miners that just had a lot of energy that they. That they had, you know, effectively prepaid for. That they're unlocking some value from.

B (43:36)

Oh, well. Oh, well.

C (43:38)

But I think that's all of them.

B (43:39)

I think this is. Yeah, this is a pretty. Now it's pretty clear. This helps you understand how AI is playing out.

A (43:45)

There's no more questions.

B (43:46)

No more questions. If you. If you can. If you can see this infographic, it really makes everything crystal clear. Crystal clear. That's great. Well, thank you for taking us through that.

A (43:57)

Great work, Tyler.

B (43:58)

And let me tell you about Addio, the AI native CRM. Addio builds scales and grows your company to the next level. And get started. One company that didn't make the cut for our infographic was, of course, amd. And fortunately, there is a great profile in the Wall Street Journal of Lisa Su, the CEO of amd.

A (44:16)

And let's pull this up because this is a fantastic picture minus a very.

B (44:23)

Weird reflection going on very weird reflection in the glass. And so if you zoom in on this, there's something that is right in her nose that looks like a bugger or something. It's a very odd choice of image. Of course it's just a reflection on the glass, the photo, otherwise the way she's positioning herself, it's powerful. She did her job. Everything looks great from her perspective. She looks great, but the photographer just didn't quite catch that one reflection and so it just gives a very odd look. But anyway, let's dig into the article the chip CEO staring down Nvidia and talk of an AI bubble At a board meeting in late 2022, Lisa Su, chief executive of chip designer AMD, announced that she was radically changing course. Quote I'm going to pivot the entire company, she told the directors gathered around a boardroom table at the company's Austin campus. The rise of AI was a once in a lifetime opportunity, and the company had to put AI at the center of its entire product line. Three years later, the Santa Clara, California based company has nearly quadrupled in size, its market value rising from 90 billion to more than 335 billion. Despite a recent pullback, AMD's strategy of positioning itself at the center of the global AI race has paid off handsomely, making sue into a billionaire and her company into one of the only viable designers of powerful chips needed to power advanced AI models. In a market that has recently been completely dominated by Nvidia, sue is showing that there may be a place for a strong number two that can compete.

A (46:06)

Insane flow, by the way. Check this out.

B (46:08)

Yeah, look at that. She's 56 years old. She has a PhD in electrical engineering from MIT and a deep understanding of the physics behind her company's products. She developed a reputation as a giant slayer by out competing market leader Intel a decade ago to take the lead in producing central processing units, or CPUs. Now she's staring down the ultimate Goliath in Nvidia, the world's most valuable company and the foremost maker of the chips that power AI data centers. AMD will have to deliver on its promise to produce chips that are comparable to Nvidia's, most of the staring down her own cousin. It is of course, true. Investors bid up AMD's stock price in October after the company announced a marquee deal with with Oracle and OpenAI, which have both agreed to buy tens of thousands of AMD's newest generation of chips, known as the Mi 450.

A (46:58)

Mica says the Jensen Sue, Thanksgiving must be wild.

B (47:02)

I wonder if they're close enough to actually have Thanksgiving.

A (47:04)

They're first cousins once removed. Yeah, I think at that point it.

B (47:08)

Just depends on how hard you go for Thanksgiving. Some people have, you know, how big is your Thanksgiving? Do you go?

A (47:15)

I have a pretty small family.

B (47:16)

I have a pretty small family too. My in laws, we're still doing, we're talking 20 people, man. Maybe is like a typical Thanksgiving, but I know there are some people that are like, yeah, let's get 100 people together, let's turn it into Woodstock, which I'm down for, honestly. It's fun. Maybe you need, you know what you need to do this Thanksgiving? You need to get everyone who's cooking for your Thanksgiving dinner on Linear. It's the system for modern software development. It's a purpose built tool for planning and building products. The product you're building this Thanksgiving is a beautiful, bountiful feast. It's a feast, folks, and you need LINEAR to help you plan. Who's doing what? Who's making the pumpkin pie? Who's making the cranberry sauce? Can you assign tickets? Can you close those tickets?

A (48:01)

Are you using AI agents?

B (48:02)

Are you using AI agents? I think that there's something here.

A (48:07)

Everyone wants to build coding agents, nobody wants to build feast agents and I think that's wrong.

B (48:12)

Get on Linear. I love it. The OpenAI deal turbocharged the roadmap, said Sue. It represents a huge vote of confidence in the MI450, which of course we've seen from Semianalysis is performing very well. And AMD has caught up a ton this year. They've just been on a tear, taking feedback from all over the tech industry, integrating it very quickly and actually outperforming Nvidia with certain models. And all of that's been chronicled by the folks over at Semianalysis in their inference Max model. So the growing AI market is a huge opportunity and we want to make sure that we have deep partnerships that enable to get us a big piece of that. The rest will handle itself. The deal making was on display again this week when the company said it would work with Cisco and Saudi Arabian AI venture firm and a Saudi Arabian AI venture firm to build a large cluster of data centers in the kingdom. Sue attended a blog that was humane. I think so. So SU attended a black tie dinner for Saudi Crown Prince Mohammed bin Salman alongside other executives at the White House Tuesday. And AMD sent a representative to a US Saudi Arabia investment forum on Wednesday. Until now, data centers have mostly been entirely Nvidia's market. One tech analyst said AMD is in a position to take a more meaningful piece of that. So $1 trillion a year here. Let's get into it. At the heart of Sue's strategy is her belief that there is insatiable demand for computing power and that as the as the market for AI grows, the companies offering the best and most reliable AI infrastructure will thrive. She said she believes AI is a is not a zero sum game and that recent concerns about an overheated market for chips and data centers are exaggerated. In recent weeks, the massive spending on data centers have intensified concerns that an AI bubble is building. AMD stock, which had jumped nearly 60% in October, has fallen about 20% this month. I'm not concerned about AI bubbles, sue said in the interview. I do think that those who are thinking that way are a bit too short sighted. They don't really see the power of the technology.

A (50:16)

It's actually only down. I don't know exactly when this article was released, but 13¾ in the last month better than some others.

B (50:29)

This is not the time to stay on the sidelines and worry, hey, am I over investing? She said. It's much more dangerous if you underinvest than if you over invest, in my opinion. Let's go. I love that. A year ago she predicted that the market for AI chips would hit 500 billion in sales annually by 2028. At an investor presentation last week in New York, she was even more optimistic. The market for AI and data center computing, she said, will reach 1 trillion a year.

A (51:00)

Interesting that she under predicted or seemingly under predicted because Jensen by himself is saying that he's that Nvidia has a line of sight to 500 in 2026.

B (51:11)

Half a trillion. And that's just for them. Revenue, full year. That is a lot of money. Wow. Very exciting. Well, speaking of Nvidia, sue is also helping to capitalize on inflection point in AI. Many executives in the chips industry, including Nvidia chief Jensen Huang, expect demand to shift from clusters designed to assist in large training runs to inference. And inference doesn't require chips with as much computing muscle as training does. According to the Wall Street Journal, as AI models like ChatGPT and Gemini become increasingly integrated into daily life, and as companies design thousands of enterprise software tools that rely on AI models, demand for inference functions is about to go up by a billion times, Wong said last month. Zuang and Su and Wong are distant cousins, but they didn't meet until both were established executives. That's interesting So I don't think they'll be hanging out at Thanksgiving together if they just met recently, but maybe.

A (52:09)

Plus, that's just a story and their entire family. Both their family lines have been colluding to corner the AI chip market for generations.

B (52:18)

It does feel like a storyline from Dune or something. AMD has a strong line of inference computing chips, but has struggled to design chips powerful enough to compete with Nvidia in training. Let's see, they go into a little bit of her backstory. She was Taiwan born, Queens raised. She worked for IBM where she focused on designing products according to the needs of big customers. When she took over as CEO of AMD in 2014, the company had a market value of less than 3 billion. It's up 100x since then. She did deals with Chinese partners that helped stabilize the company's finances, move the chip designer deeper into the processors that powered PCs and data centers, capitalizing on Intel's weakness. Today AMD sells an estimated 41% of data center CPUs, up from essentially 0% five years ago. What a run. It was that success that allowed AMD's board of directors to trust sue when she proposed in that fateful board meeting in 2022 to pivot the company to go head to head with Nvidia. One of the board members attended Said told this story. She had a vision around our ability to be AI from endpoint to data center and everything in between, said Talwalkar, one of the board members who attended the meeting. She had the conviction of the need and the need to move fast. What do you think about vision Jordi? Is it good that she had vision?

A (53:39)

I think vision is still underrated. You're pro vision, pro vision.

B (53:42)

It's a hot take. Some people don't like vision, some people are anti vision these days. I think I land on the side of provision. I think it's important.

A (53:50)

Sue was able to dispense with intel as a competitor because intel had self inflicted wounds, said Daniel Newman, CEO of research firm Futurum Group. That's in sharp contrast to Nvidia, a company that's expected more than 200 billion in revenue next year. That has gotten thousands of AI developers hooked on the proprietary software that Nvidia chips rely on. But Nvidia can't make enough chips to keep up with Demand, and the OpenAI deal proved that large AI companies are eager to diversify their supplier base.

B (54:20)

We got some major, major update here on Lisa Su as a celebrity. As AMD has become a bigger player, she's begun to embrace the spotlight. Earlier this year, at an industry summit in Paris, sue was treated like a celebrity, with many young women asking her to take selfies and sign autographs. That's amazing. Jensen Huang was signing autographs couple months ago. That was this year, maybe it was.

A (54:46)

Like five, six months ago that one year.

B (54:49)

Well, now she's also signing, also signing autographs. As, as the, as the industry grows, more and more celebrities are minted. Well, in July, sue attended an AI summit where Trump rolled out his AI strategy. When she addressed the audience, sue pulled out one of the company's chips out of her pocket and said it had 185 billion transistors and took nine months to make. Such chip can cost tens of thousands of dollars. Trump later mentioned her in his speech as one of the notable executives in attendance. She was still in the shadow of Wang, however, who Trump asked to stand while the crowd applauded. What a job you've done, the President said to the Nvidia CEO. So he's like, thanks for Lisa Su for being here, but let's give it up for Jensen. He's given it up too much for Jensen. He's not having a balanced approach.

A (55:42)

Well, I don't think, think it's, it's necessarily his job to.

B (55:46)

Well, ten times the market cap. Ten times the applause, perhaps.

A (55:49)

That's right.

B (55:50)

Market cap weighted by applause. Let's see. Over the next decade, AI will transform every industry, every business, every product, every interaction. Our chips are enabling this massive new revolution. Well, very, very interesting. I'd like, I would like to, to learn more about Lisa Su as a celebrity. I want to see more, more deep dives, more Page Six style writing is what I'm looking for.

A (56:18)

Maybe a New York Post cover.

B (56:20)

Exactly. A lot of, A lot of caught on the street. A lot of like, you know, like, what type of coffee does she like? That's what we need to know. We need to get her in the Mansion section. Should we do some Mansion section?

A (56:32)

We do. We should.

B (56:32)

There's a big question in the Mansion section today. First, let me tell you about graphite.dev, code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster. The question I know everyone has been asking is, is 200 million, the new $100 million in luxury real estate? What do you think?

A (56:49)

Many people have been asking that.

B (56:50)

So there's a $200 million house in Indian Creek, a $250 million house in Bel Air, a $300 million house in Aspen, and a $205 million house in Palm Beach. And this is funny because I we've reviewed like almost all of these houses.

A (57:07)

We've actually old rock, says Lisa sue on Hot Ones.

B (57:10)

That would be great. Maybe on Theo Vaughn as well. A surge of ultra rich buyers has pushed asking prices to new extremes. Yet many headline grabbing mega mansions languish on the market or trade for a fraction of their debut numbers. Just a decade ago, the 100 million price tag was still considered a new frontier for luxury real estate. The first nine figure home sale occurred in 2011. Wow. By 2019, there had been about 20 sales recorded at that price point. Now, real estate insiders say a new price pricing benchmark is setting the tone for the high end market. 200 million. Since 2025, at least five major US properties have listed for 200 million or more, mostly concentrated in South Florida on Indian Creek, a private island near Miami Beach, Florida, often called the Billionaire Bunker, cosmetic surgeon Dr. Aaron Rowlands and his wife Maureen Rowlands listed their unfinished waterfront estate for $200 million earlier this month. About 70 miles north in Palm Beach, a circa 2005 mansion once owned by the late Frank and Maureen Wilkins is asking 205 in Aspen, the longtime home of California billionaire Stewart and Linda resnick, listed for 300. The listings raise the question, is this aspirational pricing or as the ranks of the global billionaires expand at a rapid clip, is it possible that the market, that the top of the market has actually doubled? I would say yes, it's totally possible. Look at the run up in the stock price. The assets are more valuable than ever. There's going to be a massive wave of wealth created from the AI boom.

A (58:42)

I would not want to be in the market for a single family home within 20 square miles of OpenAI.

B (58:51)

Yeah, there was some post. I don't know if it made it in the timeline, but there was some post about there's like 40 people that are looking for houses in the five to seven million dollars range in North San Francisco and there's none for sale.

A (59:05)

Because all cash offers I'm sure from.

B (59:08)

The, from the liquidity.

A (59:10)

Also Porsche allocations are going to be pretty rough out there.

B (59:13)

You think so?

A (59:14)

I expect it. Did you see the Elect in the chat? Says Lisa sue is a car enthusiast, drives a metallic blue graphite Porsche.

B (59:22)

Ooh, I like that.

A (59:23)

She apparently names them after her, after her chips. So she has cars. Ryzen, Epic and Radeon.

B (59:32)

That's cool. That's very fun. Yeah. This Resnick, this Bel air House, the 250. I remember digging into him and he was part of the telecom build out in 2001. And so even though his company went bankrupt, he still wound up with with a fortune that allowed him to buy a huge house. You could think about the same thing happening in the AI world.

A (59:54)

Well, here's another note. AMD sponsors the Mercedes AMG F1 team which has a new co owner, the CEO of CrowdStrike. So all eyes on George's team this weekend. We'll see how they do with their new co owner.

B (60:19)

Before we move on to our next story, let me tell you about FIN AI, the number one AI agent. AI that handles your customer support. The number one AI agent for customer service. Silicon Valley estate sells to mystery buyer. I want you to guess who it is, Jordy. I don't know. I want you to make wild guesses. In California, Silicon Valley, a roughly 12 acre estate designed in classical Italian style has sold for more. Sold for $56 million.

A (60:48)

There we go.

B (60:49)

Known as Via del Plato, the property is located in Portola Valley, one of the country's wealthiest towns. The sellers are venture capital investor Bandel Carano and his wife Paula Carano, who built the estate after buying the land in 2013 for 13.625 million, property records show. Hmm. Have you looked up Bandel Carano and what deals he did to be able to afford such a wonderful house? I would love to know the story of his venture capital work. The Carano's put the home on the market for $85 million in February. The most recent asking price was 65. It went for 56. It's the priciest home to sell in Portola Valley to date. The prior record was 35 million for a 13 acre estate. The buyer, who currently lives in San Jose, searched for several years before purchasing the Portola Valley property, but they didn't want to name the client. The Kharanos didn't respond to a request for comment. Massive mess. You got to get.

A (61:54)

Mandel Carano was an investor in Polycom Inc.

B (61:57)

The Polycom? No way. That's amazing. Well deserved. One of the most iconic conference calls. Phones.

A (62:07)

We used to have one.

B (62:07)

We used to have one on our desk. It's probably in the studio. It was a prop that we used for a long time. I think it's in the back.

A (62:13)

You got to bring it back.

B (62:14)

We never actually used it. We wanted to have it.

A (62:16)

How long until somebody creates like an AI Polycom?

B (62:20)

I don't know. Yeah, we. That's so crazy.

A (62:24)

He.

B (62:24)

He did the deal in Polycom and made a bag. I love it. Listing agent said okay. They built after raising an old home, completing the project in 2021. In addition, the roughly 23,000 12,300 square foot main house property house has a pool, a pool house and then another unfinished 5,000 square foot building has a ballroom and a wine cellar.

D (62:50)

Wow.

A (62:50)

There we go.

B (62:51)

They designed the property. He's a general partner at Oak Investment Partners, a venture capital firm with offices in Wilton, Connecticut and Menlo Park. In 2024, the Khoranos listed a nearly 2,000 acre ranch outside of Bozeman, Montana. Portola Valley had a median sale price of 5.5 million. Interesting. What else is in here that we wanted to look at? So there is. There is some articles on San Francisco. What's going on there? San Francisco is back after pandemic related struggles and growing up Affluent young families demanding space have elevated the 94127 zip code which is anchored by the walkable West Portal Village center and encompasses desirable residential neighborhoods such as St. Francis Wood. This west side pocket reliably offers a rare San Francisco luxury detached single family homes with yards. The suburbia in the city lifestyle has recently propelled the zip code to rival historically more expensive city enclaves. In October, its $2.5 million median listing price was second only to the 2.7 million dollar median price in Premier Address, Pacific Heights. Interesting. So there are a bunch of shops, people are moving in. The median price per square foot is $952 and the houses are moving. 37 days on the market is all it takes for the median house to find a new buyer.

A (64:24)

Pretty good.

B (64:25)

Let me tell you about Profound get your brand mentioned in ChatGPT. Reach millions of consumers who use AI to discover new products and brands. Should we let AI ruin the em dash? I think the game's over.

A (64:38)

It's over.

B (64:39)

I think AI has already ruined the EM dash. I think it's ruined. I think we have to move on from it entirely. But Joel Stein at the Wall Street Journal has another take. Joel says don't let AI ruined the EM dash and writes a defense of the newly controversial punctuation mark.

A (64:58)

It's so funny because EM dashes are not even the way that I first identify a lot of AI writing. Like it's often the it's not this, it's that. That hits way harder. Totally, Joel says. A few weeks ago, my 16 year old son Laszlo, a student journalist, received a tip for his school newspaper. A source had told him that the administration, which had banned all use of AI had used it to write an email to parents. The email in question was certainly boring enough to be written by ChatGPT, but it was also boring enough to be written by a school administrator. We ran the text through two AI detection tools, and both determined that there was a zero percent chance a computer wrote it. I don't know how a computer figures out if it's dealing with another computer, but I'm guessing it asks which of nine photos are motorcycles. So why did this source think AI had written it? Because the email contained EM dashes. The EM dash, a punctuation mark a bit longer than a hyphen that denotes a long pause, has become the black light on the hotel sheets for AI shamers. A sign that an essay.

B (66:04)

The black light on the hotel sheets is a good turn of rays.

A (66:07)

A sign that an essay, a letter, any kind of written work was written by a machine and not a human. Yeah, the EM dash is now a GPT ism. It is not advisable unless you want people to think you're writing as an output of an LLM.

B (66:18)

Yep.

A (66:18)

On Instagram, the beauty influencer Lux Gen warned that if people don't want to be accused of using AI, they should take out the ChatGPT hyphen.

B (66:26)

They just call it the ChatGPT hyphen.

A (66:28)

Last week, OpenAI CEO posted on X the following brag, small but happy win if you tell Chat GPT not to use EM dashes in your custom instructions, it finally does what it's supposed to do. It's interesting that they just didn't release an update that eliminates it. Like, they seem to think it's a feature, not a bug. And they're asking people to kind of opt out of it by doing custom instructions, which I'm sure like 1% of the user base actually uses.

B (66:53)

Yeah, it's weird that they couldn't just have a custom instruction that goes over all of ChatGPT that just says like, you don't need to never use the EM dash, just let's turn it down by 90%.

A (67:05)

Yeah. All this M dash shaming has been upsetting because I, a professional writer, love, love, love an M dash.

B (67:12)

And he uses a lot of M dashes.

A (67:14)

In my writing, the chat chatgpt tells me I use 11 em dashes every thousand words.

B (67:19)

Wow.

A (67:19)

Other professional writers use them a lot too. 4.4 times more than the average person. These numbers, upon investigation, could be highly inaccurate Since I got there from ChatGPT, which explains one reason why my son schools bans him. It but the point is, EM dashes are loved by professionals. This is starting to make a lot more sense. What is ChatGPT trained on? A lot of professional writing.

B (67:41)

They literally have a deal with the Wall Street Journal.

A (67:43)

And so if it's used 4.4 times more than in everyday writing, it's just going to come up a lot.

B (67:49)

Yeah. I mean, if you tried to, you have to give it some direction on like where it's going to sound like. Yeah. And if it just winds up being like, yeah, make it sound like a, like a Reddit comment, it's like, okay, well that could go all over the place. That could be very, very weirdly worded and it just would not be very consistent.

A (68:06)

Yeah. So the point is, EM dashes are loved by professionals. Why? Because more than any other punctuation mark, the EM dash is deeply human. It's the breath marks of Emily Dickinson, the stream of consciousness of Virginia Woolf, the head clogging maximalism of David Foster Wallace, or the self aggrandizing asides of Joel Stein. It's also the mark of discretion. When Holden Caulfield tells his writer that he loves it when his classmates digress during a presentation, he says, I don't like it when somebody sticks to the point all the time. The boys that got the best marks in oral expression were the ones that stuck to the point all the time. I admit it. And he admits it naturally with an EM dash. Large language models consume a lot of the best writing out there, which is why they use EM dashes. Unlike most people today who think a period is a rude way of screaming, AI believes in punctuation. If there's anything to take away from the EM dash scare, it's that we should use more of them. Because while most people can't find one on their keyboard and wouldn't know when to use it if they did, that's pretty funny. The EM dash requires more nuance than a thumbs up, a heart, or an eggplant. It's less jarring than parentheses, but a bigger interruption than commas. It's the length of an M which sets itself apart from the boring shorter EN dash in the men's locker room. For writers, we make fun of en dashes.

B (69:27)

Interesting.

A (69:28)

Most of all, we should use EM dashes because they are a declaration of humanity in the face of AI's onslaught. If only ordinary people would learn how to use them.

B (69:38)

Do you, do you see this link? Shorter en dash. For some reason it linked to en-dot in and that website just goes nowhere.

A (69:48)

En dash. I don't know why on an Indian domain.

B (69:52)

I don't know. That's just like a mistake or something. They should correct that because EN-IN doesn't go anywhere. Yeah. I have never really used EM dashes. I also, I like this take. I mean, I agree with this. This makes sense. He should defend the. Defend the castle. Maybe the bigger tell to your earlier point is around that contrastive parallelism. That it's not this, it's that. That's what it's called. And that contrastive parallelism. I was wondering about that. I went through a bunch of the old scripts that I'd written for YouTube and I couldn't find a single example of ever using that style, that structure. I'd written a lot about a lot of different tech companies and I never said, you know, I've made a whole video about Nvidia in my Nvidia video. I never say Nvidia isn't just a GPU company.

A (70:48)

Google isn't just a search company.

B (70:50)

I've never. I never said that.

A (70:51)

Holding company of the future.

B (70:52)

Yeah. Yeah. I don't know. I don't know how that got baked in. I think that's maybe from marketing, from, like, marketing lingo. Because I was trying to run the benchmark of. Does this contrastive parallelism creep in.

A (71:09)

More.

B (71:11)

When you ask the. Let's see, what is it? It's antithetical parallelism or contrastive construction or correlative conjunction pairs. This is the term. I learned this from them. But that antithetical parallelism, it's not this, it's that we're not writing a story. I never use it. And it has a sound that sounds like marketing. It sounds like it sounds like marketing. Like, I feel like you could probably go through the New Yorker, the Wall Street Journal, and you'll see a lot of EM dashes, but you won't see a lot of antithetical parallelism. And so I feel like the antithetical parallelism is an even bigger tell. And Tyler, what do you think?

C (71:53)

I also think it's interesting. Like, you would imagine that a lot of these, like, GPT isms.

B (71:58)

Yeah.

C (71:58)

Would be. They would only be like, with one model.

D (72:03)

Right.

C (72:03)

It's like just four O does that where.

B (72:05)

Totally. Totally.

C (72:05)

But it's like not that at all.

D (72:07)

Yeah.

C (72:07)

It's also just like OpenAI models. Like yesterday I tried the Gemini 3 Pro and I asked write an ad for a GT3RS.

B (72:16)

Yeah.

C (72:16)

And then. So I'll just read the start.

B (72:18)

Listen to this.

C (72:19)

There's a moment at 9,000 rpm where the world stops. The noise isn't just sound M dash, it's frequency. The vibration isn't just an engine EM dash. It's a heartbeat. Like it's like the exact same.

B (72:30)

It's very similar. It's the very similar.

C (72:32)

It's very strange. I wonder like why this is like. There's a case, you could say that models are just like training off each other, but it seems that seems like not something that they're doing.

B (72:44)

Yeah, I don't know. It's just kind of landed in this zone where that is what is seen as good or something. I don't know. And it's just like sort of decided that antithetical parallelism is the greatest thing. So thank you to Joel Stein at the Wall Street Journal for writing this article. I will support you in your fight to keep the EM dash. But I want to hear your take on contrastive construction.

A (73:13)

Actions speak louder than words, John. Are you gonna just rip like 20 em dashes in your next essay?

B (73:19)

So the problem is that I have never. I'm the guy. I'm the one who doesn't know how to find it on my keyboard. I'm who he's talking about.

A (73:28)

Just hit dash, dash, space, find it on your keyboard.

B (73:31)

And wouldn't know when to use one if they did. I don't know when to use it. I don't know where to find. Wasn't taught to me when English class. I just never learned.

A (73:38)

I'll teach it. I'm teaching it to you right now. Go on your keyboard, hit dash, dash, space on imessage and it will create it for you.

B (73:46)

Okay, but when doing it, so do shift, option, dash, shift, option, dash. I don't know. I think you can't teach an old dog. I don't think you can teach an old dog new tricks. I don't know if it's possible to teach me to use an EM dash reliably. I think I might just throw them in randomly. What if I start using two EM dashes next to each other? Just throw people off. No one would expect that. Right. It's completely unexpected. So you'd be like, oh, well, this looks like AI, but it's something weird.

A (74:17)

It would be a really bad model.

B (74:20)

Whatever model he used was terrible. GPT2 for this. Anyway, let's move on. Let's do some timeline and then I think we can come back to this article and solve and see if we have more time. But first, let me tell you about TurboPuffer serverless vector and full text search built from first principles on object storage.

A (74:39)

Used by cursor notion linear anthropic, 10x cheaper, extremely scalable.

B (74:43)

I like when we're talking over each other during the ad reads. That's really the flavor we want to bring through. I love it. Okay, so there's a major shake up going on at TBD Labs. Augustus Odina is out and I'm giving him a follow for this. He's going. He's going independent free agency.

A (75:01)

Yesterday I resigned from TBD Labs Meta AI. I wasn't there for very long, but I think I got a few useful things done. It's an impressive group of people and it's especially impressive that it got assembled as quickly as it did with such a high talent bar at a large company. Quote, unquote, founder mode is real and good. I will certainly miss my coworkers there. I think now is an unusually high leverage time to pursue ambitious new projects at the intersection of AI and other technologies. Please reach out to me if you're interested in that sort of thing. And I expect I will have something more detailed to share in not too long.

B (75:38)

Is that the sound of a thousand venture capitalists writing term sheets with blank numbers on them?

A (75:46)

I messaged him yesterday and he said at the end of the show would be honored to be on the show when ready to talk about it.

B (75:53)

Yeah.

A (75:54)

So looking forward to that. Bernie Sanders.

B (75:57)

Okay, what's happening with Bernie?

A (75:59)

Firing shots.

B (76:00)

He's taking shots of the AI oligarchs. He says, here are some of the most powerful AI oligarchs in the world enjoying a private dinner with a dictator who murders his own citizens with a bone saw. Does anyone really believe they want to wipe out poverty or improve life for ordinary Americans? I don't. And he's sort of mogging Greg Brockman here because he leaves Greg Brockman's net worth. He doesn't even try and estimate it. He just says question mark, question mark, question mark.

A (76:26)

Bernie Sanders.

B (76:27)

Age 75.

A (76:30)

Guess again.

B (76:32)

78.

A (76:34)

Guess again.

B (76:35)

77.

A (76:37)

Guess again.

B (76:37)

80.

A (76:38)

Guess again.

B (76:39)

82.

A (76:39)

Guess again.

B (76:40)

I have no idea.

A (76:41)

84.

B (76:42)

84. Oh, he's up there.

A (76:43)

He's up there. Oh, wow. Okay, so Grandpa is pissed.

B (76:49)

He's pissed.

A (76:50)

He's absolutely.

B (76:51)

But what does that do with him not being able to look up Greg Brockman, the CTO of stripe? Do a back of the envelope. What's stripe trading for in the private markets? Call some secondary brokers. Get. If you're Bernie, you just need to call a bunch of secondary brokers understand where stripe is trading, then work backwards for a from what his stripe stake is. Then call Satya Nadella. Ask him based on he could have used. Are you marking OpenAI at 500 or do you think you should market more at like 400 or maybe 600? Are you bullish on that? All of that goes into the dcf. He's build a spreadsheet and then you would have a fair market valuation for Greg Brockman's net worth and you wouldn't.

A (77:25)

Need to use chat. Got pretty close. 83, 85. Good work. But if he really wanted to stick it to these American AI oligarchs, he could have used like Kimmy to do the kind of analysis that you were talking about. He would have gotten a lot closer. I think we probably need to do more to define oligarch in the AI era. Right. Does a billion dollar.

B (77:56)

I thought people would use Plutarch. Plutarchy. Plutarcy.

C (77:59)

Plutarch is a person.

B (78:01)

But what's our lives? Plutocrat. There's one that just means technocracy. Technocracy. There's one that specifically means wealth. That's like wealth where the plutocrats. That's it. Right. Plutocracy is. I think plutocracy is plutocracy.

C (78:18)

But also, I mean, does Bernie not.

B (78:20)

Realize that Greg Brockman Government by the wealthy is plutocracy. Oligarchy is something else that's just ruled.

C (78:28)

By a few, I think.

B (78:29)

Ruled by a few. Interesting. A small group of people having control of a country, an organization or institutions. Well, you know what Bernie's got to do? He's got to start a lab, start a foundation, model company and compete in the free market.

A (78:46)

I like it. I like that. Active retirement.

B (78:50)

Yes. And then start trading on public.com investing for those that take it seriously. Multi ass investing. Trusted by millions.

A (78:57)

Yeah. I don't blame Bernie for being a little bit salty at this point because he's never on those charts of the best political traders in history. That's true. Right. He's been having to sit there and watch Nancy Pelosi run it up the most insane runs and he's just missed.

B (79:13)

Out the FOMO Must be crazy. Every day he's opening up and Nancy's up another mill and he's just sitting there like, ah, I missed it. I missed it. The trade of the century. Sam 3d. Meta, we talked about this a little bit. What?

A (79:27)

Yeah. Atharva is hyped on Meta's new model M3D.

B (79:34)

I forgot about this. So this was purposely released at the Perfect time, because we were just earlier accusing Meta of slopping it up at the slop trough with meta vibes. And there's a little bit of a narrative that was emerging which is like, okay, you know, Zuck did this whole talent war. He acquired a ton of talent, clearly invested crazy amounts, huge gigawatt five gigawatt plan, Hyperion, Prometheus, he's going all in on AI. And then the first thing that he came out with was like, meta Vibes, this sort of like sloppy TikTok AI clone that didn't really break through and just kind of frustrated everyone in the timeline, did not appreciate it, did not like it. And so at that point, I think everyone was calling for, okay, you got to deliver something that's actually solid. You got to deliver something that's useful, that's valuable. Maybe it's a contribution to open source. Maybe it's even just an API for coding that's really good, or a model that does something unique, special, or just.

A (80:33)

Free ASI for everyone.

B (80:35)

That would also be acceptable. Yes, that would definitely reset the narrative on is Meta slopping it up at the trough. But this is an example of them not slopping up at the trough. This feels elevated. This feels useful. I'm very excited for where this gets implemented all over the place. This feels like a very serious organization, like the work of a serious organization, the work of a group that's taking their work seriously and not just trying to score points, go viral, just make a couple extra bucks. They're actually doing some real research. So I was excited about this. What was your take?

C (81:06)

Yeah, I mean, this is really cool. I think people have probably just overindexed too much on Vibes, because I don't even really think of Vibes as being like an Alexander Wang project. It's not like they trained a new model that's not their model. It's just project managers need a new app. So I think there's still a ton of really good researchers. I'm very excited to see what they actually put out because I assume that it's going to be great.

A (81:29)

It's hard to know without actually knowing the usage of Meta AI. You can look into it a little bit by just looking at the App Store charts and seeing the kind of traction they have, but I think in hindsight, they would have just been much better off not releasing that or not making it a moment.

B (81:44)

Dude, you're telling me you don't like Meta Vibes? Dog, what are you doing?

A (81:49)

That is Trump in an orange and rainbow suit dancing.

B (81:54)

I thought we had ptz. I thought we could zoom in on this. Yeah, I mean there's some golden stuff. You know what I was actually thinking about was they kind of, it's kind of a do nothing win situation. Right. Because I feel like at least this year the vast majority of AI slop will be consumed on Instagram. Like no matter even if it gets made in the Sora app, like it's coming over to Instagram if it's going to do well and it's going to be distributed on the other platforms. And so I would be very closely watching the actual user minutes, the time on site. But I would imagine that the amount of Sora generated content, the user minutes is like 100 times on Instagram than on the actual Sora app. Just because it's such a bigger, bigger audience. And so even though it's made by Sora, it's distributed elsewhere anyway.

A (82:50)

Yeah, let's check in with the what's.

B (82:53)

Going on app store chart genius.

A (82:55)

Sora is still number nine in the charts. Capcut and Temu Shop like a Billionaire are sitting just ahead alongside. Go wish your digital wish list.

B (83:08)

Well, let me tell you about numerl.com let numero worry about sales tax and VAT compliance. Compliance handled so you can focus on growth.

A (83:19)

Luxury watch guy is concerned about the bitcoin sell off because he says there are going to be so many effing APs on the market facepalm flooding the aftermarket. So we'll see how that nets out. But the cryptocurrency industry certainly has been known to enjoy an AP.

B (83:45)

Yeah, Google's Genius. By not selling TPUs, Google allows Nvidia to maintain high GPU pricing which in turn props up price of inference. Google then captures that inference price Premium by running TPUs for inference on GCP. This is a conspiracy theory by Dave 3 and 30 Tepper. I'm not sure how real that is.

A (84:10)

Yeah, there's also they are going to sell TPGs. Just because you have two players in the market doesn't mean you can sort of have what happens in the soda industry. Right. But effectively, like I was thinking about.

B (84:21)

GLP1 market, there's just a lot of demand. You don't need to cut prices the.

A (84:25)

Price that and you also understand that you can get in there. There can be reason not to get in a race to the bottom dynamic and effectively have a gentleman's agreement to keep prices in a certain range. Of course that can go too far but.

B (84:40)

Well, let me tell you about Vanta Automate Compliance and Security. Vanta is the leading AI trust management platform. Do you mind doing some timeline?

A (84:51)

Of course. Pier from Cal.com says shower thought why is no one doing outbound for pizza? Hey this is Gigi from Gigi's Pizza calling. You ordered last week. We have a pepperoni pizza ready and could deliver it in 10 minutes. Are you hungry? Tech sales guys has trillion dollar idea and of course it goes viral again. Some ideas and and bangers are evergreen. Morgan Housel yesterday was saying I think the majority of societal problems right now are all downstream of housing affordability. Austin asks how would you fix housing affordability? The current admin has suggested 50 year mortgages. We'll see how much of we'll see if those go through and how much of an impact that they have. But Morgan Housel says build more homes, plenty of demand, plenty of capital. Mostly a local zoning issue and that certainly, certainly tracks living here in California. John Palmer, dear friend of the show and somebody that we are actually working on a little project with says this might sound crazy but they should pivot. He's talking about Sunday Robotics which make the cute new robots with with fantastic little hats. He says this might sound crazy but Sunday should pivot to a consumer app where you just tele operate your own memo which is the name of the robot and meet up with your buddies and do fun stuff as robots. I agree. I would love to meet up with the lads in the real world in robot form. I think he might be onto something. There was a story coming out earlier this week. An investigation from the lever found that a PE firm is buying hockey rinks and banning parents from recording and taking photos of their kids games. Based. This story broke.

B (86:58)

Why is it based?

A (87:00)

Because they tried enforcing the rule on the wrong person. US Senator Chris Murphy So a little excerpt here. As the $40 billion youth sports industry comes under private equity control, Corporate owned facilities and leagues from hockey rinks to cheerleading arenas have begun prohibiting parents from recording their own kids sports games. Instead parents are forced to subscribe to these companies exclusive recordings streaming service which can cost many times more than the streaming costs for professional sporting events.

B (87:35)

That's so funny because I don't think this is. I was thinking about how, how like I, I'm such a gear nerd that oftentimes when I'm watching my four year old play soccer I'm like one of these days I'm gonna bring out these like PTZ cameras out here and like get the most cinematic footage and they're Gonna have to go up against me. I'm gonna be fighting them.

A (87:59)

Yeah. Every kid's sports team should have a dad or a mom. That's the cto and their job is just to rig the entire field with cameras and then provide crazy edits for each kid after. Easy. Even if they lose, it's like the one, the one moment that they had, you know, they finessed someone. You still think this is based. Tyler, you still saying hello base department? I mean I my mind you appreciate value creation.

B (88:30)

How else, how else do you expect private equity firms to make money? Yeah, how else do you expect them.

A (88:34)

To make hard working private equity firms? They need to make, build, you know, afford their interest payments.

B (88:41)

Yeah.

A (88:41)

In some instances, parents have been threatened that if they chose to defy the rules and record the game, they may end up on a blacklist that punishes their kids.

B (88:50)

Well, what they need to do is they need to form some sort of like union, basically where they all say we all are gonna film or none of us are gonna attend and we're gonna take our business elsewhere.

A (88:58)

Those threats were even reportedly made to a US Sitting. A sitting US Senator. I was told this past weekend that if I livestream my child's hockey game, my kids kids team will be penalized and lose a place in the standings, said Chris Murphy at a public event. Why is that? Because a private equity company has bought up the rinks. So this. He was trying to live stream himself. Like he was just trying to set up a camera and live. Like how does that. Like at some point, I think the obvious solution here is parents can take as many pictures as they want. If I'm a parent, I'm gonna take some pictures, but I would probably still get the recording. I would probably still subscribe so that family members could watch.

B (89:41)

Yeah, yeah, yeah.

A (89:42)

It doesn't have to be all or.

B (89:43)

Not competing the free market. Let us both film. Whoever gets the better vibreal, whoever gets the craziest edit out of it is going to be able to sell it. And if I can out edit you and I bring better gear than your private equity backed videographer. I'm not paying. I'm not paying for it.

A (90:01)

It.

B (90:02)

Let me tell you about figma. Think bigger, build faster. Figma helps design and development teams build great products together. You can get started.

A (90:09)

Triple glaze.

B (90:14)

Every time OpenAI says anything, it gets turned into a headline. These days, the latest one is that.

A (90:21)

The keywords start using sign language, headwinds. But I guess even that would still potentially get turned into a headline.

B (90:30)

Yeah, I don't know. It's.

A (90:32)

Sam is very tricky.

B (90:36)

I feel like people use. I feel like if you went through. If you went through a variety of, like, earnings call transcripts from plenty of public companies. CEOs say the word headwinds all the time. They're like, oh, yes, like this particular thing, the tariffs. That's a headwind. Oh, this is a headwind. Interest rates might be a headwind. That might be a headwind. Competition might be a headwind. Like, people throw this out all the time.

A (91:00)

Yeah. If you say as a business that we don't have headwinds, you're not going to be taken seriously.

B (91:05)

Yeah, yeah. But OpenAI has somehow wound up in a situation where they're so priced to perfection that even just admitting that there might be more competition is like a massive news cycle. Now, it's pretty remarkable, but the actual quote here is that Sam Altman warned of headwinds from resurgent Google. This, of course, is a story that we've all been tracking through Gemini 3.

A (91:30)

Some temporary economic headwinds for our company.

B (91:34)

Economic headwinds is pretty wild.

A (91:37)

Didn't have to say put that part in.

B (91:40)

I wonder how that's going to manifest. Are people going to be unsubscribing from ChatGPT going over to Gemini as an app, or are they going to. Is this on the API side? The API side seems a little bit more easy to slide around from one to the other. It feels like, like OpenAI certainly caught up very quickly with Codex. They were behind for a while and they caught up, but now Google's sort of out in front. Although Anthropic's doing very well. It's all just like a variety of horse races. Some of the revenue should be stickier than others, but when the goal is 100 billion in revenue so fast with this series of competition, it is very, very difficult.

A (92:22)

He also said, ChatGPT is AI to most people and I expect that that to continue. So brand that is 100% accurate, powerful element, but not enough to eliminate all competitive threats.

B (92:35)

Okay.

A (92:35)

He added, I don't want this note to be a downer. We are doing remarkably well as a company and I expect that to continue.

B (92:40)

Yes, yes. Let me tell you about Julius AI, the AI data analyst that works for you. Join millions who use Julius to connect their data, ask questions and get insights in seconds. What is this story of the FBI raising a bounty for a former Olympic snowboarder who turned alleged drug kingpin?

A (93:02)

We love these stories. We covered this guy on the show.

B (93:04)

Before okay, his name's Ryan Wedding. Ryan James Wedding.

A (93:09)

He's married to the game, is he?

B (93:13)

What is that?

A (93:14)

No, Ryan Wedding.

B (93:16)

Oh, Ryan Wedding. Okay. Yes, he's married to the drug dealer game.

A (93:19)

I suppose he's a former Olympic snowboarder from Canada. He represented Canada at the 2002 Winter Olympics in men's parallel giant slalom. After the Olympics, it is alleged that he became a transnational drug trafficker and orchestrated the murders of various witnesses.

B (93:36)

Whoa.

A (93:37)

And he's only 44, so he's kind of just kind of hitting his stride. He was added earlier this year, March 6th to the FBI 10 most wanted. So he's in the top. Top 10 from.

B (93:52)

They have any idea where he is? I wonder how he'd hunt this guy down.

A (93:56)

And they're saying he's running one of the largest drug trafficking operation and is believed to be the dominant cocaine distributor in Canada. So from the Canadian Olympic team to the Canadian narcotics trafficking MVP status, Pam Bondi. The hard thing is when they talk about because drug kingpins have just been the stars of so many different great television TV shows and movies. Pam Bondi said he's the modern day iteration of Pablo Escobar.

B (94:38)

Somebody was about to turn him in, but they hear that and they're like, oh, he's too cool to turn in. Can't do it. It's not worth it. I just gotta be friends with this guy. Living through history.

A (94:48)

Oh, wow. So after the 2002 Winter Olympics, he moved back to Vancouver and attended university. He got into bodybuilding and started working as a bouncer. After two years in university, he dropped out and began to speculate in real estate, which he financed by growing marijuana in a massive warehouse. He eventually expanded his operation and joined up with Iranian and Russian smugglers. And in 2010, he was convicted of attempting to buy drugs from a U.S. government agent. And he went to prison for it.

B (95:23)

There are so many startups these days that are going after Palantir's market. Competing in enterprise software, enterprise AI, forward deployed engineers. It's all a bunch of memes. Everyone's, everyone's keying off of the Palantir strategy. I think if you're trying to play in that market, you gotta take down a most wanted fugitive. You get the 15 million, that's your seed round and you use this as marketing the company that finds this guy. I'm gonna believe that your model's pretty good. Whatever you're doing, if you can find this guy, I'm gonna be like, yeah, yeah, that's a good. That's enough of an eval for me. You pass the benchmarks. I don't care. Your mmlu. Tell me, how many fugitives have you checked off the FBI Top 10 Most Wanted list? That's what I want to see.

A (96:12)

Well, there's a documentary series in the making titled Snow From Olympian to Narco.

B (96:18)

Okay. Yes.

A (96:18)

Based on a Rolling Stone investigation.

B (96:21)

Okay. Jack Altman says, I heard the term forward deployed VC meaning investor who's helping a company. And now I'm logging off for the rest of the year. What's funny is like there used to actually be forward deployed VCs. What's that? Genentech story. Right. Who was the VC in Genentech? It was. Who was it? I don't remember. Robert Swanson, a venture capitalist, met with biochemist Herbert Boyer in 1976. There was, I think it was either Sequoia or Kleiner Perkins, but one of.

C (96:59)

Bob Swanson I believe was at Kleiner Perkins.

B (97:03)

Okay. Is that right? There was one example of a VC who basically was in the office of the company every Friday and was basically just like their financial back office the entire time because the team was a bunch of scientists who understood the technology but did not really understand how to build the business. And the forward deployed VC did really go in. That's a lost art because we are in the era of like founder friendly. Let the engineer figure out everything. Let them hire like kind of hands off from the venture capital community. So I'm actually okay with the term forward deployed VC if it's actually forward deploying. Like you can't just be like helpful. You need to be like actually in.

A (97:47)

The office show who's like basically like probably doing like actually 20 hours a week for one of his portfolio companies and has been for years.

B (97:59)

I would give him the Ford deployed.

A (98:01)

Vc but he would never call himself.

B (98:04)

Yeah, it'd be a very bad matters. How about for deployed Privy? Because Privy helps you build on crypto rails securely, spin up white label wallets, sign transactions and integrate on chain infrastructure all through one simple API. Peter Reinhart is in the news. This article might be a little bit too long, so we might tell you to go read it. But Peter Reinhardt says my lobbyists are very nervous about me posting this. But overregulation is working against us all. The costs are astronomical to us all.

A (98:33)

But hidden there's one excerpt that I'll read.

B (98:37)

Yes, this is the CEO of Charm Industrial and he has been going viral with a new, new article about overregulation.

A (98:46)

And so he's giving an example. So as one example, one state agency asked Revoie to do certified engine testing to prove that the Revoie doesn't increase emissions of semi trucks and that Revoie must do the certification across every single truck engine family. It costs $100,000 per certification and there are more than 207, 270 engine families for the nine engines that our initial partners use. That's 27 million for this one regulatory item. And keep in mind that this is to certify that a device whose sole reason for existence is to cut pollution by 90% and which has done so across nearly 100,000 miles of testing and operations is not increasing the emissions of the truck. It is a complete waste of money for everyone.

B (99:36)

So the Revoie system attaches to the truck and adds as an extra battery pack. It's a way to retrofit a gas powered truck to include an extra electric engine. There's something where I guess if your claim is that you're reducing emissions, I don't have that much of a problem of you trying to certify that, that you're not increasing emissions because that's the base level for your claim. Like your marketing claim. You're claiming that you're reducing emissions. If you're increasing emissions, then you're doing the opposite of your claim. So there is something about like whatever claim you're making, you should be able to back it up. But the main thing is that $27.

A (100:19)

Million, any type of system that requires you to spend $27 million just to get in the game.

B (100:25)

Yeah.

A (100:25)

When you're at, you would think that.

B (100:28)

There'D be a better way to understand the actual impact of this. I mean, there's so many maybe more efficient ways to deal with this. Whether it's just like, okay, how widely deployed are these? If they're making $100 million a year, take the taxes, do some independent analysis, pay for it yourself, and then decide, okay.

A (100:50)

Also the question in this case is why does it cost $100,000 per certification? Is there like a monopoly in the certification market?

B (100:57)

Absolutely. It's the government. The government's the one that does the certification. But yeah, well.

A (101:01)

Oh, oh, I thought it was an external. I thought it was an external partner maybe, but even then it's hard to imagine that there's like $100,000 of sort of services that goes into this. You should be able to cut that cost pretty dramatically.

B (101:17)

Before we move on, let me tell you about adquick.com out of home advertising made easy measurable plan buy and measure out of home with precision. Casey Hanmer has a big post here. Should we run through some of this or. Oh, you want to skip down to this one? Okay, here we go.

A (101:32)

Got another one off Cockroach the private credit cockroaches have been scuttling around this time. Private lenders are swapping debt for equity in distressed 48. I am logging into Bloomberg so that I can actually read the article.

B (101:56)

4840 private credit lenders are poised to take control of Pallet Co. 4840 Solutions about a year after providing $1.75 billion of debt to support Summit Partners acquisition of the business, a group of lenders including Antares Capital are set to take equity in 4840 under a proposed restructuring. Other lenders include KKR BlackRock, Carlisle Group Group. The deal would shed 1 billion of liabilities from 4008-40s balance sheet. In cases where borrowers were struggling with their liabilities, private credit firms are more frequently swapping their debt positions for equity stakes to try and prevent losses. While swaps can produce upside, they also offer more unpredictable returns than direct loans. Pretty crazy to be set up to get a payment for the debt and then be like yep, actually we're just riding with you on the equity side. But it's happening more and more because it's better than the alternative of watching the company go out of business. And so this was originally the cockroach phrase that this poster is riffing off of is of course from Jamie Dimon who was identifying first brands as the cockroach and was saying usually when you see one there's more, there are more cockroaches. That was his whole point. And, and the Blue Owl CEO co CEO Executive team kind of fired back and said like yeah, maybe at J.P. morgan we're good. And so they've been kind of duking it out going back and forth.

A (103:28)

Kieran here is highlighting another fairly concerning conversation with a buddy who has a pretty scaled company in the consumer packaged goods manufacturing space. And he was saying that around a number of the people that he competes with in his category have taken on so much debt from various private credit players over the last five, six years. And a lot of that, at least a number of them he thinks are just not going to make it out of the next two years without a restructuring. So we will see. I'll let you read through this post from Will and then I'll be right back.

B (104:13)

Will Meniatis is on a generational run of Long posts He says on Becoming legible to Capital the most underrated factor in the success of firms is the degree to which they are legible to capital. Defining this precisely is hard, you know when you see it. But an idea, a firm or a person is legible to capital, which he spells with a capital C when by dint of their existence, capital forms behind them in excess. Think of an idea being legible to capital as being magnetically charged to the capital markets. Dollars are just attracted to you. Allocators discuss it at cocktail parties, funds market it at their AGMs and Twitter discusses novel financing schemes to get more dollars behind the idea. Becoming legible to capital is the single greatest superpower for a fledgling firm. Look at Ramp, look at Cognition, look at Leopold, look at Tony and Dan at Danny. These are all superhumans at being legible to capital. It's worth noticing how ideas and people become legible to capital. First, there's no divide between the CEO, management and the idea. The story is not that of an individual successfully and impressive of that of an individual individually successful and impressive individual bringing their talents to a new platform. You can see these kind of puff piece stories in many companies that are illegible to capital. The company is never about you. For a founder to be legible to capital, there can be no error between them and the expression of the idea. Their lives only matter insofar as they are building with intensity towards the pure form of the idea. Second, the company can be understood fundamentally as an equation trade. In the best case, the most legible companies grow as a super linear function of dollars in the founders are highly verbal and can clearly articulate and understand the inputs to this growth equation. Talent, capital management, etc. The companies that are most legible to capital and the management that runs them at their best feel like clockwork toys in the hand because by looking at them the levers of progress and output become immediately clear. They are immediate smack you in the face expressions of an idea that is both small enough that you can feel the thing click around in your hand, but cosmically large enough that the feel that the idea feels true in some sense beyond literal. When the trade is immensely clear, every capital provider across the stack can immediately understand their role today, but more importantly their ability to scale to putting 10100 1000x the number of dollars into the thing as it reaches maturity. Finally, a company becomes legible to capital when it is the entire firm from capital to management to talent is singing from the same song sheet. They are all living their lives in expression of the same exact idea, an idea I've stolen shamelessly from, is from Phineas Barnes, Finn Barnes, who is over at First Round. Now he's at the gp is the chocolate cake problem. Many firms have great inputs. Eggs, chocolate, flour, some beautiful icing. But management thinks they're cooking a souffle, investors think they're getting cupcakes, and talent thinks they're getting pound cake. The companies that are most legible to capital are obsessive about chasing the tolerances between competing ideas of the firm to zero. By becoming legible to capital, a firm not only can achieve singularity in the private capital markets, that is unlimited free capital forever, but more importantly, it can unify the secret that ignites it internally with the world outside. When that secret is shared, there is truly no limit. Jordy, what's your review?

A (107:53)

I think of this in the context of when you think about El Segundo specifically and watching companies become more legible to capital over time and the momentum that starts to build where early on at the seed stage, there's just excitement around an idea in a team, and then you watch as these companies like basically bifurcate into. Sometimes you can have an incredibly talented founder that really stops accumulating capital because the sort of potential still remains unclear and they're not building that momentum. Whereas another one that has the same quality of founder starts to build that legibility.

B (108:33)

Yeah, I think. Yeah.

C (108:35)

Tyler, there's something kind of interesting. We've been talking about this a little bit where I think Jordy basically said this thing where it's like maybe the role of the CEO is actually to separate the valuation from what it kind of really should be. So it's almost like it's less legible. Right. Where if you're in a sector like Quantum or something, and it's kind of very. Actually it's quite hard to understand what the fundamentals are. That's actually where you get higher multiples.

B (109:00)

Interesting.

C (109:01)

So maybe that's like kind of against what Will's saying.

B (109:07)

I don't know. Yes. How does the meme stock craze fit into that? The meme stock craze is legible to a form of capital. It is legible to retail capital. It is legible in the sense that the retail stock trader, the meme stock trader understands how the stock will move on the basis of press releases, on the basis of viral moments, CEO comments, enthusiasm. If you are in a meme stock army, the legibility is derived from the power of the community, the apes, strong together. Well, how many Apes are there and how strong are they? If you notice that your apes are growing roaches.

A (109:56)

Strong together, strong together.

B (110:00)

As you, as you grow your ape army and you get stronger and stronger, that's the nature of the legibility. When I think of the examples that he's using, I think of it almost more like just the checklist of questions about a business and the inputs and outputs with regard to the capital. But also once the TAM is so well established, there's no more questions to about it. It's like, well, of course this is a trillion dollar industry. If things execute well, well, of course the team is the best one in the market. We've tested every other team, we've tested this idea and we've seen, we have full coverage of the team. There's no incumbent that's coming behind them. We know we have the leading team and we know the market's huge. And then also, are we sure that the gross margins are great forever? Are we sure that this is monopolistic or not super competitive? Are we sure that Google wants won't copy this and successfully compete in that space? So just getting super high confidence about whatever framework you want to use, whether it's all the questions in zero to one or five forces or whatever business, whatever spreadsheet or slide deck you want to see at your firm when you're making an investment, once you've checked that off with a company, it's so many times that everyone's like, yes, this is the winner, they're pulling away, they're doing great, they're a compounder. The legibility capital thing is important because you don't have earnings, you don't, you're not in the public markets yet. And so you need to become this like, foregone conclusion and then run from 1 billion to 50 billion in the private markets. And if you're, if you're illegible and there's still this question about, oh well, like when are you shipping or what are you doing? How does, how do I put my money in and get my money out? That, and it's more of like a moonshot. It gets a lot harder to just stack up.

A (111:56)

Yeah. Even at the early stages, it's still underappreciated that people say like, I have zero faith in this idea, but I'll still back the founder just because they're great.

B (112:10)

There is one way to become more legible to capital. Go to getbezel.com, shop over 26,500 luxury watches fully authenticated in house by Bezel's team of experts. Get a hitter, put it on your wrist. When you walk into that asset management firm, they're going to whip out their checkbook. You know it's going to work. You know, it's a secret hack.

A (112:31)

Should we bring in our first.

B (112:33)

Let's do it.

A (112:34)

Guest of the show, we have Julia Steinberg, a legend.

B (112:38)

Welcome to the show from Arena Magazine.

A (112:41)

She's back.

B (112:44)

Thank you so much for stopping by. Thanks so much for coming down.

E (112:48)

Nice to see that the studio is real and it's not AI generated.

B (112:52)

It is real. It is real. We are here in Hollywood bringing media back to Hollywood. We appreciate you.

E (112:58)

Are you the only ones left?

A (113:00)

We really are. We're in a huge complex, and to our knowledge, not a single other firm actually shows up every day.

B (113:10)

I do have. There is a one benefit. There are many benefits to being in Hollywood. One is that it's in between where Jordi and I live, so it splits the commute. The other benefit is that I think it should dramatically increase our chances of winning a star on the Hollywood Walk of Fame. Because if there it is, a lifetime achievement award, you're truly in incredibly rarefied air. Frank Sinatra is there. Or Leonardo DiCaprio. As a star, you truly have to be at the pinnacle of your craft. But if there is some bonus point for actually being in Hollywood, I think we increase our chances dramatically.

A (113:49)

Lifetime achievement award of Animal Analogies applied to Artificial Intelligence.

E (113:56)

I wonder if they'll just start putting influencers on there. Like, where is Alex Earle? Why is she not on the Hollywood Walk of Fame? Because that's what Hollywood is at this point.

A (114:05)

Portnoy made the most people angry of anyone ever.

B (114:10)

Yeah, Portnoy is. I mean, as you walk down the Hollywood Walk of Fame, like, yes, there are some absolute legends, but there are some other folks who are, you know, mid tier. Yeah, I would say mid. And so I would put influencers there. The weird thing is that the influencers aren't in Hollywood. So the question is, like, does the Hollywood Walk of Fame just wind up reflecting what's happening in.

A (114:30)

Is it geofence? Are the star. Are the stars?

B (114:33)

Like, When. When does MrBeast become eligible? Because he's in North Carolina.

E (114:37)

Well, when are they gonna do a Saudi Arabia?

B (114:39)

Oh, yes.

E (114:40)

That's what I'm looking for.

B (114:41)

Don't give them the idea. They will definitely do one over there.

E (114:44)

They should hire an open.

B (114:46)

I bet the. I bet Hollywood as a. Is Hollywood the city? I don't know, but I don't think.

E (114:50)

It'S an independent city.

B (114:51)

I think that's Hollywood is. I think Hollywood. There's obviously folks here that want to raise money. They should franchise the name. They should sell the rights to the name Hollywood and put a Hollywood in Saudi Arabia.

A (115:02)

Saudi Arabia should just recreate every city in the world in the desert and then use it, and then use it to incentivize movies to get made there. Yeah, well, bring culture to the desert.

E (115:14)

That's what LA did.

A (115:16)

Yeah, yeah. No, we've been to New York on the Fox lot here in la. They have basically a mini New York.

B (115:24)

Full representation of New York. You can go film whatever you want there. Anyway, give us an update since you've been on the show a few months ago. What's latest in your world? Are you full time at arena magazine?

E (115:38)

I am full time at arena magazine. We're getting ready to launch our upcoming issue 006, an ode to capitalism. And that's all I can say. More will be out next week, so keep your timelines open. I remember I was talking with Max, I was like, oh, is like Thanksgiving week? Is that a good week to launch? Will more people be online or will fewer people be online because they'll be with their families? And we had this whole debate about how online people were going to be during Thanksgiving week. And the answer that we came down to is actually, people are going to be more online during Thanksgiving week, I think. So they don't want to be responsible for getting the turkey in or out of the oven. They don't want to argue with their family about Epstein files or whatever.

B (116:19)

So, yeah, I wonder what the online behavior is, because over Christmas, like, we noticed that that's when like, that really crazy H1B dust up happened.

E (116:28)

Well, I'm Jewish, I don't celebrate Christmas. So I was, I had, you know, I had plans. I had to go get Chinese food, I had to go to movies. And I was just totally offline. And all my Christian friends are my friends who celebrate Christmas. They're hyper aware of the H1B discourse they could recite. You know, is this person.

B (116:43)

It was very odd timing because a lot of the tech podcasts were off for the holidays and didn't need to comment on it, even though it was this, like, hot political issue that you would have expected most tech podcasts to touch on. But then by the time everyone got back from New Year's, like, the whole. The whole, like, you know, dust up had sort of moved on, so.

E (117:01)

Well, maybe that's good because the H1B issue is just, you shouldn't touch it. It's too controversial if you want to be monetized still.

B (117:07)

Yes, yes.

E (117:07)

So maybe that was some very clever Jewish marketing gimmick by putting the H1B discourse on Christmas so only the Jews could participate.

B (117:14)

Maybe. No, I know. So I do think everyone participates in the timeline over holidays. I think it's. But maybe it's more phone time, less computer time. Like people are milling about, they're stepping away from the turkey, but they're not necessarily sitting down and writing.

A (117:32)

More good screen, less bad screen.

B (117:34)

Wait, which one's the good screen?

A (117:36)

That's the joke. It's like I close my bad screen.

B (117:39)

The bad screen is the computer.

E (117:40)

Yeah, yeah, I think my good screen is the computer. My bad screen is my phone.

B (117:44)

I would agree with this on my phone, actually.

A (117:47)

I mean, you guys are not working an office job that you hate.

B (117:52)

Yeah.

E (117:52)

Yes.

B (117:53)

Yeah. I think there will be people milling about reading post, but they won't have time to sit down and really we were debating this over the stream. Like will anyone have time to watch like a three hour stream on Thanksgiving?

A (118:05)

Of course.

B (118:05)

We're taking Thanksgiving off. But not Black Friday. But we are. We are going to be live.

E (118:10)

Well, that's the most important, Dave.

A (118:11)

We're going to be open for business. We're going to be celebrating.

E (118:13)

Tell me what deals. Tell me what to buy on Amazon.

B (118:16)

Yeah, we will. We have some good people lined up. So.

D (118:19)

Yeah.

B (118:21)

What'd you write about last magazine and did the California Forever piece line up with. It's going to publish in the next magazine.

E (118:30)

Yeah. So in the last magazine I did the review of Dan Wong's Breakneck.

B (118:34)

That's right. We talked to him too, a little bit. What was your takeaway?

E (118:37)

I think I really like Dan. I think he's a really smart guy. I've hung out with him a few times. He gave me a huge stack of the London Review of Books and he's like, this is just like the greatest thing ever. So it was nice bonding over print media with him.

B (118:48)

That's great.

E (118:49)

But I do think that it's just like I'm very skeptical that the United States government should get involved in these huge infrastructure projects in its current state. I like the idea of engineering, but what makes American engineering so spectacular and so amazing is that we do it in the private market.

B (119:06)

It's distributed.

E (119:06)

It's distributed. And if we just said, okay, the United States government is going to do this huge engineering project, it's like, do we trust the United States government to do this? Huge engineering project. I like, I drove down to LA yesterday from San Francisco and it's like half the roads were terrible and it's like, you know what? I sort of wish this whole thing was a toll road. Yeah, maybe I'm not like that, like die hard libertarian where I'm like, government shouldn't do roads. At least not yet. Give me like five years.

A (119:33)

I joke that the G wagon is so popular in LA because you actually need a 4x4 to like get around some of the.

E (119:40)

Well, yeah, well, it was raining like crazy yesterday on the freeway and I was driving down, I was like, thank God that my beautiful, beautiful RAV4 has four wheel drive or else I would die because these roads are terrible and no one knows how to drive in the rain.

B (119:52)

Yeah, I was reflecting on. There are so many interesting anecdotes about like, is the government equipped to take on the project that you are interested in? Whatever your pet project is, whether it's healthcare or space exploration or whatever you want. And yeah, I keep coming back to this. Part of step one in any of that is fix the productivity problem in government projects.

E (120:19)

Yeah. I think for the next issue of Arena 006, it's coming out December 1st, I wrote about California Forever, which is the project to build a new city in Solano county, which is about an hour northeast of San Francisco. And it's sort of like an interesting concept to me. It's like, oh, it's this privately funded city. You know, all of the, like the who's who of Silicon Valley, Laureen Powell, Jobs, the Collison brothers, Reid Hoffman, Nat Friedman, Mark Andreessen, all these people put money into this. It's a privately funded city. But so much of the struggle with getting California Forever off the ground is just because of government. And I remember I was fact checking the essay. I was communicating with Julia Blystone, who's the California Forever head of comms. And I said like, okay, well the next step is there's this 1200 page environmental review. And she's like, no, it's actually 12,000 pages. I don't think I've read 12,000 pages in my life. I don't think anyone in this room has read 12,000. And I read a lot. Maybe I have in my life, but not at once. And definitely not Environmental review report.

A (121:21)

Yeah, especially a report that somebody is just going to drag into ChatGPT and say, turn this into five bullet points.

B (121:28)

Points. Please turn this into one image.

E (121:32)

My favorite memory from college actually is this is actually when Peter Thiel Was co teaching a class my sophomore year, German 266. And it's a class of like 60 people. You have to fill out a form to get in. It's supposed to be a good class and it is a very good class. It's a fascinating topic. But my favorite or one of my favorite memories from the class is the kid next to me who had to start at that was worth like millions if not tens of millions of dollars. Put in ChatGPT Summarize Hegel in three bullet points.

A (122:04)

One shot it.

E (122:05)

Yes.

A (122:05)

Don't make mistakes.

B (122:06)

That would be. Yeah, that's.

A (122:08)

Well maybe I mean Stanford whiz kid, his startup's only worth a couple million bucks. Maybe he's actually not that. Not that elite.

B (122:18)

Yeah. What is the. What is the mood in the Stanford alumni world right now? There was this question I was sending you this post from Deli and this question about like the age old question of like is this next generation cooked? Is there is AI actually turning off folks brains to the point where they can't sit down and read Hegel all the way through that everyone is doing this. Is it a systemic problem or is this just cantankerous old millennials shaking their fist at the younger generation?

E (122:49)

Well, I will say this. At Stanford there's a thing called coterming. It's like a fifth year Master's. You apply most of the time, people get in and you just sort of stick around for another year, finish up a master's and you get two degrees at the end in five years. Apparently a third of my class, the class of 2025, co termed or is coterming right now. A third. 33%.

B (123:09)

Yeah, we've seen this a lot. We saw some data that more people are applying. Applying to law school.

E (123:12)

Oh yeah, the law school thing is terrible. I have a few friends who are studying for LSAT right now and I'm like hitting them in the head with their huge LSAT books.

B (123:19)

Is that because you think they won't be a. It won't be a career path.

E (123:22)

There aren't going to be associates. Maybe they're going to be like five associates, but there aren't going to be 50 associates.

B (123:27)

Sure, sure.

A (123:29)

It's interesting because if you genuinely are excited about devoting your life to understanding and finessing the legal system, it's probably as good of a time as ever because you're going to be able to learn a lot faster. You're going to be able to be more productive. Like if you're super high agency, you'll find the right firm and attach your career to that and probably be able to have a wonderful career. But deciding to do law school because you don't know what you want to do and you trying to buy time and you want to do something that's sort of like high, like on a high status track and hopefully high earning without being necessarily high agency just feels like a train wreck, an inevitable train wreck.

E (124:14)

Well, how it seems to me my dad actually he just retired from a legal career and he was the first person to say like, yeah, all the associates are sort of screwed because they just can't out compete the AI and law, especially how it's structured because there needs to be a partner on the case. As per this weird agreement that was reached I think in the early 2000s, associates can't really be on cases anymore. So there's even less of a need for associates. And this is going to be true in all fields. And I think obviously in law specifically there are these 100, 200 year old firms that really care about legacy, they care about reputation. So they're like, ok, we need to have some young people to carry the firm on. But it's just a question of how are those people going to learn anything, anything if ChatGPT is just doing all their work or Harvey or anything like that? I think how I like. I think if you're like the top like 5% of law students, you're probably gonna be fine. Yeah, we're still gonna have judges. Judges are one thing that I'm like, maybe AI should just replace judges. We have a lot of bad judges.

B (125:18)

That's a crazy idea. Before of the courtroom, the thing to replace you would think would be the stenographer. Like we have dictation and yet or.

A (125:29)

Else I don't know, I use otter.

E (125:31)

AI all the time.

B (125:32)

It's not that great. Is that great? Okay, we can get into the bear thesis for AI in a minute. But the illustrator, what are they called?

E (125:41)

Just court Courtroom illustrators.

B (125:43)

Courtroom illustrator. We have cameras.

A (125:45)

Why don't we just take a picture? The downside of if you actually use nanobanana or any other image model, it would, it would not, it would just be like good enough to not produce a funny result. And I feel like there's so much humor.

E (125:57)

It has to be like Studio Ghibli, like murderer or murderer like defendants.

A (126:02)

No, no, that would work. But I'm saying like sometimes the illustrators like botch an illustration so badly that it's funny.

B (126:09)

That's what you're thinking About. Yeah, yeah. I was shocked by how resilient that job was to automation. Since we do have the camera and we've had the camera for a hundred years almost and it's stuck around as a. I think.

E (126:22)

I don't like a lot of court or a lot of law is like ritual. Like in the uk, the judges still have like the weird head things that they need to wear the wigs or even like the judges robes. My dad, like he had to wear.

B (126:36)

A tie around for a really long time. Even if the AI can do so much, it just. We will do it out of procedure. You, we will have these conversations with humans for whatever reason. I don't know.

E (126:49)

Yeah, procedure is very inelastic and also I think just, you know, lawyers, it's a very easy cast of person to demonize. And I think just as McKinsey will stick around so they can say, oh, sorry about the layoffs, McKinsey made us do it. It's gonna be the same with the evil lawyers.

B (127:04)

Yeah, sure. Yeah. You want the buck to stop with someone and not necessarily an AI.

E (127:07)

No, not necessarily. Because if you say, oh, sorry, the AI said this will, they'll say like, well, you used Grok instead of ChatGPT. Can you ask ChatGPT? Because that's my preferred AI model. If it's just McKinsey.

B (127:18)

Yeah. What's the. I feel like there's still probably some alpha in becoming a lawyer or going down the legal path in 2025. If you are basically setting yourself up not to just be associate fodder for high margin billing, but in fact you are setting yourself up to gain control over an institution that would be harder to build independently.

E (127:47)

Well, I think with young people, especially law, it's because you have, you know, you make like good money in your like late 20s and 30s, but you start to make really good money in your 50s and 60s and I think.

A (127:59)

For sort of like an MLM.

E (128:02)

Yeah, a little bit. But it's also for people my age who it's like, oh, I could either like work really hard for like 30 years and maybe if I make partner and I'm one of the really good partners, I can make a ton of money. Like a good amount of money. It's like, why do that when it's like, oh, if I learn how to code and after like 10 years I can make this much money.

B (128:19)

Yep. 4 years. With vesting.

E (128:22)

Yeah, with vesting. All you need is the four years.

B (128:24)

Yes. @ the right hot company at the right time.

E (128:26)

Yeah. So it's Interesting to me because I feel like with the Gen Z financial incentives, it's so oriented around, like, make money now or even like the, oh, you have X amount of years to make money before AGI hits. And the economy is permanent underclass. Yeah, permanent underclass.

A (128:42)

You're still waiting for the gambling on your student loan debt, the double or not, like double your debt or pay nothing.

B (128:51)

I've become very skeptical of the permanency of the underclass phenomenon. I feel like there will be. The strength of the underclass is growing in number, but the permanent permanency is decreasing because at any moment someone can be plucked from the underclass into the global elite through just one. Right. Like parlay or one. Right. Actually like building a company. Joining a thing like distribution on the Internet is so, so aggressive that you can be languishing in the underclass and then go viral and launch a D2C line. And if you just have like a few things go your way, you become like the market just gives you $10 million. Basically.

E (129:32)

Yeah. I think it's very rare. It's random too.

B (129:34)

It's random.

E (129:35)

Yes, yes.

B (129:35)

Yeah, yeah. It's like you were on that plane and you screamed at that person and then you went viral and you became.

E (129:40)

And then you started a podcast.

B (129:43)

Exactly.

E (129:44)

Yeah.

A (129:44)

Don't give people ideas. I think it usually goes the other way.

B (129:47)

Like lately screaming at people. On example, I don't exactly know how it all played out for her, but.

E (129:51)

Hayley Welch, like, well, she was stupid, but it.

B (129:55)

Yeah, I think she, like, she launched a coin and it did not execute well. But that type of opportunity is entirely random. And every once in a while the market just comes around. It does not feel permanent to be in the undercloud.

E (130:08)

So when I start yelling. So on Wednesday when I fly to see my family for Thanksgiving, when I start like really aggressively yelling at people on the plane, you know, it's time to buy Julia coin.

B (130:17)

Yes, yes, yes. You got to get in early on that.

E (130:20)

Sure. You have to get in early before I have my tantrum on the plane.

B (130:23)

Yeah, yeah. What else is. Has anything updated you on the get your bag culture? Are people now self aware of it? We were talking to Jessica Livingston and she was referring to the concept being actually not very new. She was saying carpetbaggers.

E (130:39)

Well, yeah, that was like a term for the Civil War and it was like anti northerner. These people are just exploiting business opportunities, which I don't know. It's like, I don't like the term exploitative business opportunities. I think we have enough regulations that it's like our. How true is that? In the United States, at least. But I don't know. Something I've been thinking a lot about is. So we're in LA right now. I went to high school a few miles from here, and I think, like, a very core memory for me was a college admission.

B (131:07)

Core memory. Shout out. Ashley Vance.

E (131:08)

Yes. The college admissions scandal. Varsity blues.

B (131:13)

Yes. Oh, yeah.

E (131:13)

I was a junior in high school when this happened.

B (131:15)

Okay, so you experienced.

E (131:16)

So my school, actually, the whistleblower's daughter was a senior, and she was gonna go to Yale for soccer. And we're all like, oh, it's sort of weird that this girl. I'm gonna call her Susie. Susie. Is gonna go to Yale for soccer, but she's not on the soccer team at her school. And there's not really evidence that she plays soccer, but she's gonna go to play soccer at Yale, One of the.

A (131:34)

Most elite teams in women's soccer.

E (131:37)

Sure. I don't know anything about soccer, but this girl was gonna go to Yale.

A (131:40)

Then.

E (131:41)

You know, my high school was very, like, elite. Everyone was concerned with if they got into Yale or not. So it was all sort of like, oh, like, this is important. Like, everyone flagged that Suzy's going to Yale, but we all sort of knew that she didn't play soccer. And I remember my high school was very small and, you know, it was a college prep school. And the day before the news came out, I had a meeting with the head of school to talk about, like, where I should apply to college, what my SAT scores was, blah, blah, blah, privilege, blah, blah, blah. I acknowledge that I had a very elite upbringing, but I remember the principal was late, and this never happened. There seemed to be something very wrong. And I was like, I don't know what's happening. And she was late to the meeting, and it was fine. And I was told that my SAT scores were fine. Blah, blah, blah, blah. The next day, first period, I have a moral philosophy class. I'm the only junior in the class. It's all seniors. I petitioned to say I can take this electricity. And we're talking about meritocracy and whether meritocracy is real. Yes, in the class, we have a reading about meritocracy. Is it meritocracy? Just. Is it something that we should optimize for? Is it something that's unjust because, you know, intelligence is heritable, so it's not earned, Blah, blah, blah, blah, blah. And we get to the class and, you know, at 7:00am that morning or whenever we all woke up, we all saw that this girl's dad was, you know, he was arrested for securities fraud. And part of his plea deal was that he was giving up this college admissions scandal where this guy, this college admissions coach was telling people to put that they were black on their college apps, that they were faking associations with various sports teams, all these elite schools. So we go in and it's like, okay, we're talking about meritocracy. What about Susie? What about the college admissions scandal? Because we're all sort of fed up with the whole college admissions process. It's a dog and pony show. My little brother's going and get through it right now. And I'm like, why are they asking you what three words you are? What could this possibly have to do with how good of a student you will be or anything like this? But I remember going through the college admissions scandal firsthand, just seeing how corrupt it is. But also in a way, it was sort of like an equalizing form of corruptness. Because if your family can donate $50 million to Harvard, you will get it. You have to be really, really, really, really bad to not get in. If your family donates 50 million. What if your family doesn't have 50 million to donate to Harvard? What if they only have 500,000?

B (133:55)

Well then you have to make 50 million before you apply.

E (133:58)

Or you can use a $5,000 to buy your spot on the sports team.

B (134:01)

Oh yeah, five grand.

E (134:03)

500.

B (134:04)

Oh, 500.

E (134:04)

But that's a lot less than 50 million. Yeah, so it's a sort of more equalized. So it was for, it wasn't like the like 0.1%, but it's still like 1%.

B (134:12)

Sure, sure, sure.

E (134:13)

Corruption. So I think that experience.

B (134:16)

Isn't there a pro libertarian hotel take that says like you should just auction off every slot.

E (134:20)

I've said this before. Like I, well my, my sort of take on this is that college admissions office should just be sold to the highest bidder.

B (134:27)

Yeah.

E (134:28)

It would be a lot more honest than what's done now. Because right now it's. They sort of, they give you the spiel about like, oh, we need to have a super diverse class where we have one cellist and one person who's really good at like studying molecular bio and like.

A (134:41)

And a sailor. You have to have a sailor.

B (134:43)

But Michael said sailor.

E (134:44)

We actually don't have enough sailors. They're, they're doing walk ons in the sailing team at Stanford. My best friend in undergrad did that. But yeah, I think we should just auction off the college admissions office.

B (134:53)

Have enough sailors in this country anymore.

E (134:55)

We don't have enough sailors in this country.

A (134:56)

So bearish.

B (134:57)

Yeah, we got to get back into sailing. That's hilarious.

E (135:01)

But yeah, I think it's. This is actually something I was talking about with my boyfriend is like, how much would it cost if you just auction off a spot? Maybe if you only auctioned off 10, if you limited.

B (135:11)

Oh, I think you would, you would auction off literally every slot, all of them. Like, what is the market clearing price? Because you know, in the luxury watch world, like there will be a certain number of. I mean this happens with like the market. What is the market clearing price for Taylor Swift tickets? But like, you know, there are, if they, if Porsche comes out with a new, a new GT3RS, there might be, you know, 5,000 made. And they will try to find the market clearing price as close as possible to what the market will bear. Sometimes they overprice them, they don't sell them all. Sometimes they underprice them and they immediately trade up on the secondary market. But you could even have a secondary market, you know, okay, yes, I bought my slot at Harvard for 2 million, but they're going to 3 million now, so I'm going to sell that, go to NYU instead. You could have an entire entire process just to maximize the price.

E (136:00)

I just don't know. It's like if that happens, what percent of the Stanford class would be international? Which is a good question.

B (136:04)

It would be hugely international, of course. I mean it would be across the entire world. Everyone would just pay to go. And I mean, there's already a little bit of that. But it's definitely like there's more design to the system than that. There's certainly, I would imagine.

A (136:20)

How do you defend that against, you know, just eliminating an entire track for young people that maybe have. Of negative capital within their family systems. And even if you tell them, even if you tell the family at birth, like, you know, expect to be able to purchase, you know, if you want to go to, if you want your child to go to Sanford or the auction range at that point, counting in inflation will probably be, you know, 70 to 80 million to get into the school.

E (136:48)

Well, probably what we'll have is.

A (136:50)

And then it ultimately actually destroys. Would, would naturally just destroy. I think some of the. Would certainly be valuable for students from elite families to network with other students from elite families. But ultimately would kill a lot of what makes universities great.

E (137:11)

Yes, I agree that it would kill a Lot of that. But probably what would happen is now you have a lot of scholarships when you apply for college. The Coca Cola Scholarship. Scholarship is a very big one. A lot of these are sort of mostly like full rides to college, like the very prestigious scholarships. And so probably what will happen is that instead of having these scholarships to college, it's like, okay, we young people of America whose family can't afford the $50 million sticker price of college admission, apply, and we will give you the $50 million to pay your way for this price. But, yeah, it's obviously a very flawed system, but it's interesting. It's an interesting mental model about how much we value college admission. Especially because at a place like Stanford, there are what, like 1700-2000 graduates in a year? Not all of them do fascinating things.

B (137:59)

Yeah. What city are you living in now these days?

E (138:04)

I live in San Francisco.

B (138:05)

What is the mood in San Francisco? Do you think it's getting better?

E (138:08)

I don't know. I've only lived there since August, so I can't really say.

B (138:11)

There was this narrative in the. In the Mansion section today about it cleaning up its act.

E (138:18)

Okay, I have a fun story.

B (138:20)

There's more demand for West Portal Village Center. I had never heard of that.

E (138:25)

I don't know what that is.

B (138:26)

It is now more expensive to live there than in pac Heights, apparently. $2.5 million median list price. There's also the story about, like, the OpenAI secondaries driving up real estate prices. But it's always interesting to feel what's actually happening on the ground.

E (138:42)

So I live in Pac Heights. I live in lower Pac Heights. And a few weeks ago, there was a homeless guy who just totally passed out on my doorstep. It was a Saturday morning. I woke up late, like 9 or 10. I checked my phone, and in our apartment, WhatsApp, it's like, hello, there is an unhoused person on our doorstep. Please don't disturb him. And then in this group chat, we had this huge debate over whether we should call the cops. And the conclusion is just, the cops won't come. He can't really do anything about it. He seems nonviolent. Maybe we should wait for him to leave. So I left my apartment at, like, you know, 10:30. I was back at 3, and he was gone. But it's still like, a weird thing to account for.

B (139:22)

You sort of want to call an ambulance or something?

E (139:23)

You want to call. You just want to call someone?

B (139:25)

Yeah.

E (139:26)

And have my building, you know, they're probably like 30 people who live in the building. So it's like someone could have called someone or maybe if all 30 of us tried. But it was very disconcerting because it's one of those things where it's like, oh, everything's better.

B (139:41)

Everything was everything noticeably cleaner during Salesforce this year. Salesforce's conference, because a few years ago, Benioff just completely cleaned up the entire city. And then the other time that they cleaned up the city was when Xi Jinping.

E (139:55)

Was that something you remember? Yes, when Xi Jinping. I was in SF that week, and it felt like it was a different place.

B (140:00)

Really. That's funny.

E (140:01)

I think SF generally or urbanism generally, it's just. It's so hard to justify living in one of these places. I think for most people.

B (140:11)

I agree. But it's almost the flip side where it's so easy to justify living in a suburb because, yes, you're in the city and it feels like, oh, this is like, we gotta clean this place up. We gotta clean this place up. But then. Then you sort of work on that. And then you also work on your life and your company and your business and whatever. And then as soon as you have money and a family, it's just so easy to go to a suburb. And there's so many fantastic suburbs. And it was always the thing that annoyed me about the discourse of, like, San Francisco has fallen or San Francisco is such a bad place to live, is like, well, people live in Palo Alto. They live in Atherton. They live in. Up north in Wine county country. They live in Incline Village. They live all over the place. And they live in these wonderful, luxurious places that have no crime, no homelessness. They have fantastically managed cities just a commute distance from San Francisco.

E (141:00)

Yeah. I think San Francisco, it's always sort of been like a young person city.

B (141:03)

Totally. When I lived there, it was super rough, but I was like, I'm young. It's fine.

E (141:07)

Yeah. I saw a guy on Twitter who's going to give me a Taser, so.

B (141:10)

Thank you.

E (141:12)

But I think.

B (141:13)

I think weapons on the Internet.

E (141:15)

Yeah. Giving away weapons.

A (141:17)

You can buy a Taser on Amazon. It'll show up this afternoon.

E (141:19)

But it's like a Taser gun, so I think it'll be easier to use. I was gonna buy a Taser.

B (141:24)

Well, those are. Yeah, those are legal. You can also just buy this.

E (141:26)

Yeah, it's.

B (141:27)

But if someone's gonna give you a free one.

E (141:28)

Yeah, someone's gonna give you a free one. So SF is sort of. It's. It's, you know, the Duality of San Francisco is it's great because some random stranger on the Internet will give me a free taser gun. But the sort of negative part is that I need this taser gun to stay safe.

B (141:40)

Yeah. It's always been so fascinating. I wonder if the OpenAI wealth will meaningfully change the budget of the city. It's just so much. And it's actually. The company's actually in San Francisco. It's not a Mountain View company. It's not a Palo Alto company. It's not a Menlo park company. It is a San Francisco company. And you have $500 billion of wealth, loosely, with tens of billions in the pockets of employees right there that should wind up doing something, even if it's just some renovations, some private security.

E (142:14)

How good are OpenAI employees at spending money?

B (142:17)

I don't know. How would they spend it? How would you recommend this?

E (142:20)

I have a friend who works at OpenAI, and he was just like, you know, I don't really know how to spend my money. And I was just like, do you want, like, I can. Yacht.

B (142:28)

Yeah, yacht.

E (142:30)

I don't know.

B (142:31)

There is a marine in San Francisco.

E (142:32)

There is? Yeah.

B (142:33)

It's beautiful.

E (142:34)

I like to go and walk around there, and there are these, like, cool boats and yachts. It doesn't need to be, like, a super ostentatious one, but I don't know.

B (142:41)

Every yacht is secondary to cleaning up the streets, you know?

E (142:45)

Yes. But all of, like, even buying a yacht would help clean up the streets because it is some simulation of the city's economy.

B (142:50)

Okay.

E (142:51)

I don't know.

B (142:51)

I suppose. Yeah.

E (142:52)

Speaking of boats, it's like, waters high, tides rise, all boots.

B (142:58)

Sure, sure, sure.

E (143:00)

So I don't know. I guess it's sort of. It's silly to say, like, rise or, like, raise the economic output of San Francisco because it's one of the most economically, like, intensive regions in the country. But I do very strongly believe that, you know, just let market forces do their thing, things will get better.

B (143:16)

Well, thank you so much for stopping by the Ultra Dome. This is fantastic.

E (143:19)

Yes, thank you for having me.

A (143:20)

It's always great to get the Update excited for 006.

B (143:24)

006.

E (143:25)

Yeah. And 007 is next after that.

B (143:28)

Where can people get it?

E (143:29)

People can. Arenamag.com arenamag.com and @barnes and Nobles near you.

B (143:34)

Yes. There's a Barnes Noble deal now.

A (143:36)

There we go.

B (143:37)

Irl.

E (143:38)

So if you want to buy any Romantasy books, which I think is the new genre. Yeah.

B (143:43)

There's Romantic fantasy.

E (143:44)

Yes.

B (143:44)

Okay.

E (143:45)

That's the new genre that women.

B (143:46)

That's not my. It's not my type. I will see.

E (143:48)

You have to get on that.

B (143:49)

Well, thank you so much.

E (143:50)

Thank you so much for popping in.

A (143:51)

Great to see you.

B (143:51)

This is great. Great to see you.

A (143:52)

Cheers.

B (143:54)

Let me tell you about eight sleep.com exceptional sleep without exception. Fall asleep faster, Sleep deeper. Wake up energized. We have our next guest. Joining right now we have Bob Bushall from Dupe. This is a very fun company.

A (144:11)

What's happening? Welcome to the show.

D (144:14)

What's up? Thank you.

B (144:15)

We made it, guys. We made it.

A (144:17)

We made it. Finally. Great, great, great setup here. You got. You're just daily driving that for zoom.

D (144:24)

That's it.

B (144:25)

Yeah.

D (144:25)

I mean, it's comfortable, it's cozy.

B (144:28)

Introduce the company. I remember when you launched, I remember seeing a whole bunch of posts go viral. But how are you describing the company today?

D (144:36)

Yeah. So Dupe is used by millions of people to find the products that they want or similar products for less. It's a search engine, it's a generative marketplace. I could throw a whole bunch of buzzwords. It's AI driven, this and that. But ultimately it's a shopping companion. It's a shopping app that helps give people a confidence pill, a little bit of a boost to know that what they're buying is at the best price and is the right product.

B (145:04)

Agentic Commerce OpenAI is talking a big game about it. Is that a headwind for you? Yeah, you know, it's.

D (145:14)

First of all, nothing set in stone yet. Don't, don't count us out. We're also talking to them.

B (145:20)

Sure.

D (145:22)

We're building an app that would be native to ChatGPT.

C (145:27)

Okay.

D (145:27)

You know, it's like the thing I like to say about ChatGPT, Google, all these tools is that they can do everything. And that is the problem. It's like you can, if you're a chef, you can use a Swiss army knife to cut a tomato. But why would you. I think verticalized AI tools are very powerful because they are single purpose and they solve a problem in a more precise way.

B (145:53)

It's odd. I have like no doubt that there's people out there that are making tens of millions of dollars on Studio Ghibli apps. Right. Or, you know, different animation apps, different, different sub apps that maybe could be done in ChatGPT if you know the right prompt, but because it's lower tier on like the prompt ladder, you kind of got to know to go there to get it. It's not coaching you through it. It's not actually giving you all the different harnesses and wrapping that you need to actually do it.

A (146:24)

How hard has it been to actually build the product experience? Because I'm assuming you're effectively browsing the web on behalf of, of, you know, if I upload a product or I paste in a link and then you guys are browsing the web or you already have and you're. And you're sorting products to find products that look the same and, or maybe from a different manufacturer, the same end manufacturer, different brand. But how hard has that product experience been? Because I've been continuously disappointed by computer use and just web browsing from a lot of different LLMs. At this point.

D (147:02)

When you say computer use and you're talking about things like the browsers doing the browsing on your behalf, basically.

B (147:08)

Yeah, yeah, go out and find this thing. The example that he gives is I.

A (147:11)

Use it for cars.

B (147:12)

That's my find me a GT3RS. It's a rarer car, It's a specific car, It's a great car, but it's not like, find me someone who sells Diet Coke. Like, Diet Coke is a commodity. It's sold everywhere. A GT3R, there might not be one for sale right now. Like, it's possible that there's just none for sale. And so, yeah, maybe a few of.

A (147:35)

Them quite a bit more common, but there are certain cars you look for and it'll say, I found you this one and it's like it sold two years ago.

B (147:42)

Yeah. And so on that eval of like, go hunt around on the Internet. Find this particular thing that's not particularly fungible, not particularly commoditized, at least. Agent mode. These, agent. Agent. These, these agentic commerce products have not been up to Jordy's standards. At least.

D (148:01)

I take it you haven't bought the, the GT3RS yet.

A (148:04)

Not yet.

B (148:05)

Not yet. And the only reason is just the AI is not good enough.

D (148:09)

Yeah, of course, of course. That's the only reason. Come on, guys. So, so, so first of all, I think it's. I, I too have been fairly disappointed with, with these like, browsers that, that browse on your behalf. That is, is ultimately not what we're trying to do with dupe. AI has some flaws in like the shopping universe. And maybe that, you know, it speaks to some of the pain points you're finding with, you know, finding your perfect car, which is they don't have live inventory, they don't have live price, they don't have Product catalogs, they don't have customer preferences, they don't have customer behavior. There's a lot of data that is missing in the shopping journey. So ultimately, what you're left with is the best guess. So to answer your question, how hard has it been? It's hard because the way you do it, the way you do it right, is you got to do it the old school way. All right? In the shopping world, you got to go press palms, kiss babies, hug people, take them out to, you know, to a drink, meet them at conferences. You got to know their names, you got to know their kids names. It's like a lot of trust building in person. It's not as simple as just kind of plugging it in because when you build that trust trust, they then trust you with their product catalogs and they give you a commission for sales because, you know, you turn out to be a partner that drives a lot of sales for them. So it has been, it has been pretty hard. But the tech that we have is product ingestion from these partnerships, but also live scraping. So it's, you know, in a sense it is browsing and it's scraping this data for you. We have vectorized millions of products or tens of millions of products at this point, so we can do kind of a live assessment of what products look like each other, what products should go together, stuff like this. Ultimately, the way to solve it, AI is not sort of a panacea. You have to pick your battle. So Dupe, for example, for the last year and a half has been focused solely on furniture. Why furniture? Well, turns out unlike the GT3Rs, a lot of furniture is made in the same few factories, sold back to brands under different names. The brands then upload them to their website under different names, different prices, but it's the same furniture. And I know this because I got scammed with buying furniture online. And it really pissed me off because I found the same product I wanted for about 40% less. The exact same, same spec, same color, everything exactly the same. And so that ultimately is like the lore of what spawn dupe. And I've seen these factories in person. I've visited China.

B (150:32)

How do you think about incorporating human feedback that feels, that feels like something that would be a really valuable resource. I was thinking about, like on Instagram, you see a lot of ads for a bunch of different stuff. Some of the stuff is like, it really grabs your attention. People just buy it, but then it's really low quality. And if you could actually have some sort of, you know, okay, feedback. Hey, we shipped you this dupe, this product. We think we got you the lowest price. But like tell us how the product actually performs because that better informs our decision making. Do you think that'll be like a compounding advantage? Do you think that's valuable over the long term?

D (151:13)

It's so valuable and we don't even need to ask the customer. First of all, we have access to all these reviews. We also have access to the checkout data and we have access to the return data because if someone returns something, we don't get paid. So these brands pay us four months after the sales have happens, three months after the sales happen, after that return window is set. So we know which products get returned the most. So absolutely. That's a compounding effect.

B (151:37)

Yeah, that makes a ton of sense, Jordy.

A (151:40)

You get a lot of hate mail from certain brands that kind of rely like you're providing a service that consumers love. And then certain brands out there that are sourcing products and setting whatever price they feel like is relevant. And then you're actually.

B (151:57)

And blowing up my spot. I had 99% margin on this couch.

D (152:01)

Yeah, we basically like killed shop dropshippers for sure. Like the drop shipping industry is a, is a total scam and they hate us. But ultimately our brands, the part, our partners love us, we drive millions of dollars in sales for them every, every single month. And so why is that? Well, let's say like, let's take a partner like Walmart.

B (152:22)

Walmart.

D (152:22)

So some people think Walmart has thousands of products. They don't. They have millions of products. Right, Millions of products. They're not always in store because a lot of those products are drop shipped. So what we do is, you know, let's say you guys are looking for, we'll take a piece of furniture, for example. That chair right there, I mean that's, that's not being sold at Walmart. But let's say you want something like that and you upload an image to dupe or you tell dupe like, hey, that's a Herman Miller. Find me something like that for less. Walmart has 15 versions of that chair. You don't know that. You would probably never even think to look at Walmart for that chair. But Walmart has a great version of that chair and they love us because we merchandise their products at that exact high intent moment where the customer knows what they want. They just want a great deal on it. We surface the partners up. Now there are some brands, for example, we were sued last year and yeah, we were Sued. I'm not going to name the brand. It's very easy to find.

B (153:18)

Sure.

D (153:19)

The brand. They are the largest furniture brand. Now the lawsuit essentially like it tried to kill our company. It was, it was like a pretty tough, a pretty tough lawsuit that we, that we had to fight. We settled it to both parties. Content like it's fine now, but what.

A (153:38)

Was the, what was the claim?

D (153:41)

The. There are a couple claims. So copyright infringement. They were pissed that we were using or letting our customers use images from their website to find similar products. They were, they were pissed that like they were part of. We went super viral when we came out. I did all of these like post, post it note ad campaigns and stuff that went super. They were pissed that they were in some of those, some of those social kind of tick tock videos. They weren't even ads. They were literally just social content. Yeah, talking, talking about our website. They were really, really upset about that. And ultimately they were pissed that we were unfairly competing essentially. I obviously, you know, we, we have a public statement on how we felt about all of these claims and so they basically said we were falsely advertising, you know, that they were white labeling stuff that we, that we were unfairly competing and also infringing on their copyright. So ultimately, you know, cold water under the bridge. We're here.

B (154:36)

Dupe is alive.

D (154:37)

It's. Well, it's working exactly as we want it to and it's crushing. And so there are some of those partners, but those are very few and far between and it tends to be like the almost the monopoly type partners, you know, that, that are, that are upset. Everyone else, everyone else and most of the Fortune 500 that we work, they are thrilled that Dupe exists.

A (154:59)

Sense. How'd you get the domain? It's absolutely fire.

D (155:02)

We. So the company was called Carrot before it was called Dupe. We had three months of Runway before we pivoted the company to Dupe. We worked with Nikita Beer on it. He was a total goat.

A (155:13)

Helped us find like spent your last, your last 20 grand on intro.

D (155:19)

Well, we had, we had more than 20 grand. We had about three months of Runway. But yes, we definitely spent more than 20.

A (155:24)

A number.

D (155:25)

Yeah, that's with Nikita. Probably the best money, probably the best.

B (155:29)

Money I've ever spent.

D (155:30)

Saved our company. Let's be completely honest.

B (155:32)

That's actually crazy.

D (155:33)

I love it like total bananas moment for us. But the domain was a very last minute decision. We knew we needed a great name to stand out in the kind of shopping space. I did a bunch of Google Research on keywords. One of the words was dupe. And I saw dupe was on. On Google Trends 100 rated Google trend keyword and growing.

A (155:54)

Yeah.

D (155:54)

And I saw that no one used. No one was using dupe.com. we reached out to a broker that was sitting on that name. They wanted just the most insane price. We then negotiated a far lower price for dupes.com, went back to the owner of dupe.com and said, if you don't sell this domain to us in, like, five minutes with this price, which I thought was a fair price, it was all of our money. And we gave them all of our money. And you know that we're going to pull the trigger on dupes.com and they sold it to us, like, on the spot. Basically said, okay, fine, go for it.

B (156:25)

Smart. That's great. That's how we negotiated it. I love that. That's great. Thank you. That's a good story.

D (156:32)

Yeah, but that was like a bet. The. Like a literal bet. The company moment. And then Nikita did the most mensch thing ever. He put out a tweet about us and instantly blew us up.

B (156:41)

It was crazy. I remember that tweet. He framed it very, very well. And then you can imagine this. It just works on social.

A (156:47)

What other categories do you think are most ripe, other than furniture?

D (156:52)

Yeah, so we do fashion now. Handbags, shoes, clothes. Really big one that we don't do very well yet because AI doesn't have a sense of smell. You know, it's only shaking kind of vision. Perfumes. Perfumes are huge in the.

A (157:07)

Basically, I look at it as anything without a logo. So, like, furniture, like, there's iconic shapes and silhouettes and, you know, materials that get used along that. But ultimately, people don't have, you know, you're not like, you don't have a couch that has, like, a restoration hardware, big logo printed on it. The challenge with fashion is that people want, like, the association with a specific, like, brand. And that can come through a logo or a. Or a silhouette that actually can't be implemented without getting. You know, you're right, you're right.

D (157:44)

Except we're going through this, like, bizarre moment in shopping where brandless is a thing and not showing logos is a thing. Like this. This kind of quiet luxury movement. Like, people. People want something to feel good and that buttresses, like, even the brands that you love, they're white labeling guys like, they are, are. They are. None of this stuff is, like, custom made specifically for a single brand. Maybe if you're going into the Couture world, of course that stuff is, is custom made. But ultimately, you know, when someone sees skims leggings, like, yeah, they, they want the look. People shop. People shop the look like that's what they're shopping ultimately. And it's, it's less about like, you know, people often misconstrue what dupe.com is like. We're not trying to sell you replicas. We actually ban replicas across the board.

A (158:32)

Yeah, yeah, yeah.

D (158:33)

And all those sites, we are trying to get you that look, that garment, you know, that furniture for less. And most people shop for the look now, especially in this market.

A (158:44)

Yep, that makes sense. What, how do you, do you think you'll have enough leverage at some point to get, to get the underlying brands to pay you at the time of sale? Is that even important?

D (159:00)

It's actually very important. I'll tell you why it's important. If we were growing 100% month over month and we weren't seeing our cash flow for three months, it stifles our marketing budget. We would have to then either raise credit or go raise a massive round just to pay down the marketing. This is actually the biggest risk that I think about. So one of the things we're partnering with the largest payment network to do agentic checkout, global checkout. And you know, we're not announcing the, the name yet, but we are working through that tech and the idea being, you know, most of these brands are really upset with agentic checkout because they, they don't get the customer information. But we do it in a way where the checkout would happen on dupe, but the brand retains the customer information and the customer relationship. And the, the only reason we want to do that is so that we can, can go from net 90 payments to net 0, net 1 instant payment back to our company. So yeah, this is very top of mind for us and it's definitely a risk that I think a lot about.

A (160:02)

How big do you want dupe to get?

D (160:06)

I think dup, we often talk about a road to a billion shoppers. The way that we're going to do this is we have an mcp so any LLM can plug us in. You know, the largest LLMs and the long tail LLMs, they can plug us in. Ultimately, we want to be the canonical answer engine for how do I get this for less? Or find me something similar to this. We want to be that answer engine for these websites. The beauty is because they don't have access to the behavioral data, the shopping data, the product Catalogs the product pricing. The moment they plug us in, it's basically zero effort. It's just, you know, a few lines of code. The moment they plug us in, in any click going out from our answers that gets monetized. We split that commission back with the LLM so it's like instantly monetizing their user base in an entirely new way. So that's something that we're very excited about. When you think about how the growth potential of that and the tailwinds for any LLM to really consider partnering with us for our links and for our expertise and the ability to monetize instantly, I think we can get to hundreds of millions of shoppers on our infra. They might not be on dupe.com but they'll be.

A (161:16)

You're still monetizing them.

B (161:18)

Yeah, that's right.

D (161:19)

That's right.

B (161:20)

That's very exciting.

D (161:21)

That's right. And in the last 18 months we, you know, we're right around sub 20 million users between.

B (161:28)

Let's go.

D (161:29)

Oh my God. Do I get the Kong?

B (161:32)

No, terrible. Were going to give us a number, but we got.

D (161:36)

Yeah, yeah. No, no, no, no, no. So, so I'll give you a few numbers. I mean we're, we're rounding a corner. We're going to come up on 100 million in GMV sold through Duke.com.

A (161:44)

Whoa.

D (161:45)

18 million shoppers.

B (161:47)

Fantastic.

D (161:47)

Between one and two monthly. One and two million monthly shoppers, which we just launched our app last week, went number one in the App Store. By the way, super stoked about. Please, everyone listening, Go download Dupe Dup. Look for it in the App Store. Le leave us a review and rating. Gotta put in the plug went number one on the App Store and of course we had a lot of fun with that. We've been putting a lot of fun content out.

A (162:07)

How did you go number one? Was it content or some type of viral incentive?

B (162:14)

We have a lot of users.

D (162:15)

These people, they want our app, they use our website to shop and have used our website to shop for many months. They wanted our app and the way the App Store was works is on a 24 hour clock. So the number of installs you get in a 24 hour period, if it's more than any other app, you rock it up to the top. I didn't think it would happen on day one. I actually had a whole. I was convinced that we were going to go number one. First of all, I called the shot in August on Twitter. I said we are going to unseat dethrone ChatGPT when we launch our app in November. So we, we shot our shot and we kind of netted it, but it happened within a couple hours of us going live. I didn't expect that. I expected it to be about 4 hours, 48 hours. So it happened very, very quickly.

B (162:58)

Wow.

A (162:58)

Very cool.

D (162:59)

I mean, we still have a lot, we still have a lot of work left. I mean, you know, obviously we're at the beginning of our journey.

A (163:04)

Call your shot again. 100 million in GMV. When are you gonna, or when, when are you gonna hit it?

D (163:11)

We're a couple months away from that.

A (163:13)

Okay, well, we got the gong ready for you.

B (163:15)

Amazing.

A (163:16)

You're in la, right?

D (163:18)

New York.

A (163:19)

Okay.

B (163:19)

Okay, well, well, we should, we should.

D (163:21)

Be able to get to that, that, I mean this, this shopping season. Let's see how Black Friday goes. Make sure none of our systems go down. But that, that would be pretty dope.

C (163:28)

Yeah.

A (163:29)

This is, this is your Super Bowl.

B (163:31)

Very cool. We'll talk to you soon.

A (163:32)

Great hanging dude progress for having me.

B (163:35)

Congratulations. We'll talk to you soon. Let me tell you about wander.com book a wander with inspiring views, hotel grade amenities, dreamy beds, top tier cleaning, 247 coner service. It's a vacation home, but better. And that, that's our show for today. That's this week in tvpn. Thank you for tuning in. Thank you for listening. Thanks for hanging out in the chat. If you've been chatting this week, we love you and we'll see you on Monday.

A (164:02)

Have a fantastic weekend.

B (164:03)

Have a great weekend. Goodbye.