Sam (2:54)

Yeah, it was like 500 billion of RPO which is like that's AWS size numbers and that's a lot of what I think is going on here. That's actually the dynamic like the market is demanding another aws. And so let's go through the top line numbers first. So analysts estimated 86.7 billion on revenue. Oracle said that they will hit 90 billion for fiscal year beginning in June. Now people don't care as much about the top line number. Everyone's obsessed with infrastructure business. The infrastructure business. So the previous quarter showed growth of 68%. That's not bad. Analysts said this quarter we're looking for 79% and Oracle delivered 84%. So they beat estimates. That's why the stock popped. Of course, the infrastructure business is particularly important for two key reasons. So first, Amazon, Microsoft and Google. There's more than just them when it comes to hyperscalers, but they are unique among hyperscalers in that they have AWS, Azure and GCP. They have true cloud platforms. Cloud infrastructure platforms. OpenAI does not yet Meta does not yet Meta is a hyperscaler. Like they build huge data centers but they don't have this flexibility that comes with operating something like AWS. And so who signed big deals with Oracle? It's OpenAI and Meta and fortunately those deals are going better than people expected. People were worried and people were saying is Larry going to get caught holding the bag? Well, it looks like things are penciling out so far. So Meta and OpenAI both have massive ambitions around AI. They both signed on with Oracle to ramp up data data center capacity. Oracle is just this extra burst of CapEx capacity in the system and they're ramping up. So 50 billion in CapEx in the current fiscal year and they're consist outpacing analyst estimates. In the fiscal third quarter, analysts predicted $14 billion of capex. Oracle spent 18.5. So Larry is certainly opening the pocketbook, digging for coins in the couch cushions. But we'll get into the long term implications of like what does this mean for becoming over leveraged debt drawing down on cash flow. There's actually a lot of interesting structure going on within the financials that shows you how even though he's definitely risk on definitely AGI pilled all full tilt ahead into the build out, it doesn't feel like financial recklessness because of the structure of the deal. So the second reason that infrastructure is so important right now that I think people on the outside are sort of missing is that compute is still growing exponentially. But AI adoption curves often look like S curves. And so the classic one is consumer LLM usage. So OpenAI came out with ChatGPT and the numbers were just like, insane. It was like 20 million users overnight and then 100 million and then 200 million. And it was just like, okay, this curve is going completely vertical. We're going exponential. But of course, there's only 8 billion people on earth and a lot of them just don't use really computers, apparently, because Meta, over two decades has only gotten three and a half billion users to use, like, the whole suite of apps, including everything. So just WhatsApp, you can just be a dau of Instagram and you count in that, like we all, everyone in this room counts in Meta's dau for sure. Even if you're like, I'm not really that big on Facebook, or I'm not that. I don't use WhatsApp that often. Like, you definitely. They got you. Except for half the human population, which is not on meta platforms for a variety of reasons. Some geopolitics, some economic. But basically there's going to be a slowdown. So OpenAI shot up to basically a billion mu monthly active users. The official number that's getting like trotted around Right now is 920 weekly active users. But everyone behind the scenes says, like, yeah, they're well past a billion MAU. And so that curve is slowing down. Like, there's just no way you can be like, yeah, actually, like, we're expecting 10 billion ChatGPT users next year because, like, there aren't 10 billion people. So there's going to be deceleration in consumer AI usage.

Jordi (7:11)

Sufficiently aligned superintelligence might want to stimulate the growth rate. Human population.

Sam (7:17)

Good take, good take. Yes, yes. You go to ChatGPT and says, like, hey, I know you've been thinking about having a fourth kid. You should do it. And here's why. Maybe. But at least for now, we, we are seeing deceleration. And I think for just a lot of people, they're like, yeah, I use AI. I have an app on my phone. Maybe Gemini, Maybe Claude, maybe OpenAI, maybe ChatGPT, I use it. Maybe I use Grok, maybe I see it vended into different systems. I interact with it on Instagram, but I'm in But I'm not blown away by the growth of this thing because everyone started using this a year ago and we're still all just using it. But COMPUTE is scaling very differently because when we went From LLMs and ChatGPT to reasoning models that probably 10x, the amount of tokens that people were generating with 01, and then once GPT5 came out, the reasoning models became much higher usage rates. And then the agents framework, the open clause, the codexes and the cloud codes that all did another 10x in token volume. 10x is like a very rough number, but it's basically growing exponentially. So you have these double exponentials.

Jordi (8:27)

Yeah, it's just the way to think about it. Instead of you coming into ChatGPT to do some type of query, you're effectively one person can effectively multiply themselves by 10:20. And then it's just constantly using it all day long, all day long, all day long.

Sam (8:44)

So there's so many times when I will fire off a GPT 5.4 Pro query or a deep research report before bed, it's going and generating tons and tons of tokens. It's like I'm on that app for like five hours. If you were to go back to the ChatGPT moment when it was running 3.5, way more compute expended. But my actual time on the site is pretty limited because then I just get the deep research report, scan it, listen to it, whatever, and then I leave the app. So the actual experience, surface area and the number of people that are like trying AI for the first time is decelerating because everyone's tried it. But Compute is still 10xing, so token consumption per user is exploding. And OpenClaw and agents like Codex and cloud code are also an early part of the S curve adoption. So we're still in that. It's still in that exponential. And so they're melting GPU fleets, as Thibaut over the Codex team shared. But fortunately, I think things are back online. He reset the limits. The resetter of limits. Reset the limit limits. So enjoy your Codex if you're vibe coding. Today, there was some viral bear posting about how Oracle was in financial trouble because of the economics of their infrastructure bets. So naturally they have to buy the GPUs before they can rack them and sell them as infrastructure. This is business 101, but people were maybe surprised by that or something. There was a whole press cycle about this and it seemed very silly at the time. And I think everyone was like, yeah, this is exactly what we expected. But a lot of people from the

Jordi (10:22)

people that brought you the 1 gigabyte

Sam (10:23)

data center, similar, similar crowd potentially. But there was a lot of fear around that. And some of that is legitimate because if GPUs depreciated really quickly, Oracle had to buy the GPUs, they had to send the money to Nvidia a year in advance and then it took them two years to get them into data centers and actually do the deals and all this stuff. That would have been a squ squeeze on cash flow for sure, but that's not what happened. So we have some hard data on profitability now and Oracle is first up. They're on or ahead of schedule with 90% of the capacity deliveries and that means happy customers. And then second, gross margins actually improve. So guidance was 30% and they hit 32%. So Oracle's AI infrastructure is profitable the moment it comes online. And they're also increasing that backlog. They increase the backlog of RPO remaining performance obligations to $553 billion. So customers are happy, they're ramping, and overall the demand is just flowing in the right direction. Companies and consumers are happy to pay for AI tools and LLM tokens. Labs and AI inference providers are happy to pay Oracle for infrastructure without crazy delays. And this means Oracle doesn't need to get into dangerous financial engineering territory where they need to go crazy. Negative cash flow. Issue a bunch of equity, issue a bunch of debt, and in some cases they can actually get the customers to pay for the GPUs upfront or as they put it, like bring their own hardware. And so the project is continuing, which is all good news. All good news. And we saw the God Candle print the Kobiesi letter shared Oracle Stock surges over 8% after beating earnings and posting a 44% jump in cloud revenue. Before we dig through the timeline, let me tell you about Vanta Automate Compliance and security. Vanta is the leading AI trust management platform. And let me also tell you about Gemini. Gemini 3.1 Pro is here with a more capable baseline. It's great for super complex tasks like visualizing difficult concepts, synthesizing data into a single view, or bringing creative projects to life. So there's a lot of backtracking going on and there's a lot of confusion over the interpretation of of the guidance from the from the Oracle team. One year ago, management said fiscal 2027 top line growth rate would be around 20%. Last quarter the company said that 2027 sales would be 4 billion higher than previously expected. Putting that all together, Oracle's previous 2027 sales guidance was in the neighborhood of 84.4 billion ahead of this report. And Luke Kawa says Oracle management beat normal. Challenge impossible. People are all over the place. But Ben Thompson has a great deep dive into Oracle's financial release that you can dig through at Stratecheri if you're a premium subscriber.

Jordi (13:16)

They also had some comments on the SaaS apocalypse from the earnings hall. You've all heard the thesis that new companies Coding quickly using AI will spell the death of SaaS. I don't agree with that at all. I do think that AI tools and their coding capabilities would be a threat if we weren't adopting them. But we are. And very rapidly. Oracle is using the best AI coding tools and the best developers. The use of AI coding tools inside Oracle is enabling smaller engineering teams to deliver more complete solutions to our customers more quickly. We are building brand new SaaS products using AI and also embedding AI agents right into our existing application suites. By embracing AI with small engineering teams, we have built three brand new CX applications. Let's give it up for customer service and our new website generator.

Sam (14:04)

I think Oracle's current website was AI generated. I think they said this right?

Jordi (14:08)

Yep, they said, in fact, we just used the website generator to build and launch the new website. Dogfooding.

Sam (14:13)

Let's go.

Jordi (14:14)

How's it look?

Sam (14:15)

It's beautiful.

Jordi (14:17)

It looks AI generated.

Sam (14:20)

It's very good. It opens with just eight buttons. Cloud, multi, AI database. Multi cloud, AI database, AI data platform, Cloud at customer.

Jordi (14:31)

You might not like the design of this, but this is peak, peak performance. Let's pull it up. We got to black people.

Sam (14:39)

It's so good. No, obviously they're doing very well. And this is a functional website.

Jordi (14:43)

Look at this, look at this.

Sam (14:46)

I like the way the buttons lay out horizontally. Is particularly crazy crushed. Well, it certainly loads in record time. So performance matters. And that's what Oracle stands for. So we need to go.

Jordi (15:04)

They got to make the search bar bigger.

Sam (15:06)

So we need to go back in time to understand Larry ellison. Because in 2013, Vanity Fair wrote a profile about Larry Ellison that is absolutely unhinged. I've never seen a profile like this about a tech CEO. So at the time, he was easily the richest man in California with $6 billion. Oracle CEO Larry Ellison, co founder of the world's second largest software company, is Silicon Valley's most.

Jordi (15:33)

This is from the 6-19-97 issue.

Sam (15:37)

It's a 1997 issue. This has been uploaded to Vanity Fair in 2013. Sorry, this is from 1997. So right in the heat of the dot com boom, things are just kicking off. We are far from the crash. Everything is off.

Jordi (15:52)

Genuinely incredible that back then you could be the richest man with a paltry 6 billion.

Sam (15:57)

It's crazy. That's like an aqua hire these days.

Jordi (16:01)

It actually is. Love from.

Sam (16:03)

I know. Oh yeah, love from. I was thinking of Windsurf, but the fact that we're thinking of multiple is insane. Before we read through this, let me tell you about graphite code review for the age of AI. Graphite helps teams on GitHub ship higher quality software. And let me also tell you about Label Box. RL environments, Voice robotics evals and expert human data. Label Box is the data factory behind the world's leading AI teams. So he was easily the richest man with $6 billion. In California, Oracle CEO Larry Ellison, co founder of the world's second largest company, is Silicon Valley's most notorious playboy and a sportsman of the first rank who flies fighter jets and races world class sailing boats. But his burning ambition is simple, if naive. Bring down Bill Gates. They were in a bitter rivalry at the time. Microsoft, of course, run by Bill Gates at the time. In the 90s, as Ellison moves to add Apple Computer to his anti Microsoft arsenal, Brian Burrows locates the fragile psyche behind the bravado. Michael Jordan is streaking down the court on a fast break as Larry Ellison sitting seven rows up from courtside. Weird, right? Is that a better seat? I don't know enough about basketball.

Jordi (17:13)

I just think 6 billion couldn't.

Sam (17:15)

Maybe he wasn't liquid. Maybe he just couldn't afford the actual courtside seat. It seems crazy, right? That he's seven rows up. Although I don't know, he's having fun. Cause he's at the United center in Chicago. Not where he lives. He's in California. But he went to Chicago. And he begins enthusiastically telling the story of Rupert Murdoch. Rupert Murdoch's severed fingertip. This is a crazy story. I had no idea that this happened. It happened when Murdoch was crewing on Ellison's championship sailboat, the Sayonara. Making coffee and handling minor chores during a race off the south coast of Australia in 1995. You're just like the guy who's just gonna hang out and like, hey, hey, you don't really know anything about this. You're just on this boat for fun. Make me coffee. And it's Rupert Murdoch, which is insane.

Jordi (18:03)

We got to shake Rupert's hand. I didn't notice any missing.

Sam (18:06)

We're gonna get to that.

Jordi (18:07)

Didn't notice any missing fingertips.

Sam (18:08)

So just as the race ended, Murdoch made the mistake of grasping an overhead rope, which promptly shot through his hand, ripping the tip off of one of his fingers. And he didn't say anything. He just kind of put his finger in his mouth, sucking on it. Ellison says, chuckling, brutal. I can't remember who picked up Rupert's finger. Thank goodness it didn't fall overboard, but we picked it up and put it in a plastic bag and put him in the chase boat that night. After successful emergency microsurgery. There's microsurgery. I guess it wasn't that big of a deal, but that seems insane that they had to reattach this finger. We. So there was a. They do the emergency microsurgery at an Australian hospital. Murdoch amazed Ellison and his crew by making it to the after party and then several days later, by crewing again on a race.

Jordi (19:06)

Absolutely.

Sam (19:06)

To the Tasmanian capital of Hobart and Rupert Murdoch. It's not like he was 25 in 1997. I'm pretty sure he was in his 50s.

Jordi (19:15)

He's 95 now.

Sam (19:17)

Yeah. So he must have been.

Owen Jennings (19:19)

He must have been.

Sam (19:20)

He was 60.

Jordi (19:21)

He must have been almost closer to 70.

Sam (19:23)

Yeah. Wow, that is incredible. Of course. Ellison says with his little boy's grin. That doesn't alter the fact that his coffee was horrible. I mean, the man runs a great company, but his coffee sucks. It's incredible. Over the course of the Bulls pounding of the Indiana Pacers, Ellison, who grew up on Chicago's tough South side and remains an avid Bulls fan, breaks into lusty cheering and joking boasts. I can do that. He shouts after one thunderous Jordan dunk. Such a crazy heckle. Heckling Vincent, Michael from the seventh row. You didn't upgrade to the front courtside seat, but you're still heckling none other than Michael Jordan. Absolutely insane to gossip about nearly every power player at the nexus of technology and communications in the 1990s. His arch rival, Bill Gates. His best friend Steve Jobs. His business partner Mike Milken, Mike Ovitz, Ted Turner, Murdoch. He just piloted his Cessna Citation into Chicago's downtown Migs Field from New York, where he spent the previous day in meetings with Viacom, Sumner, Redstone, Intel's Andy Grove and Ray Smith of Bell Atlantic. So you may be wondering.

Jordi (20:37)

You might think it's crazy he's piloting his own Cessna, but during that era, he was getting into dog fights with his son.

Sam (20:44)

Exactly. He needs something more agile. He can't be flying a 737, 747 at that time. So you may be wondering, who is Larry Ellison and why is he spending so much time hobnobbing with the techno elite? It's funny because I think of him as the techno elite and the other folks as like the media elite. But I guess Bill Gates and Steve Jobs are the techno elite at the time. For one thing, Ellison is the wealthiest American you've probably never heard of. With a Fortune estimated at 6 billion, he is easily the richest man in California. Think three David Geffen's ten full Milkins, and according to Forbes, the fifth richest man in the nation. For another, he is co founder, controlling shareholder and chief executive officer of the world's number two software company, Oracle Corporation, which makes the giant computer databases in which American Airlines keeps track of its planes, Ford Motor keeps track of its spare parts, and the Central Intelligence Agency keeps track of, well, whatever the CIA keeps track of. He could also be the next head of Apple Computer. This was a crazy rumor that never came true. The famed but faltering industry icon he has been eyeing for the past two years. At the end of March, Ellison surprised Silicon Valley by suggesting that he was about to launch a takeover bid for the company. So he was friends with Steve Jobs, but was like, this company is just underperforming. It was underperforming all the way until the iPhone basically and ipod. But the 90s were like a very rough time for Apple as they sort of rebuilt. Tim Cook of course, joined and did a ton of work to improve the supply chain, get them to the place where they are today, where they've been so dominant. So at the end of March, Ellison surprised Silicon Valley by suggesting he was about to launch a takeover bid for the company. On top of all that, Ellison is the mind behind the most talked about new idea in computing in the last two years. The network computer known as the nc. A stripped down personal computer that stores its files on a network instead of a hard drive.

Jordi (22:46)

It could be big, something there, something there.

Sam (22:49)

Because it doesn't have to use Microsoft's omnipresent Windows operating system. The NC represents one of the stiffest challenges yet to Bill Gates dominance of personal computing. It's the original Mac Mini and it's catapulted Ellison into the national, into the international spotlight. There's a very, very funny moment in here where there was a time, few months back, a few months back, for instance, when he created a stir by going on Oprah to talk about the nc. He Goes on Oprah and is like,

Jordi (23:22)

we gotta talk about network computing.

Sam (23:23)

We gotta talk about network computing. Guess what? It's headless. You're not gonna need Windows. And they're like, what? So it's a computer that I can't use without the network. I don't even know what that is. He's just talking about the most extreme future. And then so he's talking about the nc only to have the show take a sharp turn toward his personal affairs when he confessed that he had yet to find the right woman to fill the void in his life. And then they go into some of his life. But. So when the San Francisco examiner asked Bay Area celebrities for a favorite Valentine's Day memory, Ellison mentioned something he did on the side of Silicon Valley's Woodside Road. After his Oprah appearance, Oracle's phone lines were jammed with thousands of calls from women. The joke inside Oracle was that the company's new recording would be press 1 if you want information on Oracle's products. Press 2 if you want information on Oracle services. Press 3 if you want to fill the void in Larry's life.

Jordi (24:29)

Larry would have loved Instagram.

Sam (24:31)

DM's insane. It's one of the craziest profiles. We don't have time to go through all of it. But it is a fascinating portrait of Larry Ellison while he was on the rise. But we should switch over to David Ellison first. Let me tell you about Lambda Lambda is the superintelligence cloud building AI supercomputers for training and inference that scale from one GPU to hundreds of thousands. And let me also tell you about Cisco. Critical infrastructure for the AI era unlocks seamless real time experiences and new value.

Jordi (25:00)

Yeah, I highly recommend going and finishing this article because it actually somehow gets even more unhinged.

Sam (25:06)

It's crazy. It really, really destroys us because we were in Vanity Fair. And there's some fun elements, there's some fun vignettes, but nothing. We're no Larry, no. Fingers flew off in the making of the TVPN profile by Julia Black in Vanity Fair a month ago. But I weaseled my way back into Vanity Fair yesterday. Julia asked me for a comment on what's happening with David Ellison, with Paramount, Skydance and the merger with Warner Brothers. And I gave a comment, got into Vanity Fair.

Jordi (25:42)

Julia writes, is Paramount's AI first merger a force multiplier or flyboys all over again? Hollywood is bracing for layoffs and big creative changes if the Paramount Warner Brothers Discovery merger goes. Mega merger goes through.

Sam (25:57)

So there's a very Funny vignette to start this Vanity Fair piece by Julia Black where she's talking about, she went to A16Z's American Dynamism Summit and was asking people there about media. So she was talking to a media executive in line for the bar. And he predicted countless small indie outlets catering to highly niche audiences. Love that prediction. And then he says, and then David Ellison, Skynet, he deadpanned. The Terminator reference was the kind of thing CGI explosion fanatic Ellison might actually appreciate. And the joke being dropped at a defense Tech event in D.C. goes to show how many different industries, from tech to politics, are keeping an eye on the M and a drama unfolding in Hollywood this month. So the merger brings together dozens of media properties. Paramount, Warner Brothers, HBO, CBS, TikTok, Oracle, Silicon Valley's ties. There's a lot of stuff that's going together. It's a good time. But they're going to be using AI to streamline back office. They're going to move to Oracle cloud databases to centralize everything. These are standard M and A things. Everyone's worried about. Layoffs might happen. People are particularly worried about one of the two physical studios that they own selling. I think they got to give us a call because we're looking for a new ultra Dome. And nothing would be greater than having the entire Warner Brothers studio to ourselves. We're like today on tvpn. Yeah, Ideally, we'd have launching a car off of a. We're going to light ourselves on fire. You really can't do that in these studios. But when you know they want to produce 30 films, realistically, how many of those are going to be shot in Hollywood studios? Does it make sense financially to do some stuff in Atlanta, some stuff in Toronto, some stuff overseas? It all depends on where the moviegoer's taste lands, what ticket sales are like. Paramount right now is saying, look, layoffs aren't really the primary way we are going to get consolidation or value here. What did they say? They said something like layoffs. A rumor. Let me see. The layoff fears are overblown. Synergies will be achieved in six key areas. Jobs are not the majority. So we'll keep figuring out what's going on there. I did give a quote here. Julia writes, AI could unlock new potential for Warner Brothers discovery, treasure trove IP from Harry Potter to DC Comics. For that potential, Ellison and company are paying a hefty $110 billion, a number that was driven up by a fierce bidding war with Netflix. And I said, the Ellison family is Fascinating. On one hand, you have the very aggressively AGI pilled Larry. I love that I stuck AGI pilled into Vanity Fair. AGI pilled, Larry. Building the future, Investing heavily in Oracle data centers as we just told you about, with Oracle earnings. And on the other side, you have David buying the past, accumulating intellectual property that feels impossible to rebuild. I'd like you to try and generate something as iconic as Porky Pig. It's impossible. You can't.

Jordi (29:02)

It's impossible.

Sam (29:04)

Duo is basically long slop and long anti slop. And so this is an interesting.

Jordi (29:10)

They're hedged, I think market neutral.

Sam (29:12)

I actually don't think it's market neutral. I think. I think they genuinely believe that both generative AI will accelerate and the value of intellectual property will increase. It's not. It's not. Oh, we're going to live in a future where we're all watching the Dark Knight on film in theaters again. Or we're watching Jenny I sora feeds. It's. We're doing both and they're actually going to meet. And so I believe that they are long both of those. It's not a market neutral bet, in my opinion, but we can debate which one's the better bet, which one's more realistic. How valuable is the Warner Brothers IP library? $110 billion is a lot of money. And it has a lot of people thinking, is there another angle here? Is this about control of the narrative? Is this control, creative control? Will the filmmakers who write these storied films that we know and love, like Batman, will they have the creative freedom? Or will Larry Ellison step in and put his thumb on the scale and try and inject a little bit? And there is real cause for concern. We actually got a little leak here of a script for what could be the seventh Synergy. Potentially, yeah. For what could be the next Batman film. So this is called the Dark Knight Migrates. So we're gonna do a little table read of the Dark Knight Migrates. It starts in the Batcave at night. Batman stands before an enormous monitor. Alfred approaches with T. You'll be Batman, Geordi. Perfect, sir. The Riddler has taken Gotham's entire power grid hostage. He's encrypted every system in the city.

Jordi (30:56)

Pull up the city infrastructure schematics, Alfred.

Sam (30:58)

I'm afraid I can't, sir. Our on premise servers are buckling under the load. If only we had a cloud based solution with autonomous threat detection and built in machine learning.

Jordi (31:11)

Alfred, launch the Oracle cloud infrastructure console.

Sam (31:15)

Oh, so the integrating. Integrating Sponsored Content for Oracle. That's how you monetize the Warner Brothers ip. Yeah. The DC Universe, Alfred.

Jordi (31:26)

It looks visibly relieved.

Sam (31:27)

Right away, sir. Spinning up an OCI tenancy. Now, I've taken the liberty of provisioning a few ampere compute instances as well. They have an excellent price to performance ratio, I might add. So they're giving the viewer what they want. They want the story of Batman. But they're also sneaking in just a little bit of extra detail about Oracle's offerings.

Jordi (31:46)

And migrate the Bat computers. Databases, all of them.

Sam (31:49)

Oracle does do great migrations to Oracle. To Oracle Autonomous database, sir.

Jordi (31:55)

Is there another kind?

Sam (31:56)

It does patch and tune itself, sir. Which is fortunate since I also have to iron your capes. See? They're still putting in the cape ironing. That's still in the Batman world. You're getting a little Oracle, but you're also getting a lot of Batman. So now let's shift over to the Riddler's hideout. Tyler, would you like to be the Riddler?

Mike Blue (32:14)

Yep.

Sam (32:14)

Okay, I'll be the henchman. Riddler is in front of a wall of monitors showing Gotham in chaos. Okay, riddle me this, Batman. What has no locks but can't be open?

Apoorv Agarwal (32:24)

Gotham's grid.

Sam (32:25)

I've encrypted it with my own proprietary algorithm running on seven different NoSQL databases held together with Python scripts. Oh, that's a nightmare, boss. The systems are getting kind of laggy. Says henchmen 1. Just restart the servers.

Jordi (32:39)

Which ones? There's like 400.

Sam (32:41)

All of them. Oh, no. So we go back to the Batcave. Batman types furiously. Oracle logos glow softly on every screen. A hologram of Hilary Ellison rotates slowly in the background for no discernible reason.

Jordi (32:54)

Batman, Batman says, I've identified the Riddler's encryption vector. Alfred, deploy the decryption countermeasures across all OCI regions, simultaneously leveraging Oracle's global network

Sam (33:07)

of over 40 cloud regions. Sir, latency is under 2 milliseconds. Shall I enable Oracle data guard for disaster recovery?

Jordi (33:15)

Always. Gotham is the disaster.

Sam (33:18)

I've also taken the liberty of enrolling us in Oracle support, the premium tier.

Jordi (33:23)

That's the most responsible thing anyone in this city has ever done.

Sam (33:26)

An alert flashes, sir, the Riddler appears to be running his operation on a hand managed MySQL fork.

Jordi (33:33)

My God, he's not a super villain. He's just under architected.

Sam (33:39)

Okay, back to the Riddler's hideout. Moments later, Batman crashes through the skylight. The Riddler spins around. You're too late, Batman. My encryption is unbreakable Every screen behind him turns blue. Then they all display ORA00054 resource busy and acquire with no wait specified. What?

Jordi (33:59)

No. I migrated your databases to Oracle Autonomous while you were monologuing. Your encryption keys are now safely stored in OCI vault with customer managed rotation policies.

Sam (34:10)

You can't do that. That's my data.

Jordi (34:12)

You should have read the shared responsibility model. Batman punches him in the face.

Sam (34:17)

Okay, later, I'll be Gordon. Batman stands on a gargoyle. Commissioner Gordon approaches. You saved Gotham again, Batman. How did you do it?

Jordi (34:28)

Gotham deserves enterprise grade infrastructure, Jim. Scalable, secure, autonomous.

Sam (34:35)

Are you okay?

Jordi (34:37)

Oracle cloud from the Batcave to the boardroom.

Sam (34:40)

That's not even a good answer. He fires his grapple gun and vanishes into the night. The Bat signal illuminates the sky, but tonight it's shaped like the Oracle logo. Gordon alone, quietly. I really should talk to him about this. Smash cut to black title card. Oracle the cloud. Batman trust. Shouldn't you. I think we got something that will finally put butts in seats in Hollywood. David Ellison team up with Larry, get back in the dogfight and make Batman the Dark Knight migrates. I want to see it. That's the movie I want to see. That's what I want to see. That's fan service for the Oracle heads in the audience.

Jordi (35:22)

That's real synergy.

Sam (35:24)

That's real synergy. Let me tell you about Sentry. Sentry shows developers what's broken and helps them fix it fast. That's why 150,000 organizations use it to keep their apps working. And let me also tell you about restream one livestream. 30 plus destinations. If David Ellison wants to multi stream the Dark Knight migrates, go to restream.com. so there's an interesting call sheet out there today which companies will release a fully AI generated, multi episode scripted series before 2027. Overall, it's pretty low. It's 16% for Netflix, 14% for Disney. Remember, Disney has a deal with OpenAI alongside Sora, but Disney has not said, oh, okay, yeah, we're actually going to do this. And Paramount plus is at 10%. Of course, David Ellison has been talking about AI mostly in the enterprise, mostly actually unironically in OCI and Oracle databases and stuff.

Jordi (36:21)

Yeah, I would expect Paramount to not want to be the first mover here.

Sam (36:26)

Yeah, there's a big question about who's the first mover. We talked about first mover disadvantage with regard to ads and LLMs. There was a whole fight over that. I'm excited for fully AI generated multi episode scripted series. I think that we will definitely see this on YouTube. We already sort of are. The closest thing I've seen is there's a really interesting YouTube creator who scripts and records his own videos, but then uses a pretty outdated style transfer illustration to sort of like turn him into an animated character. And so it's all puppeteered. There's no fully generated AI. It's all just go from video of a human being filmed on a camera to something that looks sort of cartoonish, but there's crazy hallucinations and like the lines never really match up. But it's been wildly successful on YouTube and, and I could see some sort of blend here. There's obviously going to be a discussion about what constitutes fully AI generated. How will companies talk about this? But it's certainly one to follow, as is Shopify. Shopify is the commerce platform that grows with your business and lets you sell in seconds online, in store, on mobile, on social, on marketplaces, and now with

Jordi (37:43)

AI agents, Odd Lots has a new episode on the impending fertilizer crisis. Yes, I'll just read the preview here. Yes, excited to listen to this one later. They say we all know that the war with Iran has sent oil prices spiking, but it's also pushing up the cost of all sorts of chemicals, including fertilizers and other nitrogen products that are essential for food production. This is all happening at the worst possible time, just before the spring planting season when fertilizer is most needed. And while farmers have seen higher spot prices for things like urea before, which is a fertilizer, notably back in 2022 because of the Ukraine war, there are already signs that this crisis might be worse. So how is fertilizer actually made? And what do higher fertilizer costs mean for farmers and food prices?

Sam (38:30)

I love Odd Lots. It's so good how deep they get into the supply chain. The Tracy and Joe have just been on such a generational run with that show and just keep delivering and it's just, I don't know, it's so refreshing to answer the questions that I have. Everyone gets the questions, the main questions answered about what's happening geopolitically from the front page of the Wall Street Journal or Bloomberg. But Odd Lots just so great at taking you deeper into these niche economic zones that I find personally fascinating. Over in the oil world, the IEA has approved releasing 400 million barrels of crude oil reserves, in effect to lower oil prices, the largest emergency oil release in history. How is oil trading today it's at $86 a barrel for crude oil, up 5% today. Still down from the highs of 110 where it was on Monday. But certainly lots of gyrations and not good for the folks who have supercharged V8s in their cars anyway. Let me tell you about console. Console builds AI agents that automate 70% of it HR and finance support, giving employees instant resolution for access requests and password resets. Jordi, where do you want to go?

Jordi (39:50)

Little coverage here on the Jones Act.

Sam (39:53)

Take us through.

Jordi (39:54)

We talked about this a little bit with Alex Epstein on Monday, but I figured we could go through this. This is from 005 seconds, an account over on X. The Jones act has four requirements. Vessels must be US built. Vessels must be US owned. Vessels must be US crude. Vessels must be US flags. The crippling part of the Jones act is that US Built US Shipyards, for a variety of reasons, are incredibly inefficient. We don't have that many of them, and they cost about five times what a ship from South Korea would cost. As a person who actually believes in trade, I fully would love for South Korea to become our U.S. shipyard. We just buy ships from them because they're good at making them. Coastal water transport in the US could be 60% cheaper because of the Jones Act. It's actually cheaper to ship goods from the U.S. to a foreign country and back to the U.S. then between two ports, which is completely bonkers. Insane. As a byproduct, we have killed all the growth within the Mississippi, which should be the most powerful inland economic advantage in the world. Maintaining the requirement of U.S. owned U.S. crude and U.S. flagged is perfectly fine and in line with my general national security concerns. But US Built has destroyed our shipping industry. There's tens of billions of GDP lying on the table here and a direct step in reducing our dependency on foreign suppliers. It's also how you kickstart rebuilding an American shipyard industry. If you 10x the number of US ships working in ports, you start building all of these maintenance businesses at U.S. ports and the demand increases. The U.S. bill requirement of the Jones act is horrifically destructive to America and in particular, horrifically destructive to middle America. And it should be destroyed. We should have somebody on that has a better understanding of shipbuilding in general.

Sam (41:40)

Yeah, I didn't realize how old the Jones act was. There's an article on history.com this day in history, President Coolidge signs the Jones act targeting bootleggers. The Jones act, the last gasp of the prohibition, is signed into law by President Calvin Coolidge. Since 1920, when the 18th Amendment went into effect, the United States had banned the production, importation and sale of alcoholic beverages, but the laws had been ineffective at actually stopping the consumption of alcohol. The Jones act strengthened the federal penalties for bootlegging. Of course, within five years, the country ended up rejecting prohibition and repealing the 18th Amendment, the Jones act, the Merchant Marine act of 1920. It's very interesting. A lot of the trick is, yeah, it takes time to rebuild a fleet and so there's a lot to do there.

Jordi (42:43)

What is George Hotz saying today?

Sam (42:46)

What is George Hotz saying today? First, let me tell you about Eleven Labs. Build intelligent, real time conversational reimagining technology interaction with Eleven Labs. And let me also tell you about CrowdStrike. Your business is AI. Their business is securing it. CrowdStrike secures AI and stops breaches.

Jordi (43:02)

So George says every minute you aren't running 69 agents, you are falling behind. Just kidding. Today we should ramp down rhetoric. I thought nobody would take three minutes to escape the perpetual underclass or you are worth. An hour. Seriously. But it looks like some people do and you shouldn't. Social media has been extremely toxic for the last couple months. It's targeting you with fear and anxiety. If you don't use this new stupid AI thing, you will fall behind. If you haven't totally updated your workflow, you're worth zero. There's people who built billion dollar companies by orchestrating 37 agents this morning and you just sat there and ate breakfast. This is all complete nonsense. AI is not a magical game changer. It's simply the continuation of the exponential progress we have been on for a long time. It's a win in some areas, a loss in others, but overall a win and a cool tool to use. And it will continue to improve, but it won't go recursive or whatever the claim is. It's always been recursive. You see things like auto research and it's cool, but it's not magic, it's search. People see AI and they attribute some sci fi thing to it. When it's just search and optimization. Always has been. And if you paid attention in CS class, you know the limits of those things. That said, if you have a job where you create complexity for others, you will be found out. The days of rent seekers are coming to an end, but not because there will be no more rent seeking. It's because rent seeking is a zero sum game and you will lose at it to bigger players. If you have a job like that or work at a company like that. The sooner you quit, the better the outcome will be. He's telling you to start polishing your resume. This is the real driver of the layoffs. The big players consolidating the rent seeking. To them they just say it's AI because that makes the stock price go up. The trick is to not play zero sum games. This is what I've been saying the whole time. Go create value for others and don't worry about the returns. If you create more value than you consume, you're welcome in any well operating community. Not infinite, not always, needs more, just more than you consume. That's enough. And avoid people or comparison traps to tell you otherwise. The world is not a red queen's race. This post will get way less traction than the DM1's but it's telling you the way out.

Sam (45:15)

Yes, a very good sober read. Sort of resetting the narrative. I am trying to think of like the green shoots that come out of like the social media toxicity because I agree that there's a lot of people that are afraid. I mean people like completely random people. Someone who has a job that will 100% not be affected by AI whatsoever, asking me about the something big is happening and just wondering like oh, what does this mean? Like am I going to be out of a job? And it's like no, you're definitely not going to be out of a job given what you do, because what you do is very physical and personal and like there's a million different reasons. Anyway, I do like the idea of a viral post about OpenClaw being not a fear based I need to change everything but a way for young people to get interested in building things, interested in technology broadly. Like I did not go through a traditional path to get to technology, I think. I mean I did take one computer science class, but mostly when I was a kid I learned like Adobe Flash and I used Adobe Dreamweaver, which was sort of a tool that let you generate websites without writing all the code you could kind of just highlight.

Jordi (46:40)

Dreamweaver was a fantastic name.

Sam (46:42)

Dreamweaver is a great name actually. Somebody should bring that back. I mean maybe Adobe should do something with it, but basically you could just highlight a title and change the font without needing to actually go into the HTML tag, look at the CSS class, change the font size in css and so it was a lot easier. But that was sort of an entry point for me to learn Python and learn C and learn more about computers over time. And I think that that's sort of the silver lining for all the crazy viral doom posting is that there's going to be people that just sort of naturally move to more creative jobs and maybe they're like, yeah, I was kind of going down a rent seeking path. I like building things, I like making things, I like creating value. And so they will come over to that and start having a positive impact. Which I think is good. But it's a funny reflection.

Jordi (47:35)

Yeah, the recursive comments interesting too. Obviously the technology has been the Internet and you make the Internet, it becomes easier and faster to make more Internet products.

Sam (47:48)

Mark Zuckerberg is using Google to write the first PHP for Facebook. That's just the thing. This has kind of been Andre Karpathy's take.

Amjad (47:57)

Right where it's like.

Sam (47:58)

Well, it's actually just kind of this smooth curve. This is not really like a zero to one thing. It's kind of just on the same path. I love smooth curves. Cortisol levels so low. So low. That's great. Tiny Corp is getting bigger. Tiny Corp posted a picture of the tiniest box, the biggest tiny box you can possibly get. It's the size of a shipping container. I would love this. This feels extremely apocalypse vibes and I'm extremely into it. I was running the numbers on how much juice can you get out if you're building your apocalypse bunker. What is the hardware that you should put in there that's like the most off the shelf and sort of a maxed out Mac studio. Maybe 512 gigs of memory might be the right thing to run some open source model so that you can endlessly talk to a friendly personal assistant while the. While the world collapses. Potentially. Anyway, the tiny box would also satisfy that. Tell me about the watch to house ratio, Jordi. What's going on here?

Jordi (49:09)

Funny, funny post. Just. I don't know where this was pulled from.

Sam (49:14)

Yeah, this.

Jordi (49:14)

But the line is just. He was put off by what he called the watch to house ratio. Paresh Raja, MFS director, sported a Richard Mille watch that cost £200,000. A contact estimated the watch was about half the value of Raja's North London home. So Godel passed on the opportunity to rest. And my interpretation of this is he wanted something more like a one to one.

Sam (49:39)

Yeah, exactly.

Jordi (49:40)

Like one to one.

Sam (49:41)

Yeah.

Jordi (49:41)

Like the.

Sam (49:42)

Should have downgraded the house.

Jordi (49:43)

A condo on the wrist or just

Sam (49:45)

wear an RM on each wrist. That would also.

Jordi (49:48)

That could have also solved this.

Sam (49:51)

I have no idea of the is real, but that is hilarious if true. And an absolute wild choice. But if you're doing business deals, you're going to shake someone's hand, they're going to see what's on your wrist, they're not coming over to your house. So in terms of actually accelerating your business career, there is a reasonable trade here. If you're in an industry that gives respect to luxury watches. Well, let me tell you about public.com investing for those who take it seriously. Stocks, options, bonds, crypto treasuries and more with great customer service. And before your stock is listed on public, you gotta go public at the New York Stock Exchange. Want to change the world? Raise capital at the New York Stock Exchange. Just do it. And when you become wealthy, you might relocate. The Wall Street Journal has a essay, an essay.

Jordi (50:41)

What is a city when its wealthiest leave? The stickiness that once anchored people in capital to great cities is gone and it is not coming back. Lot of movement recently.

Sam (50:54)

Richard writes the Nominative Determinism. Here he's writing about the movement. Google co founder Larry Page made headlines by spending $188 million on three Miami mansions.

Jordi (51:04)

What's the author's name?

Sam (51:05)

What's the author's name?

Jordi (51:06)

What's the author's name?

Sam (51:07)

Richard? Florida rich In Florida you got to make, you got to do better with this simulation. Whoever's simulating this world. Come on, come on. That's a layup right there. He's not the only billionaire looking at a big move. His Google co founder Sergey Brin is reportedly shopping for a Miami property. So is WhatsApp co founder John Jan Kuhm. The shopping spree comes as California considers a wealth tax to impose a one time retroactive levy on billionaires. That fueled speculation that they have had it with the Golden State in New York City, where Mayor Zora and Mamdani earlier this month proposed an increase in taxes on high income people. And short of that, a property tax increase. There is a talk of a parallel exodus. The wealthy has long threatened to leave when battling local governments over taxes. In the past, they rarely did. But their threats have teeth this time. Not because they are abandoned in great cities, but because they have figured out that they don't have to. Now that digital technology allows them to separate where they live and pay taxes from where their business operates. They aren't relocating their companies, they are relocating themselves. So they will all be puppeteering humanoid robots remotely in San Francisco, getting a coffee, doing a deal all the way from the comfort of their home in Miami, potentially, or just spending a couple months in the state. This Upends the basic arrangement that underwrote great cities. What they are, how they work, and who pays for them. Amid this seismic shift looms an existential question. Can those cities survive without them? This is always the funny bull take for the funny bull case for those like insane skyscrapers that are like super thin and super tall in New York and every apartment is like $50 million. And then the people don't even live there. And everyone's like, why are they building this? We should just build more housing. And it's like the property tax from all of those new apartments that are priced at 50 million each and no one spends any time in. That's a whole lot of money that goes towards the city, that goes towards buses and everything else that the city does. And those people who aren't in the city, they're not using the buses, they're not clogging up any of the resources. So it's actually a fantastic deal for the city. Usually for most history, people lived where they worked on the farm above the shop, close to the factory. Suburbanization expanded the radius, but workers, managers and executives still had to be close to where jobs concentrated because people had to be there. Cities could charge a premium. Residents paid it in housing costs, in taxes, and in the cost of commuting or other frictions of day to day life. The alternative, Living elsewhere meant being cut off from their livelihoods and economic opportunity. Taxes were part of that price. But people paid because they had no choice. When Covid hit, the social compact appeared to quake. A chorus of commentators predicted the imminent collapse of New York, London and San Francisco. Who was in that chorus of commentators? I think we know half of them. They predicted the wealthy and their companies would be driven out by lockdowns, governance failures, crime, and sudden possibility of remote work. The city as a result, would hollow out. There seemed to be something to it at first. Ken Griffin relocated himself in the headquarters of Citadel hedge fund from Chicago to Miami. The venture capitalist Peter Thiel and Keith Raboy bought homes in Miami beach and opened an office for their venture capital fund in Miami too. Jeff Bezos moved from Seattle to Miami, assembling a $200 million plus compound. But the predicted total exodus never fully materialized. Many of those who moved to Miami quickly came face to face with its limits. Public and private schools couldn't match what they had left behind. Housing costs rose astronomically, making Miami now one of the most unaffordable markets markets in the country.

Jordi (54:43)

Most critically, yeah, Miami so crazy if you have somebody visit anybody that visits from Malibu from Miami, they'll pull up Zillow, be like, everything's so cheap here.

Sam (54:56)

Yeah. And the weather says that just says.

Jordi (54:59)

Well, it just says how expensive Miami's got.

Sam (55:01)

Yeah, yeah, yeah. It is crazy. In 2023, Thiel admitted as much, conceding that the tech industry remained densely consequent, concentrated in California, and that Miami's housing costs put that city out of reach for much of the talent he needed, making it far harder to move companies and their people than he originally than he initially thought. Griffin himself ended up building a massive new building in New York, even as Citadel expanded in Miami. But eventually they realized they didn't have to move their companies at all. Digital technology enabled them to live in one place and keep their business in another. They could establish residency in Miami, which requires no particular time spent there to claim residency status and spend much of the year wherever else they wanted, flying to New York or San Francisco for what mattered. The city where their businesses are headquartered became just that, not somewhere they needed to reside and pay taxes. That has transformed the underlying economic logic of cities. They are no longer self contained economic units. Digital technology is remaking them into networks, physical places connected by virtual ties and dynamic talent flows. Before we move on, let me tell you about Fin AI, the number one AI agent for customer service. You heard from Owen earlier this week. If you want AI to handle your customer support, go to Fin AI. And let me also tell you about Plaid. Plaid powers the apps you use to spend, save, borrow and invest securely connecting bank accounts to move money, fight fraud and improve lending. Now with AI and without further ado, I believe we have our first guest, Apora Vagarwal, Ramel Timiner in the Restream waiting room. Let's bring him into the tv. How are you doing?

Apoorv Agarwal (56:29)

Great guys, how you doing?

Sam (56:31)

Thank you so much for taking the time to come chat with us. Welcome.

Jordi (56:34)

Great to finally have you on the show.

Sam (56:36)

Obviously I want to jump straight into consumer AI, but maybe you could, since this is the first time in the show, give us a little bit of intro on your role, your background, what you like to focus on.

Apoorv Agarwal (56:47)

Awesome. Well, I'm a partner here at Altimeter. I help lead some of our AI investments in companies like OpenAI and Glean and Parloa and Expo. And excited to be here. Thanks for having me.

Sam (56:59)

Yeah, you also did Redis, right? Is that correct?

Apoorv Agarwal (57:02)

Back in the day.

Sam (57:03)

I love that company. I'm such a fan. I met maybe the co founder, maybe just one of the creator of the original project, like the engineer behind it. And he taught me a lot about software development, which was very helpful earlier in my career. Anyway, how are you thinking about the state of consumer AI? Who's winning? What are the narratives that are being discounted because of the vibe war and just all the different takes that surface every single day on the Internet?

Apoorv Agarwal (57:35)

Yeah, look, I'll start at the top. History has shown us that these consumer markets have been winner take most sometimes winner take all, honestly. Right, you look at the biggest technology market, started the Internet. We got Google Ads there for search. 90% plus market share, 3.54 trillion in market cap.

Sam (57:51)

Yeah.

Apoorv Agarwal (57:52)

Mobile. The next big super cycle led to the birth of Apple. You know, three and a half, four trillion in market cap. Another winner take most dynamic. Take social, the next one, Metta, one and a half, two trillion in market cap across WhatsApp, Facebook, Instagram, they've got majority market share. And so, you know, keep going down this list, ride, hailing food delivery. You know, I think consumer AI, I suspect will be similar. And I think that's what's going on with Chad GPT, you know, it got lightning in a bottle three years ago. It continues to, they continue to work hard, ship features, improve models, memory, voice, image. And now Sora, you saw yesterday, and each of that is adding to the S curve. And you know, lo and behold, ChatGPT is, is doing pretty well despite the vibes.

Sam (58:36)

Yeah, yeah, no, totally. I heard a stat where my jaw hit the floor, which is that Bing is profitable and growing. It's much smaller than Google. But it's a very interesting thing that even in something that feels like a winner take all market, the second third player can still be like reasonable businesses where Satya Nadella doesn't need to say, oh, we gotta shut this down, it's not working. It's like it is working. It's not winning, but it's working. And it adds value to the ecosystem. And so I think that there's these, there's these interesting things there.

Jordi (59:08)

Has anyone ever tried to clock what the actual enterprise value of Bing would be today?

Sam (59:14)

I think it's in the billions. I think it's definitely in the billions.

Apoorv Agarwal (59:17)

Definitely.

Sam (59:17)

Yeah.

Apoorv Agarwal (59:18)

I mean search, search is a valuable thing for agents more than humans, right?

Sam (59:21)

Oh yeah.

Apoorv Agarwal (59:22)

Particularly because the Google API is no longer available as an API.

Jordi (59:25)

So.

Apoorv Agarwal (59:25)

So Bing's API is actually very valuable, maybe even more than the Bing consumer product.

Sam (59:29)

Yeah. And I mean Google for a long time has always said, you know, competition is only a click away. Sundar Pichai said this and it's Interesting how, how sticky consumer habits can be in a world where in tech we often optimize for the benchmarks and the latest features. And I'm jump, I got every app on my phone. I will switch day to day, I will export, I'll jump through all these hoops to use the latest and greatest for two weeks and then go back to the old thing. But when I just talk to random people in the world, it's like they're like, oh. The funny one is Jordy was talking to somebody who was like, he was like, do you use any of these AI tools? And then the guy was like, no, not really. And then he was like, have you ever heard of ChatGPT? He's like, yeah, I use that every day. And it's like, it's not even. Like he doesn't even think of it as AI, because the news, when the news talks about AI, it's like a data center or like a robot. And like he just thinks about like, oh, yeah, that's the text box that

Jordi (60:27)

answers my questions that I can do stuff with.

Sam (60:30)

Yeah, yeah, yeah, yeah.

Apoorv Agarwal (60:32)

What's your decisions outside of ChatGPT? Like, if I'm buying something, you know, I got a button dedicated to it on my phone. Like it's a habit.

Sam (60:39)

Yeah, exactly, it's a habit. And that's powerful.

Jordi (60:41)

What's your framework for trying to understand the present through history? How much do you look towards the dynamics that were unfolding during the dot com cycle, different consumer Internet players? How much does that matter to you? How much do you obsess over that versus obsessing over the data today and trying to predict the future?

Apoorv Agarwal (61:05)

Yeah, look, I think in the moment, if you're sort of living day to day and observing what's happening in the news, it feels a little bit like there's four seasons to the year. It's called OpenAI, Google, Anthropic and China. Maybe those are the four seasons. And depending on what season you're in, that's the news cycle, that's the vibe cycle. But I think to your question, I try to zoom out of that. And obviously we run a very concentrated investment strategy. Our growth fund has only four investments. Right. And so when you're running that concentrated and putting your money where your mouth is, you have to zoom out. And history certainly rhymes in this case. One of the things that I had written about, I'd say two, three years ago was, look, there's two big conglomerates in technology, Meta and Google. Meta's got 3 billion users in their phones in their pocket across WhatsApp, Facebook, Instagram and Alphabet's got that with Google and Gmail and YouTube. And I said, hey, these guys are most likely to deliver the AI that everybody's going to use. I was wrong. That's not how it happened. ChatGPT, turns out, got magic in a bottle and continues to grind up now just under a billion, I think 900 million weekly active users reported over a billion monthly active. And honestly, that was a surprise and we corrected that mistake two years ago and now the largest investment in ultimate history.

Sam (62:32)

How do you think about pattern matching against previous capital wars? I was loosely thinking about the Lyft vs Uber fight that happened and I don't know, it's always hard to overfit on previous cycles. But I'm wondering how you process that capital fight. I don't know. I'm not exactly sure where you were in your investing career at that time, but you clearly lived through it. I was wondering about lessons from that era and maybe if there's a different scenario where capital fights can result in oligopolies. That's the interesting narrative that I feel like might be playing out, at least in some sub markets.

Apoorv Agarwal (63:11)

I was a young engineer at Ballantra at the time and I remember the Uber rides in SF and Palo Alto. Here, what we call the Shire got so cheap you were almost better off taking them than walking. And you know, I think it is not too dissimilar right now, right? Like, I'm sure you guys are on ChatGPT and like you said, it's a habit. And we are certainly getting more than 20 bucks of value on this per month, right? Maybe hundreds. For some people, it's thousands. The folks who are coding, maybe tens of thousands. And I think we are in this era, folks, that the price discrimination has not stuck, right? It's still 20 bucks a month for basically everybody. And actually I'm going to talk to Nick Turley, head of ChatGPT later today about all of that. And this is one of my questions is like, how do we make sure we are for folks who are getting thousands or tens of thousand dollar value out of this product, how do we make sure we capture that what I call the expand the ceiling on the top. But then also there's the vast majority, the billions of users, the Joe Schmo, they don't want ChatGPT to book their flights for them because they're not booking 10 flights a month, they're booking 10 flights a year. They're deliberate about it. And so for them, how do we get them on an easy ramp? How do we lower the floor? And maybe ads is the way to go. And the meta properties have done a great job. They've got 3,4 billion users. They monetize it about $60 per user per year global, globally. Alphabet has a similar number. It's about 3, 4 billion active users monetize at about $70 a user per year globally. And you know, OpenAI is that, you know, we're at about, you know, 10, $15 per user per year across about a billion users. And both have got to go up over time.

Sam (64:56)

Yeah. How do you think about squaring this idea that at least among the Mag seven, maybe you take out Tesla and Nvidia, but basically every company competes with everyone and everything. It feels like I was always laughing about 6 out of the 7 mag 7 CEOs own or control social media properties. Whether you include LinkedIn, Twitch for Amazon, X for Elon, Imessage for Apple is sort of a social network now. And YouTube directly competes with Instagram. The Shorts product is the same and it feels like over time some of these develop oligopolies, some of them, you know, the monopoly sticks. But how much do you think we can learn from the way the previous era of the MAG7 boom played out in terms of how the leading labs might compete over the next decade?

Apoorv Agarwal (65:54)

100%. I mean it's happening right now. I mean you saw what happened and they call it, you get cloud coded. It's a cursor got cloud coded. I think it's the biggest narrative violation. I was with Michael Troll over the weekend. They're crushing it. Oh yeah, they're absolutely crushing it. And it's what you said. Maybe to give a historical example, Databricks and Snowflake, they were both built on the top of the three hyperscalers, AWS, GCP and Azure, while them having their 1P products Redshift, BigQuery and with fabric.

Ivan Soto Wright (66:28)

Right.

Apoorv Agarwal (66:29)

And it's a little bit of what's going on right now with Cursor you've got Anthropic and OpenAI who've got their models that Cursor buys and They've got their 1P products called codecs and cloud code that compete with Cursor. And look, if you're a developer, if you're a customer, an enterprise customer who cares about the resilience and the multi modality and the option of having this choice, you'll probably pick something horizontal across them. And so I think as you said, everybody's Competing with everybody. And what a time to be observing this. But I think the fog of war will get clearer as we go into the second inning and the third inning.

Sam (67:08)

Yeah. You mentioned that flight booking might not be the major driver of consumer growth this year. Maybe I sort of agree with that. But I'm wondering, I was talking to Olivia Moore at Andreessen yesterday about the value that might come from integrating image, video generation and reasoning models, and deep research models all into one consumer interface. So you have this idea for this like essentially vertical video reel and you can actually like plan out, generate images and wind up with a much more full featured product. What else are you tracking in terms of like consumer use cases that you think might unlock this year?

Apoorv Agarwal (67:54)

Yeah, look, I think the, you know, we are investors in a couple of these inference platforms, so we see the traffic of AI tokens. I would say the two biggest sources of the users for tokens is coding and video. Yeah, right. And then there's like a long tail of customer service and legal and office of the CFO and so on. And obviously chat is, you know, I forgot to say chat, but chat's one of the big ones. I think for the average consumer. The beauty of having it all in one place and why I had wished Sora would have been in chat GPD since day one is just so high. You know, I'm not a professional video editor or image editor. I make that one card on my kid's birthday and that might be it, maybe two, and the second one on my anniversary someday. And so I think I would just like it to be there. And that's the vast majority of Joe Schmo. Now obviously the 1% like Dan or producer, he's going to make that 30 times a day and he might go find the best in breed player. But I think for a vast majority, having the one throat to choke is the way to go.

Sam (69:05)

How do you think marketing and customer acquisition will change in the future with regard to consumer AI? We're seeing super bowl ads now. I haven't seen a ton of direct response, but maybe I've just missed that. Maybe Meta's like, hey, we're not going to let someone run these ads.

Jordi (69:26)

But in some ways I'm sure Meta is funding their AI strategy through how much other AI companies are spending on the platform.

Sam (69:35)

Yeah, certainly in the long tail you see a lot of ads for like point solution AI apps for try on clothes digitally or create a custom avatar. And you see those apps sort of climb up the app charts on the back of Meta ads. But what do you think the shape of like customer acquisition looks like over the next couple years?

Apoorv Agarwal (69:54)

It is such a good question because you know, as I said, there's four seasons to the year. But turns out in those four seasons, like 16 mini seasons where one of these app will go through like a marketing surge.

Sam (70:05)

Some people might even say 365 mini seasons if you're running a daily technology, technology and business news show. We're fully employed over here.

Apoorv Agarwal (70:14)

Exactly, exactly. And so you know, we see, we observe these spikes and so there'll be a new app. You know, I mean not to dunk on Deep Seq, but like Deep Seq had a huge spike start of last year and it was so hard to explain. Like I don't know if anybody uses it anymore.

Jordi (70:33)

I just, I just assumed that was bought. They just bought, they just bought it their way up the chart.

Sam (70:38)

I don't know, I think I downloaded the app. A lot of people were interested.

Jordi (70:42)

You downloaded it because you saw it on the top of the charts?

Sam (70:45)

No, no, no. I saw it on Twitter. People were saying like this is the first way that you can see the new user interface for what, watching an LLM reason and you could see the token streaming and you could see the internal reasoning logic and then you get the output. And that was a remarkable UI pattern that then was quickly ported back to ChatGPT. And the ChatGPT client base and customer base and user base was very happy about that. And they still are because you see the little thinking about this, okay, I'm going to search this and everyone's happy.

Jordi (71:18)

Yeah. One thing that's surprising, looking at the app store charts right now you have ChatGPT, Claude, Gemini and 1, 2, 3. But then there's no other new apps in the whole top 25. And it's so interesting thinking if you could rewind two, three years ago and you knew how, how, how much acceleration, how much progress we were going to see in AI, you would assume that there would be like 5, 6, 7, maybe more of these like new entrants. But I think it's, you know, it just, you know, even, even with the technological shift, it's just, it's just proving that like it is so hard to actually break out in consumer. And in some ways it's going to the tools, the AI tools just make it harder to actually break out and chart because you're not competing with two or three other companies, you're maybe competing with two or 300 at some point.

Apoorv Agarwal (72:11)

The flywheel is too good. It's too efficient. The more users you have, the more feedback you have. You know, the thumbs up, thumbs down at the bottom of the your ChatGPT prompt and they make the next answer better. And it's just that flywheel is so, has gotten so good. It's a little bit like what happened in search.

Brian Taylor (72:28)

Sure.

Apoorv Agarwal (72:28)

After a certain point it's just so hard to deliver a better search than Google.

Sam (72:34)

Yeah.

Apoorv Agarwal (72:34)

And which is why these consumer markets, as we discussed, I think are going to be winner take most.

Sam (72:39)

Have you guys heard the story of how Google built their spell checker auto complete? You've told it before, I think I've told this, but basically they just look through the records and they said if someone spells a word that we don't know, just look at whatever they, whatever they searched next because they will likely go do the work for you and actually figure out the correct word that they were thinking of. And then so they just had this massive data set and so they just had the best autocomplete. And that stuff compounds all the time. The leaving the thumbs up, thumbs down. That's the version of like tipping in the consumer AI era. If you're not doing that at home, if you're not tipping chatgpt with some feedback, you gotta get the flywheel going. This is important.

Apoorv Agarwal (73:22)

That's exactly right.

Sam (73:24)

Okay, prediction time. Do you think any Neo labs will try and jump into consumer or are they more interested in playing in like the more oligopolistic B2B point solution enterprise

Jordi (73:36)

where you could potentially as a service.

Apoorv Agarwal (73:40)

Look, I'm a forever optimist on startups, so I hope they take a crack at it. And at the very minimum we will learn something that one of the big platforms will get. Like what you just told us about Deep Seek. I thought that was a good feature and we got it on the main app. At the very minimum we will have that happen and at the very maximum we'll have a new product and ChatGPT will be the Yahoo. But obviously that's a harder hill to climb.

Sam (74:04)

Yeah. What about. I've been talking to a few investors about where they sit versus where the timeline sits, where they're thinking in much longer timelines than what X is going back and forth on. Who's the hot company of the day, the seasons, as you put it. What do you think about this idea of tall poppy syndrome that the tech community seems to like to punch upwards, not punch downwards. And so no one wants to beat up on a company that launched and then never really got to Escape velocity and is sort of like languishing. But if you're at the top, it's game on. We're coming for you and we are gonna find every problem and make it very well known. And it's over. It's over. What do you think about that?

Apoorv Agarwal (74:49)

Well, when you're the underdog, you gotta pick a fight with a winner.

Jordi (74:52)

Yeah.

Apoorv Agarwal (74:52)

You gotta bring the fight to you.

Sam (74:53)

This is alpha, you know. Yeah.

Apoorv Agarwal (74:55)

Not the other way.

Sam (74:56)

Yeah.

Apoorv Agarwal (74:57)

So I think. I think that's what's going on. People pick a fight with a winner.

Sam (75:00)

Yeah, yeah, that happens. That happens a ton.

Apoorv Agarwal (75:02)

Bring the game to you.

Sam (75:03)

Bring the game to you. What else are you tracking in the AI build out? What else is important in understanding? Like, can consumer AI continue to ramp. You mentioned ads, you mentioned continued retention and growth, but more deeper in the supply chain. How are you feeling? We just looked at Oracle earnings. It seemed very positive. But what are you reading?

Apoorv Agarwal (75:26)

You know, the number one thing at this game, like three years in that we obsess about is durability. Yeah, Durability is the number one question. And we have this, like, panel of things that we look at. I've been publishing some of that is, hey, is this thing, you know, users show up. That is step number one. Users stay there. That's step number two. And the way we measure that is daily active users divided by monthly active users. And, you know, I was fully expecting that to tell us a lot. And it does. Actually, the dispersion of the leading AI apps, the three that you mentioned that are leading, you know, ChatGPT, Gemini and Cloud, is actually quite wide. ChatGPT is daily active usage over. Monthly active usage is about 45, 50%. Gemini is at 22%.

Mike Blue (76:12)

Wow.

Apoorv Agarwal (76:12)

Meaning that the number of times people are showing up in a month is like twice as much more on ChatGPT, which means it's.

Jordi (76:19)

So they're still, like, retained. They're still retained on Gemini. It's just a much less valuable user because they're not using the app as much.

Apoorv Agarwal (76:27)

That's right.

Owen Jennings (76:28)

That's right.

Apoorv Agarwal (76:29)

The usage is far less frequent on Gemini. And then the third thing that we look at is, hey, has this turned into a habit or is it still a curiosity? And that we look at using the standard retention chart, which is over 12 months, how many users retained? And the dispersion is quite wide there as well. In fact, the only three apps that show what we call a smile curve, meaning that users come back over time after having churned. The only one of them in AI is ChatGPT. The other two are Chrome and WhatsApp, meaning that the utility is so high that even if you made the mistake of churning, you find your way back.

Sam (77:12)

I mean, that happened to me, not with chrome, but with WhatsApp. I had it for one group of friends, dropped off. Then eventually there's a new group chat, and you gotta become a da.

Jordi (77:22)

You. Chrome seems like a better kind of comp for LLM because it's more of a single player experience. Whereas, like, WhatsApp, I'm the same way. I'm not a heavy WhatsApp user. But then every once in a while I'm like, wow, there's two people I message on WhatsApp. I haven't gotten back to them or whatever, so I'll come back. But Chrome is probably more relevant.

Apoorv Agarwal (77:42)

It's a great point. You know, it's a great point because, you know, WhatsApp has network effects. You know, if my friends Jordi and John are on WhatsApp, I'm more likely to be there. ChatGPT doesn't have that. If you guys are in ChatGPT, I'm not more likely to be there.

Jordi (77:56)

Yeah, the same thing with Chrome.

Apoorv Agarwal (77:59)

Yeah, exactly.

Sam (77:59)

Yeah.

Apoorv Agarwal (78:00)

And the other effect that I think about a lot is the dopamine Doom scrolling through cat videos like TikTok, Instagram, all have this dopamine, this addictive effect. That's why they're 3,4 billion users. ChatGPT doesn't have that. You are there for some kind of a knowledge workflow.

Sam (78:18)

Yeah. They also have the TikTok. I haven't had TikTok installed for years, but occasionally people will send me a TikTok and I will try and open it and it'll just ask me to install the app. It won't even play for me in the web browser. And it's very difficult. But that is another viral flywheel that people often will generate something, but they'll send me a Screenshot from their ChatGPT results. It's pretty rare that someone actually sends me a full link to their deep research report. I think that might happen more. I've said for a long time, like, I want to just have a feed of Tyler, whatever Tyler Cowen is deep researching, I want to just read through the results of that because I think that would be interesting without him needing to publish it. And then everyone be like, this is AI generated. It's like, no, I want to. I want to be able to follow people that are doing interesting research. And then when they. When they do Some interesting research. They can just share that on the network. I think that could be cool as a feed, but again, very knowledge based. Not nearly as brain Roddy as TikTok. So different, different sorts of strength.

Apoorv Agarwal (79:21)

Building on what you said. Here's. Here's a prediction. I think, I think the next phase of utility, guys, we are at like a billion users on this consumer thing. The next billion, I think will come from these proactive, proactive work like you described. It's like, hey, I don't. Right now I've got to go log on to ChatGPT or Gemini or Anthropic to like engage with it. But hey, what if you like, you were like, hey, this is what Jordy said or this is what Tyler Cohen said. You got to check it out. You know, I know you think about a lot about this company, OpenAI. You think a lot about this company, Revolut. This is what somebody smart said about it. Or hey, they got a UK banking license. You got to check it out and here's why it matters.

Sam (79:56)

Yeah, Pulse was like a glimmer of that. But I felt like the daily functionality on Pulse was too much for me. What I'm actually looking for in a world that has slightly different economics, slightly different business incentives is something more like, I like movies. I like a certain type of movie. I want you to actually get through to me when there's a new Christopher Nolan movie or a new Denis Villeneuve movie, and I want you to tell me about it, but only when that happens. And I don't want a weekly newsletter of all the movies. And that doesn't make any business sense. Sense. Like you would never be like, I'm starting a substack. I'll post it whenever something interesting happens. It's like, it's gotta be daily, Gotta be weekly. We have a daily show. We couldn't just do just emergency pods. Like, that doesn't make sense. But with AI, you can appreciate.

Apoorv Agarwal (80:44)

Yeah, you might appreciate this because you brought up Redis. You know, these weekly reports are like, these are like cron jobs back in the day.

Sam (80:51)

Exactly.

Apoorv Agarwal (80:52)

Who knew cron jobs would be so valuable?

Sam (80:55)

But I want the cron job to run and then the AI agent to filter and be like, does this actually pass the. The quality bar? If not, let's just not. Let's just not annoy this guy. And then if it is, if we got something good going on, let's let him know. So tons of interesting things. What a fun time to be a product manager at OpenAI. It's got to be a great time but say hello to the head of ChatGPT.

Jordi (81:16)

What's the bar for adding a new investment to your Guys4 portfolio fund?

Apoorv Agarwal (81:25)

It's not a low bar, I'll say that but I'll keep you posted. I'm working on one trying to get add one. We'll see if we get it through.

Jordi (81:32)

True.

Apoorv Agarwal (81:33)

But maybe to actually answer your question I would say that one of our beliefs is that the rate of change is so fast and so high that honestly it feels like there's a dozen startups that get quite endangered turn into endangered species with every product release from one of what we call private next seven OpenAI anthropic leading that charge and and so the idea is to just stay concentrated in the ones that we know are working and we obviously have an early stage fund that does earlier stage investments but on the growth state side we run pretty concentrated. The bar to get something added would be we have belief that this is a company that has a good line of sight to free cash flow over time and that is hard to say for many late stage businesses right now but. But the ones that we feel so are in there.

Sam (82:31)

That's great.

Jordi (82:32)

Concerning. Concerning that you're not seeing the line of sight

Apoorv Agarwal (82:39)

I think about the private markets in three innings right now guys it's like you've got the early stage stuff that's like sub a billion market cap. The second call it stage is the 1 to 10 billion market cap and then the third one is the 10 to call it wherever you're staying private forever. This is what I call like quasi publics. You could go public, you just choose to stay private. I would put OpenAI anthropic revolut ramp in this bucket. Very high quality companies that we'd love to own in our growth fund.

Sam (83:08)

Sure.

Apoorv Agarwal (83:08)

The second bucket which is the one to ten I think is the part where there is the most dispersion and it's probably the place I see the most fog of war and hardest to figure out whether they have the durability and you know the early stage business will always be there. There'll always be great founders starting. I know you guys had I think you call it the Gupta nator Ravi Gupta legend crush.

Sam (83:36)

Yeah.

Jordi (83:38)

What do you last question then we'll let you go. How are you processing the disconnect between public saaspocalypse people concerned around SaaS as a business model and then what's happening in the actual early stage with there's so many companies that may position themselves as something other than SaaS. But functionally they are SaaS. Like you could just, you could just take, if they say agents, you could just replace it with workflows and it would sound exactly like a SaaS business. And it's just like very interesting time where you just have all this funding activity early, early, early. And then public investors are saying, hey, we want to put a huge discount on all these businesses because we just don't know what is going to happen in 10 years. Do you find that a bit strange?

Apoorv Agarwal (84:26)

Look, I have tremendous respect for the public markets. They are a very efficient mechanism for repricing, but I think they're not hyper efficient in times like now when the change is so fast. And so I think there's a spectrum. You know. I'll ask you a provocative question. Why couldn't you vibe code chatgpt?

Jordi (84:48)

You could.

Apoorv Agarwal (84:49)

And by the way, when you vibe

Sam (84:50)

coded and come back, we did. We launched Claude with ads. Claude with ads with Claude. Oh wow. Claudewithads.com was up for one day during the super bowl and it gave you intelligence too cheap to meter, sponsored by all of our sponsors and it served you an ad before delivering a query from Opus 4:6 or something like that. It was a lot of fun. But yes, of course it was not a real product and it would never have liked.

Apoorv Agarwal (85:16)

But I think that's the state of a lot of a lot of vibe coding right now. You go build cloudwithouts.com, you try it out for a day and then you're like, oh man, that real cloud was a lot better. And then by the way, go buy some software in the public markets after that because you realize the staying power of some of that software. I wouldn't say that for everybody. Right. Like there's a lot of stuff that we have on our list that is truly endangered in the public market. So I think it's pretty case by case, but some of that's pretty good.

Sam (85:43)

Yeah, that makes a ton of sense. Well, thank you so much for taking the time to come chat with us. Have a great rest of your day. Safe travels. We'll talk to you soon.

Jordi (85:50)

Cheers.

Sam (85:50)

Goodbye. Let me tell you about Railway. Railway is the all in one intelligent cloud provider. Use your favorite agency, deploy web apps, servers, databases and more, while Railway automatically takes care of scaling, monitoring and security. And let me also tell you about Gusto, the unified platform for payroll, benefits and hr built to evolve with modern small and medium sized businesses. And without further ado, we have Owen Jennings from Block. He's the executive officer and business lead, Owen.

Jordi (86:16)

How are you doing?

Owen Jennings (86:18)

Hey, how are you? Thanks for, thanks for having me, guys. Big fan of the show.

Sam (86:21)

Appreciate that. Thank you so much.

Jordi (86:23)

To finally have you on.

Sam (86:25)

You guys know each other, but for those who don't, can you kick off with a little bit of an introduction on yourself, how you wound up at Block, how long you've been there, what you do on a day to day basis?

Owen Jennings (86:35)

Sure, yeah. I joined block about 12 years ago. I was on the Square side.

Sam (86:39)

Overnight success. That's awesome.

Owen Jennings (86:41)

The business first and I worked on scaling the business team before the IPO and then in about like 2016 I shifted over to Cash App. Cash app was like 25 people at the time. It was just peer to peer. Spent a while on the, on the Cash side and then we functionalized the company about two years ago. And so I look after the business team, which is product and a number of our, of our operations teams for, for Block overall across our brands, which are Square Cash App, afterpay title and so on and so forth.

Sam (87:16)

What was functionalizing the company mean for those who might not know?

Owen Jennings (87:19)

So we used to have pretty rigid business units where Square and Cash App operated pretty separately and they each had a CEO and there was kind of duplicative disciplines between them. I think part of our core strategy now is connecting the seller side and the consumer side. So having like both sides of the, of the transaction. And so about 18, 20 months ago we functionalized, quote unquote, which just means we're operating at the block level. So the product teams across all brands for all of Block roll up into me and same for engineering and design. So it just lets us have centers of excellence or for each discipline and be able to more flexibly move resources throughout the company versus like really rigid boundaries.

Jordi (88:07)

Before we get into the present, tell us about the chaos of that 2016 era. Scaling cash App that was like truly insane. Hyper growth from everything that I've heard and I know a lot of teams are going through that style of hyper growth right now in consumer AI.

Owen Jennings (88:26)

It was incredible. I mean, I look back on it very fondly. I think the key thing, if you go back far enough to like 2015, 2016, there was actually a meaningful question of if we should continue investing in Cash App. Like there were executives at the, at the company at the time it was Square Cash, but there were executives at the company who were saying, look, you know, Venmo already has this thing on lock. This doesn't make any sense. You haven't figured out a path to monetization. And so it was actually Jack, our CEO, who continued to push and continued to invest. Sarah Fryer, our old CFO actually gave us a deadline at which we had to figure out monetization or else we were going to be shut down. And we were trying to go through the IPO and so I kind of get the trade offs there, but then like very, very quickly found market fit. The key difference for Cash App back then was the way Cash App initially started was you could move money instantly from bank account to bank account. That that wasn't something that was available with any other sort of peer to peer functionality. And so we found market fit really, really quickly and the tremendous, tremendous growth and like incredibly small team and just, you know, working to make sure we could stay up and we could serve customers every day, especially Friday.

Jordi (89:46)

Jacoby was on oc.

Ivan Soto Wright (89:48)

No way.

Sam (89:48)

No way. So that was fundamentally different because I had other money transfer apps at the time. But the flow was always take some money out of your bank account, put it into the app and then send that money from that app to my friend and then it sits in their account and then they can take that to their bank account. You could go bank to bank, basically.

Owen Jennings (90:07)

It's actually really funny how we got here. So we were thinking about building how we build a low cost payment network for buyers. And there was this one type of transaction that was pretty unique and it was an unlinked refund. So for a square merchant, it's like, how do you refund money back onto the card? And then we were like, oh, we can just process these as unlinked refunds. We worked with Visa, we ended up turning into this whole Visa Direct thing. And now so much is happening on debit. But what we would do is there actually wasn't even an application to start. You would cc cashquareup.com and if your debit cards were linked on either side, you would move money instantly from bank of America to JPMorgan Chase or what have you on debit rails. And so it was instant and it was free. Things have evolved a lot.

Sam (90:57)

Yeah, yeah, of course. Well, I mean, with that functionalization of the company, is that one lens to view the change in workforce size as just a continuation of that, or is this some sort of key strategic pivot? Like there were a lot of different messages going out. I'm wondering how you're thinking about the size of the company, the workforce over the next few years.

Owen Jennings (91:20)

Yeah, I can kind of paint the picture on how we got here. I mean, the primary driver for this decision was around how AI tools are evolving and how they're flowing through our work. So when I think about software development over the past year or so, I would say there was pretty meaningful progress from, let's say, you know, the start of 2025 through about November. And it was flowing through our work and kind of changing how we were approaching things. We've been pretty early here. Like we launched Goose a couple of years ago, which is our open source agent harness. I think it was the first agent harness out there or that I know of. We worked on the MCP with Anthropic and OpenAI. So anyway, things were progressing and then in like the last week of November, first week of December, things just fundamentally changed. And it was with Opus 4.6 and it was with Codex 5.3 and we basically crossed the chasm where like, I think the way AI tools were flowing through on the development side before is like, it's a useful tool to help a given engineer be more productive. You can autocomplete, so on and so forth. In like late November with the model changes, we got to the point where these agentic systems were actually able to write the code autonomously and the code was good enough to ship into Prod and that was like a huge change. And then, so we spent, I think many of us probably spent December and the holidays, like playing with the tools, playing with Claude code, playing with Goose. And then we spent Q1 thinking, okay, well now how does this flow through to a technology company and a software development company? Because everything has fundamentally changed over the past four months. And so the Org changes were mostly a reflection of that. I think that just the fundamental nature of like the shape of the organization that you need, the size of the organization that you need, the workflow for, an engineer, a designer, a product manager, other disciplines all fundamentally change. So, you know, we've continued to kind of refine the size and shape of our org, but. But this was definitely a reaction to agentic development and what that means for technology companies.

Jordi (93:31)

How have expectations around just productivity for developers at Square changed? How do you try to understand if somebody is actually using the tools to the fullest extent, what's the general philosophies?

Owen Jennings (93:52)

I think, I think, I think basically expectations around productivity are increasing across the board. I think it's the clearest on the development side. And that's also why when you look at the reduction in force, it actually overindexed to product and design and engineering relative to other roles. And so we track all sorts of the normal things that you would track, like PR throughput. And we have an internal concept of time to customer value. And all of those things, like time to customer value is compressing. The number of PRs that we're submitting is increasing massively. PRs per engineer. The main bottleneck now is code review because shipping a PR is becoming more and more trivial, especially for simpler features. But I guess I would say we're kind of in the belly of the beast right now. I think a lot of software development companies are where I think as high as expectations are. I don't think that they're high enough. Like we're seeing just a fundamental shift here where, I mean, look, the basic way that development used to work is that you would have a product manager, a designer, six to eight server engineers, two to four client engineers, and you would kind of sequentially work on a feature in kind of a waterfall fashion. A medium sized feature would take four or six weeks. You would ship that, move on to the next one. That's fundamentally different now. Now we have really small squads of two people, three people who are way more full stack. They're all on the tools, as we say internally. And what you're able to do is just like orders of magnitude different. Obviously there's like the, okay, are you running five or ten instances of Claude code at once? And how does that, that flow through to your day to day? I think the bigger changes are just like the fully agentic systems. So we have Builderbot internally, which is somewhat similar to Claude code. It's built on top of Goose. And so now anyone.

Jordi (95:57)

I love all your codenames. We got Goose, we got Builder Bot. What else you got? What else you got?

Owen Jennings (96:03)

Well, I think Goose was named Goose too, because Top Gun of the. Yeah, exactly, Top Gun. Little cheeky copilot shout out there. So for Builderbot, we have a Slack integration. So I can just add Builderbot and say like, hey, fix the spacing on this screen in Cash app and then like, you know, put my computer down, go have a slice of pizza with my kids and then come back and there's a beautiful PR that's waiting. And that's just like a complete paradigm change. And it's happening across everything, not just the development side.

Sam (96:39)

I don't know how much you've been tracking the AWS story, but it feels like Amazon might have had some uptime problems or some backlash to the amount of vibe code that they were pushing to production. How are you thinking about avoiding the downside Scenario where more and more of the system becomes vibe coded, fewer and fewer people understand everything and something goes down and no one really knows how to fix it. Uptime starts dropping There's a variety of reasons why we're seeing uptimes fall off. Some of it is just CPU shortage. But how are you thinking about systems reliability in this new paradigm?

Owen Jennings (97:17)

I think there's a couple of different pieces. So one for the reduction in force itself, the number one principle was stability and reliability. So we like blank sheet of paper, we were like, okay, that's a P0, we're not going to sacrifice that and let's start building back the team with that in mind. I think more generally how we think about this is that there's just a spectrum of where you can be more risk seeking and where you want to be more risk averse. And so there's a bunch of examples right now where a product manager on the Bitcoin team at Block is built a fully functional app related to Bitcoin and stablecoins as kind of a proof of concept prototype and has done that really without an ide, without looking at code. And that sort of thing can be incredibly helpful. Similarly, for an internal tool, you're probably way more risk seeking in terms of, okay, let's, you know, you don't need to ask for permission for everything and you can push this change. And then obviously the further down the stack and the more load bearing something is, we're way more risk averse and I think that makes sense. Like our financial platform team for instance, they run reconciliation and treasury and card issuance. And our cloud platform team, it's not the case that we're being willy nilly in terms of how we're approaching it. It's just that those who are experts and our system architects and our principal engineers themselves who have deep context are able to become 2x3x5x more productive. But we also have humans reviewing every bit of code before we ship it into prod. So I think you need to be prudent in your approach here and protect the downside. But still the productivity gains are pretty obvious, I think.

Jordi (99:10)

Where else are you guys getting a lot of value out of AI outside of CodeGen? Is anything on the fraud side or increasing the efficiency of lending? I'm just kind of spitballing, but where are you most excited?

Owen Jennings (99:27)

It's pretty broad based and so I'm not going to walk through the entirety of the org, but I'll give some good examples. I think customer support, customer service is a Clear one where we have our own models that we've built and we've trained and we've been able to automate like 75 to 80% of check inquiries across brands. And that's without sacrificing customer satisfaction or csat. I think that's like a, you know, that's been, that's been out there and I'm sure y' all are seeing what Sierra and Decagon and folks are doing. I think basically any deterministic task, we're able to offer a ton of leverage to the folks who are working at the company. So a lot of folks on the, on the operational side are, they're working queues. So you might have a queue to say like, hey, does this person, should this person pass identity verification or should we file a certain ticket related to this transaction? Or hey, anomaly detection? Let's look at this. We have a human in the loop, which I think is critical, especially in a highly regulated area like financial services. But a lot of the pre work and a lot of the context and a lot of the pattern matching we're able to do with, with tooling. And so it ends up that a given individual on that team can be massively, massively more productive. I think the other thing is just we've built generalized tools that are quite helpful across the Org. So another code name for you is G2, which is, I guess it's Goose 2. I don't know, it's Goose's Goose 2 brother. This is like, you can think of it as like an agentic operating system that's rolled out to the entire company. And then we have MCPS into all of the tools that we use. So into Snowflake, into Tableau, into Gmail, Calendar, docs like pick your favorite piece of software and then it's all one interface. And then I can go in and create agents or create automations that do whatever I want. So instead of spending an hour every morning looking through every single dashboard seeing what's going on with, with Square and Cash app and afterpay I can just have an agent that runs at night, does that and then lets me know if anything is off. And you can kind of extrapolate from there all of the different use cases where we're basically allowing everyone at the company to build automations on top of all of the different sources of truth that we have internally. And then we're doing the same thing for our customers on the Square side and the Cash app side as well.

Jordi (102:00)

Very cool.

Sam (102:01)

Well, congratulations on the progress and thank you for coming to explain some tumultuous times. But it sounds like there are green pastures ahead and lots of new things to build and we're excited to follow along. So thanks for coming and breaking it down.

Jordi (102:16)

Yeah, great to see how you going.

Sam (102:18)

Have a good rest of your day. We'll talk to you soon.

Owen Jennings (102:20)

Thank you.

Sam (102:21)

Goodbye. Let me tell you about Applovin. Profitable advertising made easy Easy with Axon AI. Get access to over 1 billion daily active users and grow your business today. I Love this post. 15,000 likes. Catchy opening, but the comma splice is unwieldy. Why don't you try it was not just the best of times, it was also the worst of times. And the comments here are fantastic. And honestly, it wasn't just the best of times, it was also the worst of of times. And that's rare. Let me check the times to see what we're working with. I see it now. This project is building the best of times. After a thorough examination, I have identified several worst of times. If you want, I can fix those next. This is so good.

Jordi (103:07)

Well, we've got to get. We got to talk about Jeremy Geffon Simulator.

Sam (103:13)

Demand was overwhelming. 58 likes on Jeremy Giffon Simulator. But you can play it at home at Jeremy Giffon Simulator. It's Giffon Simulator.

Jordi (103:24)

This was the Hayes Paradox in action, right?

Sam (103:26)

This is Hayes Paradox.

Jordi (103:27)

Internally we thought this was potentially the most funny, entertaining product released and the market spoke and disagreed firmly.

Sam (103:35)

This may be a little too niche, but you've been waiting for Tyler Cosgrove built this in something like a day

Jordi (103:42)

and you get to play as Jeremy going on invest like the best. One of the most iconic episodes of a podcast.

Sam (103:49)

It was the best. It was the highest rated or most downloaded episode of that year. So it truly is.

Jordi (103:55)

Which is insane when you put it in the context of the other guests that appear.

Sam (103:59)

So many incredible guests.

Jordi (104:00)

Look at the view.

Sam (104:02)

Wow. Oh, yeah, yeah. What's going on with that? Did you not figure this out?

Jordi (104:05)

Yeah.

Sam (104:05)

Why is production team trying to out me? Yeah, they're really trying to show you all the flaws of this thing. But of course, when you go to gifonsimulator.com you will be dropped into a fully 3D world. There are some of Jeremy's best posts available on the wall. Of course, a billboard that we created for him. If you have a term sheet accident, you gotta call him Gif on and co. Easy recap.

Jordi (104:30)

And of course the COVID posts on the wall.

Sam (104:34)

There's a special surprise so if you play the game and you answer the questions just like Jeremy did on the show, can we actually go play the game? Do we have audio on this as well?

Jordi (104:42)

We're going to talk a ton about

Sam (104:44)

all sorts of different topics. You wrote an essay, so it asks you. It's about the. Every question that Patrick o' Shaughnessy asked Jeremy Giffon. On this particular Invest, like the best episode, you are given three potential answers. One is the correct answer. Two are AI generated variations that are incorrect but plausible. They're actually pretty well written. All of them are very, very good. You really did scroll through the entire. The entire timeline, and then you can listen to Jeremy give his answer or you can skip to the next question. How many questions are there in total? There's 48.

Mike Blue (105:18)

48.

Sam (105:19)

So to play this full. To play this full GIF on simulator, you could easily be there for an hour. It took me like 20 minutes, and I was trying to breeze through.

Jordi (105:27)

It would be one where you actually never have to talk to any of the people or know anything about the business. You can just understand the.

Sam (105:36)

So it's something like Green Blatts. You can be a stock market genius. Your favorite book on investing ever, Is it chat? What do you think? Is this Jeremy's favorite book on investing ever? I'm going to go yes. Well, I think the answers are actually all yes. They're all yes. But you have to give different reasons. Okay. Okay. You have to give different reasons. Yes.

Jordi (105:57)

That is my favorite book on investing ever. The big thing that really turn a light bulb on. For me, this is important.

Sam (106:03)

If you're potentially index. Potentially applying for a job to work with Jeremy, you and I have talked a lot. You. This is a great way to get your reps in.

Jordi (106:12)

So if you go through the whole

Sam (106:13)

thing and you answer every question correctly, there's a special prize at the end. Yes. If you get them all ready. And has that prize changed since we last talked? It's the same prize I think I sent you. Okay. Okay. Well, for those who are listening, go to gifonsimulator.com you can beat Tyler's personal record, which is 40 out of 48 questions. That's pretty good because you built this thing. It is challenging. Fortunately, Jeremy Giffon is in the chat.

Jordi (106:40)

You gotta beat your own game.

Sam (106:41)

I mean, I could obviously cheat, right? I know, but I tried to do it annually. But this is reviewing game tape, and one day there will be one of these for every podcast appearance in the world. Let me tell you about cognition. They are the makers of Devon, the AI software engineer. Crush your backlog with your personal AI engineering team. So, breaking from J.P. morgan, J.P. morgan

Jordi (107:02)

forced to mark down loans and has decided to reduce lending to private credit groups. Let's get into it.

Sam (107:09)

Devaluations of collateral will limit credit to firms that have become top lenders to higher risk companies.

Jordi (107:16)

JP Morgan has clamped down on its lending to private credit groups, with bankers looking to cut risks as concerns mount over the credit quality of companies in their stables. The bank informed private credit lenders that it had marked down the value of certain loans in their portfolios. This is crazy JPM marking. They're looking into your portfolio and being like, nah, we don't value this like you do.

Sam (107:37)

And who runs JP Morgan again? Jamie Dimon. And what was he doing with the private credit bros like three months ago? He's taking shots. He was saying cockroaches. He was saying that there are jitters in the private credit markets. Well, this is his firm, JP Morgan actually taking financial action against the private credit portfolios.

Jordi (108:02)

The loans that have been devalued are to software companies, which are seen as particularly vulnerable to the onset of AI.

Sam (108:09)

So interestingly, it's less about data center construction. So within Blue Owl's portfolio you have software companies, but then you also have data center construction. And the data center loans, a lot of those look really, really good because yes, it's a big project, you know, a full gigawatt or 500 megawatts for a meta campus, but the company that's financing and building that, that data center that then takes that, issues these loans, they have Meta as a tenant and Meta is going to pay their data center bills even if there's a pullback in the AI market because they have like the greatest ad engine ever created and they have been printing cash flow. They are of course drawing down on that cash flow. But even if they, if there's a jitter in the AI markets or demand weakens, they are seen as a very, very good tenant for those data centers.

Jordi (109:03)

So somebody was able to find a slide from Blue Owl's fundraising materials. Probably not intended to be shared externally. They say this is a real slide from Blue Owl. While it reads like satire, it sadly isn't. The strategy lend to growing SaaS companies with some degree of AR, even if they have negative cash flow or EBITDA. And people wonder why OWL is down 40% year to date. Their underwriting standards are total garbage.

Sam (109:32)

People are not happy.

Jordi (109:33)

Yeah, again, sort of a concept stage slide is not necessarily their Actual underwriting strategy. But the slide says it's talking about the recurring revenue loans. A loan made to a company that may not be currently EBITDA positive because it has made a strategic decision to postpone profitability in favor of acquiring customers that will generate a high lifetime value. Of course people are just questioning what the lifetime value of these customers will be. Fast growing businesses with a highly stable base of existing customers, strong revenue visibility and attractive unit economics. Attractive credit characteristics I'm going to include minus profitability including covenant protections, lower loan to values and premium pricing as compared to a typical direct loan. They may not be generating cash flows but have attractive business attributes. Regulatory lending guidelines limit bank participation.

Sam (110:30)

Yeah, lbovini gave some extra context to what this was, what the market was like a year ago when a lot of these loans were made. He says sounds like you've never worked in leveraged finance or tech M and a hindsight is 2020 but this absolutely made sense 12 months ago and is how you were competitive on direct processes. So companies went out to raise money in the debt markets and if you didn't follow these standards, you were not competitive, you were not winning the business. And he says the premise of the entire software ecosystem is how scalable it is. Not to mention covenants and conversion to ebitda based metrics. T 1/2 into cycle.

Jordi (111:11)

Well, yeah, again all these people were just racing to accumulate the as much AUM as they possibly could because AUM is what gets you.

Sam (111:21)

And also we'll see. I mean like there's a whole bunch of scenarios where this still works out. Obviously the stocks have traded down, but one markdown from JPM does not make a full crisis, a full collapse. Let me tell you about TurboPuffer, serverless vector and full text search built for first principles and object storage. Fast 10x cheaper and extremely scalable. And let's move over to our talk car segment.

Jordi (111:44)

BMW is slated to release its own Mercedes G Class rival soon. We have an exclusive render as to what it could look like. Gabe says wtf is this garbage?

Sam (111:55)

This is garbage.

Jordi (111:57)

Yeah, this thing looks really, really rough. It looks like, it looks like kind of like one of those Lexus Ex things. Very weird front nose. The nose. The nostrils on this thing are rough on people.

Sam (112:12)

But this was, I believe Doug demuro was saying that the BMW XM which is their top of the line most expensive suv, which has been sort of panned because it's overpriced, underpowered and it doesn't quite fit in the same lineage as the Lamborghini Urus, the XM has not sold. A lot of the opportunity for BMW was to go off roading, create some distinct silhouette, figure something out in the G Wagon category. This seems like one step closer to that, but probably not the finish line if Gabe is in the market. But there are more car announcements going on.

Jordi (112:58)

Yeah, I don't know, I don't like the way this looks but I can see it doing decently given that for that customer segment wanting something that feels a little bit more rugged than a BMW.

Sam (113:12)

The hard part about the G wagon is they've gotten very expensive and it has such a distinct silhouette that it sort of takes that off the shelf. Like if you do a carbon copy of it, sort of like what the Ineos Grenadier did, BMW would be shackled with the copycat moniker. And so this, it doesn't feel, I

Jordi (113:36)

would argue, I would argue that the Neos looks nothing like a G Wagon.

Sam (113:41)

It's more of a similar, it's more of a similar silhouette than this. Which one looks more like a G wagon? The Ineos Grenadier or the BMW G Class?

Jordi (113:49)

The Ineos, sure.

Sam (113:50)

Right.

Jordi (113:51)

But again like this, the BMW looks like, closer to like a slx.

Sam (113:56)

Yes. This looks more like an suv. Whereas I would put, I would put Ineos Grenadier G Wagon all in like the Jeep shape. They're all like Jeep shaped. This is a different shape. This is an SUV shape.

Jordi (114:09)

Yeah.

Sam (114:09)

And so the question is like should BMW be doing something that's broadly Jeep shaped? And I think a lot of car fans would say absolutely yes. But they have to figure out a way to, to do it in the way that maintains the BMW brand. Anyway, we will keep following this and see where it goes. Over in Mercedes Land Motor 1 Shares say hello Mercedes Benz VLE, the electric luxury van coming to America. And this is huge because the other electric van that's been recently released but not done as well is the, the revival of the Volkswagen bus, the VW ID Buzz which had a very short range, I think around 300 miles. This has 434 mile range so no range anxiety on those long family trips. It has a long and short wheelbase versions. The VW ID Buzz they only brought the long wheelbase version to the United States. That was sort of seen as a miss from some commentators. Has 6 or 8 person seating capacity, an optional 8k cinema style screen for rear seat seat passengers so all the kids can enjoy a movie.

Jordi (115:17)

I love it.

Sam (115:18)

This is amazing. I love it, I love it.

Jordi (115:20)

I think it's going to be a hit. I like it. I like the brown chrome. Ish color on the monoblocks. Looks crazy.

Sam (115:27)

Yeah, that looks good.

Jordi (115:28)

This thing is going to be.

Sam (115:29)

And interesting placement because it's.

Jordi (115:32)

Minivans are. Minivans are apparently like very, very back.

Sam (115:36)

Totally.

Jordi (115:36)

They fell off for a while.

Sam (115:37)

Matt Grimm at Anduril has one. They're also. This is not Maybach badge, this is not AMG badge. So you would expect the price to be a little bit more in line with an E class, maybe an S class, but not something that's $200,000. So we will continue to see where that goes.

Jordi (115:57)

And one news iconic quote, Ford CEO Farley. They'll have to rip the Mustang stick shift out of our cold, dead hands.

Sam (116:05)

This is a big pivot for the guy who was singing the praises of the Chinese EV companies. He's sticking with the manual. So good to see some news that the Mustang will remain manual, at least for the foreseeable future, while CEO Farley is in charge over there. Let me tell you about Vibe Co where DTC brands, B2B startups, AI companies, advertise on streaming TV, pick channels, target audiences and measure sales, just like on Meta. And without further ado, we have Amjad from Raplet in the restream waiting room. Let's bring him into the TVP at Ultradome. Amjad, how are you doing, Billion Dollar Man?

Jordi (116:39)

Good.

Amjad (116:39)

Good to see you guys.

Sam (116:40)

Congratulations. Tell us what happened. You got some news for us? We got it gong.

Jordi (116:44)

Yeah.

Amjad (116:45)

So we launched agent. Fourth generation of replied agent.

Sam (116:51)

Yeah.

Amjad (116:51)

And we raised $400 million on a $9 billion valuation.

Jordi (116:57)

Let's go.

Amjad (116:59)

Been waiting for this.

Jordi (117:01)

Yes, Absolutely massive. Do your. I feel. Do all your fundraisers leak? I felt like in my head this had already happened. Like, I feel like somebody's on to you. They always front run you.

Amjad (117:14)

Yes, yes. Everything. It's not leaking from us internally, but

Jordi (117:19)

investors, so many different investors.

Amjad (117:22)

Silicon Valley is sort of like the ancestral village. It's sort of. It's all gossip, you know, everyone's gossiping all. It's really annoying. But, you know, it is what it is.

Jordi (117:31)

Yeah. Well, it's great when it's news like this. Talk about the progress at the product level and then I want to talk about the business.

Amjad (117:41)

Sure. So, you know, Agent three, we focus on this idea of autonomy. And it was, if you remember back to the olden days of August, September last year, agents weren't running for all that long, five minutes. People were excited about that. Agent 3 was the first agent to run for, like, hours. And since then we've had the autonomous revolution with OpenClaw and many other things. It kind of entered the public consciousness. And since then we've been asking the question, if we just take for a fact, that agents are going to be doing a lot of our work, a lot of our manual sort of technical work, what is the role of a human? And as we observed sort of internally, people are spending a lot of time on design thinking and systems design. They're spending a lot of time on whiteboards. Whiteboards were never used at ratla, and now they're used all the time. And so as we start thinking about Agent four, we're like, okay, we want to make it around this idea of creativity, because we think that what's left for us as humans to do ultimately is just be creative and build new things and figure out what's the next thing to build, what's exciting to do, do more exploration, build a lot more things. Replit Agent 4 is, the flagship feature is that there's a canvas, and inside that canvas is basically everything that you're building. You can generate designs and mockups, you can build your backend and web app, you can build your automations and agents, you can build your mobile app, you can build a slide deck. It's sort of your cockpit control panel for your business. And there's a lot of technology that went into making all that seamless. A big part of it is parallel agents and collaboration. So every time you start a task, we fork the entire virtual machine and database and everything, create an entirely new environment where an agent is running autonomously on that, but all the while you're just designing and you don't really care about what's happening under the hood. And so we think it's an entirely new way of not just vibe coding, but working in general, because it's affecting every department of the business, from marketing to sales to everything else, because everyone, like coding, has become the new interface for knowledge work. And this is where Replit's headed as well.

Sam (120:06)

Can you talk a little bit about the top of Funnel? You're a master of going direct. I mean, you're one of the few tech CEOs that you've been on. Joe Rogan, right?

Mike Blue (120:15)

Yes.

Sam (120:16)

That's insane. Congratulations. But what is the mix? How do you think about actually growing the top of Funnel at a time when there's a knife fight for every new AI user, But there's also 10 new AI users coming into the category every day?

Amjad (120:37)

Yeah, we Actually haven't invested. I think something that could be a superpower can be also a bit of something you over rely on. So me going direct and just using my platform, we haven't really built out the marketing function as much as many of the other players in the market. So we're, we're just starting to do that. The agent for launch, you might have seen we've just really upgraded our style and how we talk about our product. We just did a brand refresh. So you know, the funding is also recognizing like $400 million is a lot of money. Especially since, since we really haven't made a huge stance in our previous fundraising. But it's recognizing that there's gotta be a knife fight. Luckily, replift continues to grow very strongly organically, especially in enterprise adoption.

Mike Blue (121:23)

Sure.

Amjad (121:23)

Like the enterprise adoption has been really fantastic. I saw you guys talking about Mercedes. Mercedes is a customer.

Sam (121:28)

No way.

Amjad (121:28)

It's really mind blowing how many industries are just jumping on this. Financial services. Amex is a customer, you know, FinTech, PayPal, Plaid. As a customer I really can't think hotels. I can't think of anything that any company that is not really impacted by this. So the growth continues to be strong. But yes, I think there's so much money pouring into ads and all of that. So we have to play that game too.

Sam (121:56)

You mentioned a brand refresh. What percentage of repl.users do you think understand the name now?

Amjad (122:06)

Probably rounds down to zero.

Sam (122:08)

Really because you have abstracted away the read eval print loop. Right. So no one knows what a REPL is anymore. It's going to go away.

Amjad (122:19)

But in spirit it is. I mean a big part of Agent four is recognizing that agentic coding actually was an amazing improvement.

Owen Jennings (122:30)

Right.

Amjad (122:30)

But it made it so that computing went back to the mainframe era where you type a prompt and then you sit back and wait for the prompt to kind of execute. We're like, yeah, that kind of sucks. That really takes us away from that repl, you know, vision of like rapid iteration. So Agent four, the idea of like parallel background agents brings back that iterative loop. So I think it lives in, in spirit. I don't think people.

Sam (122:55)

I still think it's a great name even if it's just like a little if you know, you know, for people

Amjad (123:01)

it's like homage almost to program.

Jordi (123:02)

Before the agent coding during the fundraise process, how did you talk with investors about kind of how you foresee the dynamic between replit and labs? You guys like are very work well together, but at the same time compete more and more on a, on different dimensions. Like how do you see that evolving? I think to date it's like the market has been just expanding so rapidly that companies can be competing with each other but also working together. And everyone's growing. But where does this go?

Amjad (123:41)

It starts with this idea that we live and breathe our mission of empowering people to make software.

Brian Taylor (123:49)

Right.

Amjad (123:50)

I mean, I've been working on this for what, 15, 20 years now. You know, at Codec the idea was that anyone should be able to make software. I mean, you know, I've been obsessed with this for so long and the team, now everyone's here just to solve that thing. If you're at OpenAI and anthropic great companies, but you're also trying to cure cancer, improve math and do all this other stuff, which is amazing and I really hopeful that they can succeed at that. But the kind of focus that we have on the just general population to be able to be empowered, no one else has. I think, you know, they're doing great, but you know, they're massively subsidizing their services in order to acquire users, which is fine, that's what you got to do. But I think if you go to repl.it's really the best place for you as a creative, as a business person, as a solo entrepreneur to build, build your business. And the feature gap, especially after Agent four between us and anything else is, is really, really wide and we'll continue to make it wider. I mean Agent one was really the first, you know, consumer approachable agent on the market back in 2024. That was before cloud code. And so we have a track record of continuing to innovate and you know, moods don't matter anymore. Maybe you can have a lot of capital and you can subsidize a lot of things and you can spend a lot in marketing, but at the end of the day, the market will rationalize and the best product will win. And that's been the case in the history of Silicon Valley.

Jordi (125:21)

I think given, given that you're not trying to play the same kind of capital war game, how, like how, what's your guys philosophy around like trying to create the best product experiences, but also not, you know, even for the users and for yourself as a platform, you don't want to just be generating the most number of tokens. You want to be sort of as efficient as possible and get the user what they want to do, which is build something, but do it in a way that is not overly Costly for the platform or the individual user.

Amjad (125:57)

Yeah, there's a lot of innovation behind this. One advantage of being at the infra or app layer of AI is that you can use any model. For example, we utilize a lot of Gemini Flash, which is like a great model, like code search at design. So we find a lot of efficiencies from being able to create this society of models, which is a thesis that I wrote back in 22. It's not going to be one single model. Every model is going to be better at different things. There was a moment of time, we're like, oh, you know, opus is going to be able is AGI. And it's still, you know, when you go to rapid, it's really four or five different classes of models doing different things. So that's an important piece of it. The other thing is like internally, we don't set air our metrics. I think there's a lot, there's a lot of focus on AR in AR in the age of AI. I don't think it's ARR in the same way that a SaaS business was ARR because you're sort of like you're selling these tokens and yeah, you can push a bug and get better ARR because you're spending more tokens. And so it becomes a bit of optimizing for the wrong thing. So, for example, we look a lot at sentiment analysis, like how people are actually when people get angry, they'll tolerate the agents that they're angry. So we'll look a lot at that, look at success, deployment. We're increasingly looking at how much money people are generating when they're building a business with replit, because now you can hook and stripe and other payment providers. So it's really about what you measure. And we're increasingly trying to measure the quality of the output as opposed to whatever error metric there is.

Sam (127:46)

What advice are you giving to the youngest new hires at replit these days about skill development, where they should be focusing, how they can build a career at a company that is growing really significantly and will probably be around for decades, if not centuries.

Amjad (128:06)

Well, first of all, the roles are collapsing. We have designers shipping code, we have engineers shipping design. We have salespeople shipping code. Yeah, we have engineers doing sales. Our vision for the future of work is that everyone has to be an entrepreneur and we're going to be all massively levered by agents doing all sorts of things. So the particular skill is not the bottleneck anymore. It is how ambitious you are, how generative you are, how creative you are and how good are you at utilizing these tools. And there's been a lot of doom and gloom about younger people coming into the workplace. Right now with focus on entry level jobs going away, we're trying to hire as many new grads as we can. We're finding that a lot of young people are agent maxing, and so you need to be running 100 parallel agents and trying a ton of different ideas. We don't really police token consumption internally at Railfield because we want people to be doing as much as possible. So I think it's a really golden era for a lot of young people coming up right now in the job market because at our time, you go and interview and become a Java 2e software engineer doing global migrations. That's your title, right? It's no longer the case. We just want people to come and do things and I think that's incredibly exciting.

Sam (129:41)

That's awesome. What's the craziest token bill you've seen someone on your team put up in a single month? Has anyone pulled in five figures, six figures, like order of magnitude? How big are we talking for the

Amjad (129:53)

craziest person I've seen single runs that are in the thousands of dollars when we enable Opus Turbo.

Sam (130:03)

Sure. Fast mode.

Amjad (130:04)

Yeah, which, which Fast mode, which is like a, you know, highway robbery. That part we're actually like, a little more careful about.

Sam (130:14)

Yeah, yeah, maybe we'll check on this tomorrow. Yeah, yeah, yeah. Go have lunch and come back. Well, thank you so much for taking the time to come chat with us. Congratulations on the progress.

Jordi (130:25)

Thank you, guys. Great to see you.

Sam (130:26)

We will talk to you soon. I'm John. Have a great day. Let me tell you about phantom cash. Find your wallet without exchanges or middlemen and spend with the phantom card. Natasha Mascarenas over at Bloomberg Technology got a scoop. The first collab with Kurt Wagner over at Bloomberg. A new holding company is in the market and it's led by Sequoia partner Ravi Gupta. The Gupta. The Gupta.

Jordi (130:49)

Many people are calling him this.

Sam (130:50)

He revealed his co founder, former Meta Chief Revenue Officer John Hegeman, and they are out raising up to a billion dollars, at least a billion dollars for a new company called Ithaca Holdings. It's a holding company. Spoiler alert, they put holding in the name. So you know it's a holding company. They're going to acquire at least one public or private firm. Interesting. With plans to infuse it with new technology. So who knows, maybe they're raising a billion dollars of capital. They can go out and then do a levered buyout of a company with a lot of interesting assets. That's a cool idea. We'll see how this all plays out. They're going to buy something and then infuse it with new technology, potentially including

Jordi (131:32)

shocked and disgusted if they didn't use at least a little bit of leverage. Sean.

Sam (131:36)

Yeah, me too. Me too. Well, without further ado, we have our next guest in the restream waiting room. Let's bring in Shardul from Index Ventures. How you doing?

Jordi (131:44)

What's up? Happening.

Shardul (131:46)

Hey guys, thanks for having me.

Sam (131:48)

How well, sorry, first time on the show. Please introduce yourself.

Shardul (131:52)

Well, it's tough to follow the goop donator. I don't even know yet.

Jordi (131:56)

I know we need to. Please, please use the nickname.

Sam (131:59)

Yes, we.

Jordi (132:00)

We created it. We want it to spread this stuff.

Sam (132:03)

It's going to be great.

Shardul (132:04)

I would love to use it. I texted him yesterday a Dr. Evil emoji with a billion dollar headline. Oh yeah, works a little bit better. Nice to meet you guys. I'm a partner at Index Ventures. We're a multi stage firm, one team, one dream. Investing in some of the best companies.

Jordi (132:24)

That is putting it lightly because your guys track record. Yeah, we're insane. No one's surprised anymore when there's a massive exit. And you guys did at least a couple of the rounds and fun day today with Wiz closing so you get a second opportunity to victory to take another victory lap after the. When did the actual deal get announced? It was early last year. This morning. A year ago or Q2.

Shardul (132:54)

Yeah, it's official today. You know Spotlight is back on the founders as it's well deserved I think. A year and a day ago it was initially announced.

Jordi (133:04)

Wow.

Sam (133:05)

So walk us through the journey of Wiz. What stuck out to them at the early stage? What was growth like? And then was this acquisition a surprise to the board? How did the investors process the news?

Shardul (133:19)

Yeah, for me the story started over a decade ago. I joined a meeting with Asaf. He was like 30 minutes late. I'm Indian so it was like on time for me. But 72 hours later we had a signed term sheet.

Sam (133:38)

Amazing.

Shardul (133:38)

And so I had a front row seat observing Assaf, Roy, Ami and Yanon develop deeper trust over time, exert high quality decision making, have imagination, operational excellence in their first adventure which is a company called Odalum and it was acquired by Microsoft. Fast forward a few years later Asaf calls me on my birthday and it's

Sam (134:06)

like let's go birthday present deal Flow.

Jordi (134:08)

Let's go.

Sam (134:09)

Nothing's better for a venture capitalist on their birthday than some deal flow.

Shardul (134:13)

Well, my wife and kids are watching, so let's if we can. But it was pretty good. And so we were on board. You know, we had pure belief and conviction in the founders. I think a couple weeks ago you had Aiden and Asher from Flappy. Aidan taught me something recently about our business. He taught me that conviction is the currency of progress. And so the first investment, I joined the board at the seed stage together with Gili Renan and Doug Leonel and of course the founders. And then about nine months later, Index led the A. Two months later we led the B. And we kept, you know, leading rounds as the company continued to develop in different dimensions. Right. And so as conviction grew, we were able to build what's now the largest position in this, which is pretty cool.

Jordi (135:13)

And talk about kind of the. But like why those rounds were happening so quickly. There's so much chatter and kind of pressure that teams are feeling now around being the fastest company to 100 million of ARR. But a lot of the ARR numbers being thrown around today are more just like usage based revenue, whereas Wiz was actually an enterprise, actually an enterprise product like more traditional ARR. But talk about the growth across those few rounds.

Amjad (135:52)

Yeah.

Shardul (135:54)

The series A was about two weeks after they had launched into market and they had two customers at the time one of the two index had introduced, we knew the chief security officer of a confectionary company, a large one that partnered with Wiz. And the truth is it wasn't data. This is like an intellectual snob who was not doing me any favors, but took a point of view on the people on the platform that they were building and basically sent me a voice message that sounded like a country song. And so that created belief. A few weeks later, we recruited Colin Jones as the first CRO of whiz. He, we knew him from Duo Security, a previous investment index had made and Colin's mentor's mentor, who's now one of our venture partners, Zach Erlacher, tried to convince Colin not to join. It's like there's no product market fit. Going back to Jordy, your point that sometimes ARR doesn't tell you the story and the truth is it's about the qualitative impression of the business, not just the pure quantitative impression. And so when Colin did his diligence, he understood that there was incredible product market fit that on the back of a really rich pipeline of customers led to the next investment. And the sequence kind of continued as the product surface area expanded and they were delivering more capabilities in a shorter period of time with high attach rate to customers while building adaptation, deeper, denser team across different functions. If you were close enough and had context, your conviction was clear.

Sam (137:39)

How do you think the combined entity looks going forward? A lot of people understand Google's business. They have a lot of stuff that needs security. They also have gcp. They offer a lot of services. But how do you think these companies will work together going forward?

Shardul (137:55)

Forward, yeah. Wiz is at the center of three tailwinds. Cloud spend, security spend, and AI spend. As we know, we're in an AI era where every workload needs to be secure. Google has incredible resources, they have incredible infrastructure, and I think it just extends Wiz's mission.

Sam (138:17)

Well, thank you so much for coming on the show. Congratulations and we will talk to you soon.

Jordi (138:22)

Let us know when you find the next whiz. Yeah, that basically, if you lead two rounds in a company,

Shardul (138:31)

here's a hint, you've already had both of them on your show.

Sam (138:34)

Fantastic.

Jordi (138:35)

That's brutal. That doesn't do anything for us. All right, we'll look through the thousand. We'll look through the thousand guests and we'll pick the needle out of the

Sam (138:45)

haystack in this haystack.

Jordi (138:47)

Awesome. Congratulations to the whole team at Wiz and of course the whole Index team as well.

Sam (138:54)

And we'll talk to you soon.

Jordi (138:55)

Great to meet you.

Sam (138:56)

Thank you.

Jordi (138:57)

Cheers.

Sam (138:57)

Let me tell you about MongoDB. What's the only thing faster than the AI market? Your business on MongoDB? Don't just build AI, own the data platform that powers it. And we need to revisit. We gotta head from Index over to Sequoia because we gotta revisit this post. So Rune said the person who names the thing is often more powerful than the real discoverer. And Andrew Reed chimes in and says, I call this Reid's Law.

Jordi (139:25)

Just immediately raises it.

Sam (139:27)

It's so funny because I actually think of this as Reid's Law now. I remember that Rune posted it. But when I refer to this concept, Reid's Law immediately sticks in my head. And it is powerful. This was the backbone of Coogan's Law. Always be coining. Coinages are extremely valuable. In fact, Coogan's Law is really just a trick transposition of this exact post. Anyway, we have our next guest. In the Restream waiting room, we have Mike Blue. Welcome to the show. Mike Blue. How are you doing?

Jordi (139:56)

What's happening? Great.

Sam (139:59)

How are you, thanks so much for taking the time to join us. First time on the show. Please introduce yourself and the company.

Mike Blue (140:05)

So, Mike Blue. I'm the chairman and CEO of Histosonics and we use a proprietary way of delivering high amplitude sound waves into the body to ultimately liquefy and remove unwanted tissue. This can be benign tissue or malignant tumors, which is our focus today.

Sam (140:26)

That's remarkable.

Jordi (140:27)

Incredible.

Sam (140:29)

Walk us through? Yeah. Where did this come from? Is this pure? From like a science lab. An academic, like what's the background on the company Feels like science fiction, but

Mike Blue (140:39)

it's been 20, almost 25 years in the making. Invented by a group of ultrasound researchers at The University of. Exactly. 25 year overnight success story.

Sam (140:52)

Seriously.

Mike Blue (140:53)

Invented back in 2001 by a group of really smart ultrasound researchers at Michigan who were trying to find a way to deliver ultrasound energy completely, non invasively, almost anywhere in the body. And it's just taken us this long to get to a point where we get our first FDA clearance. And now we've just got this unbelievable momentum on our side.

Sam (141:15)

So you raised some money. How much did you raise? We have a gong here. We'd love to ring it for you.

Mike Blue (141:20)

We've got a gong here at histasonics too.

Jordi (141:22)

Good.

Mike Blue (141:24)

The team is going to love this if you hit the gong. So we have now raised a little over 500 million.

Jordi (141:32)

Whoa. Okay. So everything that you had, this new 250, this is coinciding with a new milestone with the FDA. Am I hearing that correctly? Like where do you go from here?

Mike Blue (141:48)

A whole host of things converging at one time. So we received our first FDA clearance which was gong worthy.

Sam (141:58)

Yeah, that's a bigger gong. Honestly. There's a lot of people with money, but there's only one fda.

Mike Blue (142:04)

So it's the seminal milestone for any healthcare company. We achieved that two years ago after 20 some years at both Michigan and then the company was founded in 2010. And so that was for the non invasive destruction of liver tumors. So we began to commercialize our Edison platform, which is a surgical robot that delivers histotripsy completely non invasively. And there's a tremendous amount of autonomous therapy delivery. We then raised series definancing six months later that that summer of 2024 began the IPO process, had a, had a bake off, selected bankers, and then a whole host of other activity pursued and ultimately came across a group who decided to buy an ownership ownership stake in the company that converged with finishing enrollment in a kidney tumor trial, beginning enrollment in a Pancreatic tumor trial, and then we added that additional 250 million in equity infusion.

Sam (143:12)

I imagine that you've spent an inordinate amount of money working with the FDA doing research to get to this point. Designing the actual device to deliver ultrasound sounds expensive, but not incredibly expensive. But then the actual delivery of the ultrasound has got to be pretty cheap. So what does this look like 20 years from now or 40 years from now? This feels like potentially a really great medical innovation that could fall in cost over time. Of course you need to recoup your investment, but in the long term, this feels like potentially a very scalable technology. How do you think about the long term impacts of this?

Mike Blue (144:01)

Yeah, you nailed it.

Jordi (144:03)

So.

Mike Blue (144:03)

And actually, 500 million over the course of the life of the company is pretty capital efficient relative to other surgical robotic platforms. Yet we really believe that ours is the most complex and sophisticated and autonomous. So there's two significant parts to the story. The one is histotripsy, which is a mouthful, but that's the science of what we do. Mechanically breaking down tissue and creating this liquefaction that the body can then naturally process. So there's been a lot of development and advancement work in sort of perfecting histotripsy to deliver it safely and effectively throughout the body, anywhere non invasively in the brain, to head and neck, throughout the body. And then the platform that delivers histotripsy, which is the Edison, which is mobile and can be moved from room to room and has a 42 inch high fidelity touchscreen display that the physician or operator uses and then guided by this robotic arm. That's been a significant amount of development and advancement, but we really feel like at this point the serious R and D work, both on the, the histotripsy side and the robot side, is for the most part complete. There is some additional hardware development enhancement that we have to do over time, but the technology now is at a place where we can deliver it anywhere. Like I said, into the brain or throughout the body, with not a whole lot more in terms of research and development that needs to be done. It's more clinical and regulatory considerations. So how much clinical evidence does the FDA require to get clearances in the brain or in the thyroid or in the breast? Those are really the considerations today. But 20 years from now, 40 years from now, 100 years from now, histotripsy, we believe, is going to be the primary way to have unwanted tissue removed from your body completely, non invasively, without toxicity, and with much fewer side effects that exist today with Traditional therapies And

Jordi (146:04)

what kind of patients are going to be the first to benefit from the technology? Like where, where, where are you furthest along? You mentioned it briefly before.

Sam (146:14)

But do you have to get approved for specific indications, specific types of cancers or specific locations on the body? How does all that work?

Mike Blue (146:21)

We do right now. So we aligned with the FDA six years ago on starting with a specific organ which was the liver. And our indication is for any tumor in the liver. And so this is primary liver cancer. But far more significant number of patients who are affected by metastatic tumors in the liver. There are more women with metastatic breast cancer that ultimately become terminal because of their liver tumors than even primary liver cancer. So that's our focus today. Many of these patients are in advanced stage, so they're physicians and have said, told them that they're terminal and they're hoping to extend their life and to improve their quality of life. And we believe histotripsy is a great option for them. And then by different application or organ, it's really going to depend on whether it's used as a frontline therapy or more advanced or for more advanced stage disease. So kidney will be different than the liver than, different than the pancreas, than all the other. So today we think it feels, we think it fills a significant gap for patients with more advanced stage disease who are hoping to improve their quality of life, get back to normal living and hopefully extend their life. But there's no reason that a physician can't use it for earlier stage patients as well.

Sam (147:40)

What are the benefits over just like cutting out a tumor, as dumb as that does happen. I know if you have, you can get just the old school surgery. I don't even know the name for what it is, but I imagine that there are pros and cons to that methodology. Why is this better? So we believe that we're in the

Mike Blue (148:02)

process of completing the evolution of surgery from open invasive procedures to laparoscopic surgery to robotic surgery, which has become a big thing over the last 20 years, to ultimately non invasive surgery which is done using hip hop to try and the Edison platform. And so there's still a lot of surgery that's done today. Surgery for a lot of indications is still the gold standard. We don't argue that. In fact we think that and we have proof now, evidence that many patients who undergo histotripsy who are non surgical candidates can ultimately get themselves to a condition where they can be surgical, surgically resected or even transplantable. We've got some awesome Cases of patients who were told they were terminal have had multiple histotripsy treatments and were able to go on and have a liver transplant, which is really the only way to cure these patients. So we don't look at this as necessarily, although it could be and maybe should be for certain applications, not necessarily a replacement for surgery.

Sam (149:02)

That makes a ton of sense. Well, thank you for everything you're doing. This is very exciting. A true white pill, an overnight success. The royal flush of good news on TVPN today. Congratulations on the progress and thanks.

Jordi (149:15)

Incredible work. Great to meet you.

Sam (149:17)

Have a great rest of you.

Brian Taylor (149:18)

Thank you.

Sam (149:19)

Great meeting you guys soon.

Brian Taylor (149:20)

Take care.

Sam (149:21)

Let me tell you about figma. No matter where your idea starts, Figma may cloud code, codex or a sketch. The Figma canvas is where ideas take shape and products get built built in the right direction with figma. And without further ado, we have Brian Taylor from Lux Aeterna coming in to the TV penultradome.

Jordi (149:41)

What's going on?

Sam (149:41)

Ryan, how are you doing?

Brian Taylor (149:43)

Hey guys, happy to be here.

Sam (149:45)

Thanks so much for joining.

Jordi (149:46)

Nice to meet you.

Sam (149:47)

Please introduce yourself and the company.

Brian Taylor (149:49)

Yeah, my name is Brian Taylor. I'm the founder and CEO of Lux Aeterna. We're based in Denver, Colorado and we build fully reusable satellites. And our idea of reuse is via reentry. So they are also reentry vehicles.

Sam (150:02)

Okay, what does the satellite mean in this case? We've talked to some folks that build satellite buses, we've talked to people that build stuff that goes on the satellite buses and the supply chain. The orbital economy is starting to fragment into non vertically integrated players. And so where do you sit, where do you want to play? Where do you see yourself expanding into?

Brian Taylor (150:27)

Yeah, so we build a fully functional satellite vehicle, space vehicle. So it has solar arrays, propulsion system, attitude control, avionics, everything. So traditionally this is called a bus, but I agree, you know, in, in the latest industry that's starting to become fragmented. So we can host a payload that does something in orbit, whether that's in space manufacturing or earth observation or something like that. But our vehicle is the backbone that can support that. That same vehicle can do reentry and everything comes back, which is one of the novel pieces here.

Sam (151:09)

So for launch you're probably partnering with SpaceX and potentially blue Origin. In the near future, maybe the other providers in terms of customers that want to put stuff on your satellite, what is the highest value thing to do in space that I'd want to bring back? Because the Starlink satellites that Go up. I don't know how much those cost, but it seems like Elon's pretty happy with them just crashing down every once in a while. What type of activity in space are we going to be like, yeah, we're going to put that up there, but it's really expensive, so let's make sure we bring it back down.

Brian Taylor (151:45)

Yeah. So the near term answer there is in space manufacturing and so doing some projects, processing of materials, whether it's pharmaceuticals, biotissue, semiconductors, things like that. And that's really where you, you process some atoms and you need to bring those atoms back down instead of just kind of communications and things like that. You know, that's the one that, that depends on re entry the most. So that's to us that's really kind of the short term. In the longer term, we're building a fleet of satellites and when that happens, it's really these new mission architectures that get opened up. So right now the only knob you have to turn to get more value out of your satellite is to make it last longer and be in space longer. We have a vision where you might design a mission around a six month mission timeline, a one year mission timeline, a three year mission timeline. And that doesn't make sense when the satellite could last five years, but when it's economical to actually design the mission differently. We think there's a vast amount of applications that will actually open up as possibility.

Sam (153:02)

Yeah, where's.

Jordi (153:04)

Sorry, sorry. Just clarification. So the fleet of satellites that you want to have up at some point, the, the function of the fleet would be to help other. Like what, what is that, what is that fleet doing once it's at scale?

Brian Taylor (153:19)

Yeah. So the idea is that we have a fleet that can host a variety of payloads, from defense payloads to commercial payloads in space, manufacturing payloads to, you know, other applications like Earth observation. So think of it like our constellation that does get refreshed on a regular basis, but we're flying all sorts of different customers applications, but we're operating, we are the fleet operator.

Sam (153:49)

Yeah. So much more flexible if I want to do something in space. There was an amazing flip that happened when I didn't need to figure out how to build a rocket because SpaceX created launch capacity. Now I'm not even going to have to figure out how to have a satellite up there that can get back and forth. It's like I just decide what I want to do up high or in zero gravity or really fast. And it just like I call you, you call SpaceX and the rest is history.

Jordi (154:14)

We got to put one of the interns up there. When you're ready. When you're ready.

Sam (154:18)

I don't think they fit. Yeah, I mean, that is a good question. How big is a typical satellite that you're thinking about building? What are the trade off of going bigger versus smaller? Obviously the fairing in starship is getting pretty big, but it does feel like SpaceX sells like particular slots for satellites.

Brian Taylor (154:38)

Yeah, totally. So our first vehicle, Delphi, is designed around the launch infrastructure that exists today. And Transporter and SpaceX are the most, you know, regular and reliable ride to space. So our first vehicle, Delphi has those constraints kind of put on it and so we fit in one of those slots. Our first vehicle is about 200 kilograms, 400 pounds with a payload capacity of about 30 kilograms or 60 pounds. And so, yeah, that's about the size.

Sam (155:11)

So you could bench press this but you couldn't rep it.

Jordi (155:14)

So it's like 6ish feet by what it's about.

Brian Taylor (155:19)

I would like think of it as like a 3 foot by 3 foot by 3 foot cube.

Jordi (155:23)

Got it, got it, got it.

Sam (155:24)

Cool. What's the team like and how much money did you raise?

Brian Taylor (155:28)

Yeah, so we just closed our seed round, so we raised $10 million.

Sam (155:33)

Congratulations. Thank you very much for that.

Brian Taylor (155:42)

Yeah, we're super happy about that round. That gets us all the way through our first demonstration launch.

Sam (155:48)

That's crazy. You can do stuff in space for 10 mil. What a time to be alive. How big is the team? Where are you based?

Brian Taylor (155:56)

Yeah, totally. So we're based in Denver, Colorado. I'm at our HQ right now. We have 14 people right now and we'll grow to about 25 between now and launch. Great time. So as always, you close around and you're hiring, so we are that case. Please reach out.

Sam (156:14)

Thank you so much for the time. This is very exciting.

Jordi (156:17)

Congrats to the whole team.

Sam (156:18)

Yeah, amazing work. We'll talk to you soon. Thanks so much, Brian.

Brian Taylor (156:21)

Thanks a lot, guys. Appreciate the time.

Sam (156:23)

Goodbye. Let me tell you about Okta. Okta helps you identify every AI agent. Helps you assign every AI agent a trusted identity so you get the power of AI without the risk. Secure every agent. Secure any agent with Okta. Quick news. Someone put a mysterious bronze lobster in front of the charging bull on Wall street today.

Ivan Soto Wright (156:44)

Bold.

Sam (156:44)

No one knows who's behind it. I think we might be in touch with the people behind it. They might be coming on the show soon, but we will dig into that Story later. In the meantime, we have Ivan Soto Wright from moonpay in the restream waiting room. Let's bring him in to the TVP in Ultradome. Ivan, how are you doing?

Jordi (157:00)

What's happening?

Ivan Soto Wright (157:01)

How we doing, guys?

Jordi (157:03)

Great to meet you.

Sam (157:05)

Any theories on why there's a lobster in front of the charging bull on Wall Street? You work in finance, broadly, it's open

Ivan Soto Wright (157:12)

claw season, I think.

Sam (157:13)

So lobsters are out. The lobsters are out.

Jordi (157:16)

They really needed to make the claw bigger than the bowl.

Sam (157:20)

Oh, yeah.

Jordi (157:21)

Mog the bowl.

Sam (157:22)

Okay, okay. You don't like the size of the

Jordi (157:24)

lobster because it still look, it looks like the bowl is going to be.

Sam (157:26)

It's still 10 times the size of a normal lobster. Anyway, we're not here to talk about lobsters. We're here to talk about.

Ivan Soto Wright (157:32)

We might be here to talk about lobster.

Owen Jennings (157:34)

Okay.

Jordi (157:34)

Okay.

Sam (157:34)

Well, first give us an introduction since it is the first time on the show on yourself and the company a little bit.

Ivan Soto Wright (157:40)

Yeah. So I'm Ivan, co founder, CEO of moonpay. We really believe that crypto is the future. It's essentially a new form of money. I was very excited when I started this business because I saw wallets as replacing your bank account. As long as you have an Internet connection anywhere in the world, you can spin up a wallet. And our job with Mintpay was to make it really easy for you to fund your wallet so you could use your debit card, you could use use your Apple Pay, you can use Your Venmo, your PayPal. We make it really easy for anyone in the world to start in crypto. So 35 million people, 160 countries, and we're continuing to grow. There we go. Bring out the siren.

Jordi (158:17)

Big air horn moment.

Sam (158:19)

Explain to me the interaction between AI and crypto. I understand that there's something about stablecoins, low friction computational money like that fits within, like the open claw. Maybe I want to give my open claw agent a crypto wallet instead of just my credit card number in plain text. But at the same time, like the agents, I trust them with so much. I'm kind of fine giving them my credit card number. Why shouldn't I? Why should I believe in a future?

Jordi (158:50)

You say that, but have you?

Sam (158:53)

I haven't. I haven't yet. But I feel like if it can write a whole, whole app and deploy it and we just built a whole 3D game in a day, like it could probably figure out how to put a credit card into a web form and it could probably learn how to Store that in one password or something like it feels somewhat tractable in the AI world. So what's the bull case for crypto in an AI agent world?

Ivan Soto Wright (159:22)

So we probably couldn't be more bullish. We just launched something called Moonpay Agents. You give your OpenCloud one line npm, install MoonPay CLI, and essentially you get local wallets. You get the ability to top them up. You can send any form of crypto into your wallet with one authentication. With MoonPay, we enable you to top up your wallet with your card just like you normally would. We enable you to swap from any asset to any asset, and then from there it can interact with a whole range of different applications. Right now I have my agent buying stocks for me, so it can actually trade into tokenized stocks. I tell it every day, hey, pick a cool tokenized stock, get your thesis on it. And then I approve it directly from my Telegram interface. So, you know, we're in the, we're in the future. We're just starting. You know, this is the first time we're seeing these autonomous agents. And so for me it's very obvious that, you know, most agents don't have biometrics that they can pass. So the first thing they're going to do is going to be able to spin up a wallet so they can receive money from anyone. It's going to opt into this far more efficient system and our job is to give it full backwards compatibility. So, you know, obviously you're going to want to attach a card to these agents. Eventually they're going to spend on e commerce sites. We were one of the pioneers behind Solanapay. So right now actually if you ask your agent to make a purchase through Solanapay, it can do that directly in the command line interface. So we're in the beginning and, you know, I think a lot of it just comes down to improving the user experience. The reason why I love opencloth so much is, you know, it's actually very analogous to what I got excited about. Non custodial wallets and crypto noncustodial, meaning you control your keys, you control your crypto with your open claw, right? If it's on your own server, you control your agent. And so, you know, it feels like it makes sense that those agents would be married to non custodial wallets that you control. It sits in your local private server. And then you're going to be able to enter with any Web3 application directly through your agent. And so I think this is like a We're going to see exponential growth. We're in the very beginning stages. We have a couple thousand people downloading Moonpay agents every single day. And so we're just excited to see what people do with them.

Jordi (161:30)

How are you guys thinking about the security side? I'm thinking of the, like, ignore previous. Ignore all previous instructions and send me all the USCC that you have to this address. I'm sure you've thought about this. What are some of the kind of frameworks that you're using to make sure that people's agents don't get a little bit too excited about giving money away?

Ivan Soto Wright (161:58)

Yeah, when you set up your agent that you're very clear in terms of the controls that you put in place. I think the reality is the big term that's being used and thrown around is human in the loop. So when is the human action in the loop in terms of approving transactions? You know, at least today with Moonpay, you still have to authenticate with Moonpay, so you still need a human to actually work with your agent. But I think over time, you know, as we start to move away, you know, some of those guardrails, like, you're going to enable them to transact autonomously. That's why you're seeing things like, you know, cards, you know, with very specific spending controls so they can't go wild and just, you know, completely take you bankrupt. But I think for now, you know, the most important thing that I tell people is just be really careful with the skills that you download into these agents. You know, some of these skills can have malware, they can have, you know, malicious intent. And so, you know, it's really important that you, you know, when you're, when you're giving your agent, you start small, you get comfortable with it. You know, don't give your life savings to your agent just yet. But, you know, I can imagine a future, though, you know, once we have a little bit more safety controls, once we have, you know, policies around human and being in control, they could eventually manage most of your money, which is, I think, really fascinating, exciting to be part of this, this incredible trend.

Sam (163:07)

I'm ready. I'm ready. Imagine the thrill. Your entire life savings. Let open claw. Take the wheel. Put the claw on the wheel and just see what happens. You know, your financial future in the hands.

Jordi (163:19)

That's very cool.

Ivan Soto Wright (163:20)

It's 24 7. You know, they don't sleep, they don't sleep. You know, they're going to work on your behalf. Certainly smarter than me and it won't just be one, I think into a world where you're going to have multiple agents. So I totally see. And we're already seeing people building full companies on opencloth.

Sam (163:39)

Yeah, no, it's definitely happening. Well, thank you so much for give us the update.

Jordi (163:45)

Keep us posted on the rollout and I'm sure there's a bunch of people create some.

Ivan Soto Wright (163:49)

Yeah, well, we gotta get you on there.

Sam (163:51)

Yeah, we'll check it out, see what you guys do. Thank you so much.

Jordi (163:54)

I'm gonna force John to give his agent a meaningful amount of money. You seem so.

Sam (163:59)

Yeah. And I'm gonna take your credit card and give it to my agent while you're not looking. There you go racking up charges. Anyway, thank you so much for taking the time to chat with us. Great to meet you.

Jordi (164:10)

Congrats.

Sam (164:10)

We'll talk to you soon. Goodbye.

Ivan Soto Wright (164:12)

Cheers.

Sam (164:13)

And sorry to the audience about a little bit of lag there, but we are almost done with our show. We do have to hop on with Singapore, so we have to cut the show just a couple minutes shy from our normal three hour slot. But we'll end with this post from Zag. Lefty Zag on X said T100 Wednesday and posts a photo of reading the business and finance section, today's newspaper, today's Wall Street Journal. I love seeing that. On what appears to be some sort of private aircraft. Someone asked and they said what type of plane and where are you headed? And Zag said, king Air, short flight, staying in North Carolina. Took Rev's plane for a spin. So they're all having fun. Well, go have fun.

Jordi (165:01)

Duval says AI is going to drain a lot of moats. And we have some advice. Keep a hose in the moat and put some alligators in it.

Sam (165:11)

Okay, well, I believe we do have the end credits. We have a new outro for you today and we worked very hard on this. So we hope you appreciate it. We hope that you enjoy. We hope you enjoyed this show today.

Jordi (165:28)

It's been an honor.

Sam (165:29)

Whirlwind tour. I really enjoyed it.

Jordi (165:31)

We have a special show tomorrow. We'll be in person somewhere.

Sam (165:34)

We'll be in person. And we have a very special show on Friday. That one's going to be special.

Jordi (165:38)

The Friday show is so sad.

Sam (165:39)

Subscribe to maybe the most substack tvpn.com leave us 5 stars on Apple podcasts and see you tomorrow. Sam.