Who Is Soham Parekh? Erebor News Leaks, Microsoft Scales Back AI Chip Plans and Cuts 9,000 Jobs, Sam Altman Slams Meta’s AI Talent Poaching | Deedy Das, Emily Sundberg, Mert Mumtaz, Zak Kukoff, John Kim, Amit Jain

You're watching TVPN.

Today is Wednesday, July 2, 2025. We are live from the TVPN Ultra Dome, the temple of technology, the fortress of finance, the capital of capital. Welcome to the show. The current thing has come in hot off the presses. It's Erebor, a new crypto bank. PayPal co founder Peter Thiel and Anduril founder Palmer Luckey are launching Erebor, a digital bank built for for crypto and tech startups.

You know what I'm saying, John? The trade of the century was going back and buying up all the Tolkien domain.com.

That'S right. That's right.

You would have become a billionaire yourself if you just bought all the dot coms.

There was a lot of debate over whether or not the Tolkien, the Lord of the Rings naming scheme had kind of run its course. Right. Cause people were knocking it off. Non founders Fund companies were starting to use the names. Not gonna say any names. O that is insane. That is insane. I can't believe no one's ever said that. I'm not gonna dig into that anymore because.

We'll let you dig into it.

Let you interpret that joke. But we have news from Zach Kukoff on the show yesterday. He's coming back on the show today. He's our Washington correspondent. He says this is great news about the new Lord of the Rings themed bank Erebor. Zach Kukoff says this is great news. JP Morgan, hsbc, et cetera haven't fully filled the gap that SVB and Republic left an insider run risk forward tech bank is much needed. Backers include Palmer Luckey, Founders Fund, Peter Thiel, Joe Lonsdale, etc. I think Lux is in this too. I don't know if that's public yet though. We'll have to figure that out.

Scoop.

But they're obviously the Silicon Valley elite are rallying around. We love to see it this this new project. So they're calling it the next SBB for hard tech. The filing says says Erebor wants to be the most regulated entity conducting and fac facilitating stablecoin transactions. An unusual boast meant to soothe OCC examiners while still holding stablecoins on its own balance sheet. Who it serves and why. Columbus. This bank's gonna be in Ohio. Target customers are AI, crypto, defense tech and advanced manufacturing startups. A digital only model will be legally headquartered in Columbus, Ohio with a satellite office in New York to tap Wall Street.

The capital of capital.

The capital of capital. Very exciting. So yeah, we'll have to dig into this we're having Palmer Lucky on the show soon. Would love to hear more about the plan here. And I agree SVB was a special, special institution. It was very sad when it went away.

There's been a gap in the founder dinner market. They were a huge funder of founder.

That and also mortgages.

That's right.

Big on the mortgage front. I mean it really was a crazy failure where you could have like $40 million but illiquid and be making get a $200 million mortgage and you couldn't buy a house and it's just like okay, and then, and then you're strapped with this, with this question of like, okay, do you raise your salary and start raising burn just to get a mortgage and stuff or do you put it on the balance sheet of the company? That's been done before there.

Yeah. It's crazy that, that within whatever a month we lost SVB and First Republic, the two funds that really served. So it makes sense to build back. It is funny that Zach said we need a risk forward tech bank because it was risky. Well, yeah.

What's interesting is they didn't get burned on the startup stuff. Yeah, they got burned on Treasuries, the least risky thing ever. But no one was expecting a duration mismatch. Yeah, the duration mismatch. Anyway, in other news, Microsoft scales back ambitions for AI chips to overcome delays is the information. And NeoMorgana here is asking Grok what were, what were they trying to do with their chip ambitions? And Grox says Microsoft chip ambitions aimed to reduce reliance on suppliers like Nvidia Optimize for performance for Azure AI and Copilot at lower costs. They developed chips like the Azure Maya AI accelerator for AI tasks and Azure Cobalt CPU for cloud workloads. However, delays such as the Maya 200 chip postponed to 2026 and design challenges led to scaled back plans. So Google has the TPU, Amazon has Inferentia, Apple has the M1, M2, M3 chip. All the hyperscalers getting in on the custom silicon action.

What was the name of the quantum chip that Satya was teasing?

Majorana.

Majorana.

Majorana.

But that's still in the works.

That's still in the works. But that's a completely separate thing. And that, you know, that is such a different fork in the tech tree that you're not going to be writing, you're not going to be like, oh yeah, we have some Cuda code that's inferencing some, you know, you know, large language model and let's just Run it on Quantum. It's like you have to rebuild the entire, the entire system to run on Quantum. And that's very far out. But people are optimistic, people, sharp people that I talk to are saying like Quantum is. All the companies right now aren't delivering products that are in use. But there's no like, like the physics, the science is, is solid. It just might be 10 or 20 years until this is really in production. So it's all about, you know, TSMC fab capacity right now. So Microsoft scales back ambitions for in house AI Silicon Maya AKA Braga chips slip for a full year. You hate to see it. We love, we love custom silicon here. We want it to arrive as fast.

As possible when timelines extend.

I hate it, I hate it, I hate it.

It does happen though.

So I mean we are big tech defenders here and we are very sad to report.

The delayed part is expected to trail Nvidia's Blackwell accelerators once it finally ship and undermining Satya's goal of cutting GPU spend.

And that's gonna be hard because the new Blackwells are gonna come up. They're gonna be amazing. He's gonna need some. You don't wanna be GPU poor right now.

Not right now.

You might, you, you can be a leaser but you can't be GPU poor. And so while Microsoft pauses, Amazon is planning Trainium 3 for late 2025. And Google already has its 7th gen TPU in production leaving Azure the slowest of the big three to deploy second generation custom AI software. And in other Microsoft news, they are reorgang this entire company right now. They are announcing that they are cutting 9,000 workers in a second wave of major layoffs. We already reported on this, we talked about this. But they have 4712 applications for Hilary.

How many full time employees does Microsoft have? Isn't it in the hundreds? Thousands?

I like that Tyler's wearing the bezel hat today. It's a good one. Just throw on the team.

Tyler clearly doesn't have cluli up otherwise he would have.

No, he really should be running that during the show.

Let me think for a second, but let's approximately 228,000.

That's right, 228,000.

So this is what they're cutting 9,000 employees.

It's 4% of the 4% of the headcount. This is the second cut in 60 days. The July round follows 6,000 job losses in May and hit sales, gaming and engineering. Flattening the org chart. Management layers are being Removed to raise the coder to manager ratio. Love that. You know, I mean, it really is this world where the line between coding and managing is blending and programmers can do a lot more and act like PMs because there's AI tools and they can roll out stuff. So you're not in this cycle anymore. If like the manager decided that we need to implement a function, go take a month to build that, go take a week to build that. It's like, okay, I built it. Now what's next?

You got the afternoon?

Yeah, yeah, exactly, exactly. So. So that higher throughput engineering should come through in a delayering of the organization. And in general, these, these big companies, they often get really huge and then they need to kind of delayer and streamline the company. Anyway, seems like we have some technical difficulties. We will sort that out while I do the next story. Okay, great. Google, Amazon and Meta have all trimmed VPs and middle managers this year. Redmond is citing dynamic market conditions rather than performance. Interesting. Anyway, in other news, China hosted the first fully autonomous AI robot soccer game. I'm hoping we can pull this video up. Ian Bremmer says they've still to go. I want your review. You haven't seen this video yet?

Not yet.

Okay. First time we're reacting to it live on the stream. China's been working on this for a while and the reason I'm bringing this up is because there's a story in the Journal about China is rapidly eroding America's lead in the global AI race. Deepseek and Alibaba's open source Quinn are now running inside hsbc, Saudi Aramco and even on AWS and Azure, forcing Western vendors to justify higher closed source pricing. Very interesting development. We'll see what happens with the new Llama models, whether those can catch up to Deep Seek and Quinn unseat them. We're also going to have the founder of Mistral on the show soon. Anyway, this is the latest robot soccer game out of China. How do you think they did? It's not bad.

They're not flopping enough.

Yeah, yeah. They don't understand the carried out on a.

They know showbiz.

Yeah, that's good team.

That's good showbiz.

So if you watch the beginning of this video, the beginning shows one of the robots falling down. Look at this guy. Oh, oh, oh, oh, oh, oh. He's down.

Okay.

He's down. He's down.

Okay. They would have got a whistle for that.

China knows how to pump up attention on AI. They got the Robot marathons. They got a variety of stuff going on to draw attention to AI. And it's certainly, it's, it's, it's the hottest industry. They got the hedge fund folks moving over and they love robots. And we're seeing all those displays of drone shows, like the drone light show that is going to inspire our generation to go, want to work in drones. Right. And so in saying to work in killer drones. Maybe, maybe.

No, there is. Remember when we covered DJI in detail earlier this year? DJI puts on challenges where they people can compete to make killer robots. So they don't even hide the fact that these robots can be used as death machines.

Yeah, yeah. I mean we're going to see a lot more of this. But the government is fueling a scale up of AI over there. Beijing is pouring subsidies into cost efficient domain specific AI while Washington chases AGI widening a practical capability gap despite US chip controls. US officials worry Chinese guardrails bake censorship into exported systems. The fractured ecosystem could undercut global security cooperation and US influence over AI norms. And so yeah, this is what we've been talking to about Aaron Ginn. He'll be on the show tomorrow. We're also talking to Chris Miller, the author of Chip War Tomorrow.

And we will dig into the US allows millions of humanoid robots controlled by Chinese companies to be sold in to the United States. Yeah, it will be a colossal failure.

Yeah. I was listening to Dylan Patel's presentation at AI Engineer on YouTube on the drive in today. And the number of H100 equivalents that Huawei was able to fab at TSMC like in the midst of the chip controls was crazy. He breaks it all down and it's fascinating. So first off, there was a, there was a company that was basically a front for Huawei that was like, yeah, we want a ton of chips, tsmc. And TSMC was like, sure, you're not based in China, we'd be happy to sell these to you. But they sold them like a million chips or something like that. So the latest Huawei Ascend is crazy. Then they have the Cloud Matrix 384 or something like that, which is like 384 chips all networked together. Specifically instead of there's two different types of wires that can, they can wire the chips together. Basically there's copper and then there's fiber optic. And fiber optic is NV link. That's what Nvidia touts as. Like when you get a rack of servers and they talk to each other really, really fast. That allows for Better training and inference too, but it's very important. It's basically like you're taking a bunch of different chips and you're making them all work as one. So it's super valuable.

Yeah.

Apparently Nvidia tried to do a rack scale. I think it was 64 chips all networked together with optical cables and they couldn't get it to work. Like they couldn't get it out. And Huawei actually figured it out. And so even though their chips aren't as solid as Nvidia, like on the networking side, they're outperforming there according to Dylan Patel. And then the other interesting thing is that like the chip controls don't just affect like TSMC and Nvidia. Like obviously you can't just go buy H1 hundreds or H2 hundreds if you're in China anymore, but you can go to TSMC through a shell company. But then you also need high bandwidth memory and that comes from Hynix SK Hynix in South Korea. That's also controlled. So what they would do if they had another company, they would say like, yeah, like we definitely want like all the best HBM high bandwidth memory in like this device that we're making, like just build it for us and install it and then they would ship it and then they would disassemble the device, rip the memory out and then put it on a GPU chip. Yeah, it's crazy, crazy, crazy. So he has a bunch of good details, we can dig into that a little bit more later.

But Deep Seek was built in a cave with scraps anyway.

James Cameron has been talking about AI copyright fears. He says fears are overblown. Humans aren't original either. A lot of hesitation in Hollywood and entertainment in general are issues of the source material for the training data and copyright protection. I think people are looking at it all wrong. Anybody that's a human being is a model. You've got three and a half pound meat computer. That's the brain here for three and a half pound meat computers. As a screenwriter, you have a kind of built in ethical filter that says I know my sources, I know what I liked, I know what I'm emulating.

I know how I'm hallucinating.

I also know that I have to move it far enough away that my own, that it's my own independent creation. This is the great artist steal instead of copying. What's the difference between stealing an idea, transforming it, turning it into your own and just copying something derivative? And we'll go into that a little bit. So I think the whole thing needs to be managed from a legal perspective is what's the output like? Is the output transformative? And so the reason I want to talk about this is because the Wall Street Journal is covering the fact that studios, artists and tech giants are clashing.

Over copyright in Washington versus the Algos.

Yeah, it's a big one.

Meta recently scored a courtroom win. Judge Vince Cabria tossed author's claims but warned you are dramatically changing, even obliterating the market for that person's work, signaling more litigation ahead. I don't know if I agree with that. Right. You've used, Maybe you've used ChatGPT to summarize the key takeaways from a book. Were you really going to buy? And you know, does that stop you from, from wanting. Is it. How different is that from, you know, reviewing, you know, reading a book review or things?

So with regard to existing intellectual property, like I have gone to ChatGPT and said, hey, I'm trying to tell a story, a bedtime story to a four year old. Let's throw Mickey Mouse in there. And then why don't we mix it up with Batman? Oh wait, they are owned by two different companies. One's owned by Warner, one's owned by Disney. How will ChatGPT handle it? Well, it tells me a story with both of them. It's fine. This could never happen in the real world because of the IP is owned by different companies. But it does it on ChatGPT. But is it, is it, is it a truly a subs, a substitute? Or does it actually just make my four year old beg for more Batman and Mickey Mouse toys? Like I think it's probably net positive brand. Exactly. It's probably, it's probably net positive. Probably not fully a substitute. We're still buying and probably more likely.

To want to go to Disneyland.

I agree. And also Universal Studios go see the Batman ride. But the, the other interesting thing, I guess Warner. So it'd be Six Flags anyway. Correction, correction. But the interesting thing is that like what about new intellectual property? Like there are people that are saying I'm going to build a new intellectual property on, you know, and I'm going to use crypto rails like pudgy penguins, like they're going to try and use that technology to, to accelerate the development of a new franchise, new IP that they own. There are also folks who are saying I'm going to use artificial intelligence to create new ip. We saw this early with little Michaela. Yeah.

Now we're seeing there's A community owned IP creation. What is the, what is the Italian. I don't know, somebody.

Oh, you're talking about the Italian slop lords or whatever. The brain rot.

Italian brain rot.

Yeah, yeah, Italian brain rot. Where like someone probably owns some portion of the IP there and try and turn that into.

Feels like a free for all.

Yeah, I mean, all this stuff is changing, is changing really quickly. But I think at the end of the day, all that matters is that like, everyone gets a steak at some point and everyone gets a slice of the actual, like economic profit and it's somewhat fair. And so that's why the courts exist. They will sort this out.

Meanwhile, over on Capitol hill, a proposed 10 year federal ban on state AI laws was stripped from Trump's big, beautiful bill after creatives and child safety hawks Black blasted it as a big tech giveaway. We heard from, from Zach Kukoff yesterday that this was a win, effectively a win for Anthropic, who was, you know, interested in seeing this get blocked. The issue is that states like California have so much influence that if California passes AI related laws, they'll likely, you know, lead to widespread compliance.

So is pressure from both flanks. Tech lobbyists argue mass scraping is essential to US AI leadership. That makes sense. I mean, Deep Sea is not certainly scraping the entire Chinese Internet and so you're going to get left behind, but.

And the entire Western.

Yeah, everything. And YouTube, we'll take the whole thing. While Writers Guild and SAG AFTRA are pushing for compulsory licensing and outright data mining bans, leaving studios caught in the middle. Anyway, we will dig into that more in a little bit. There's more news about the trade deals. So Japan is saying we're not going to do a deal or we're going to push back. They're calling Trump's bluff. We'll see. It's heating up. Trump on Japan. I will write them a letter saying we thank you very much. We know you can't do the kind of things that we need and Therefore you'll pay 30, 30%, 35% or whatever big whatever number that we determine because we also have a very big trade deficit with Japan. Let's go to the video from Donald Trump, the president. So what I'm going to do is I'll write him a letter saying we thank you very much and we know you can't do the kind of things that we need and therefore you'll pay a 30%, 35% or whatever the number is that we determine because we also have a very big trade deficit with Japan, as you know. And it's very unfair to the American people. So they maybe will be happy. They maybe won't be happy. But some countries we won't even allow to trade. But for the most part, going to determine a number, just very simply write them a nice letter, probably one page or a page and a half at the most. And it's going to be essentially congratulations and it's going to be an honor to allow you to go and do business in the United States of America, because it really is an honor to be able to do that. But we never viewed it that way in this country. We're in the Gulf of America hat, you know, Mount Fuji's right there. Might want to turn it into a different name as part of the dream.

Maybe that could be a chip. Chip, you know. Yeah. Bargaining chip.

Yeah.

Naming rights to Mountain Fuji.

Yeah. You can keep the name of your own mountain named Mount Fuji. So a lot of this is going about cars, obviously. NSX is the top of mind. The Acura.

Yeah.

Remember when. Remember when also the ls when they drop deep research.

Sam said he used deep research to find like a one of one.

Yeah. Nsx. Yeah. Yeah. These are important.

This is any. I like the hat. Unclear where this goes. Where this goes. And I think. I feel like we've just had such a good thing going with Japan.

Yeah.

Feel like it's.

They should be a logical humanoid robot partner. You know, they got Asimo. Honda's been building robots for two decades.

Yeah.

I was really hoping that something would happen there.

We never comment on politics, but I would. I would like to see the United States leaning more into that partnership and figuring out how we can make a lot of things together.

So what else here? Tokyo refuses any pact that preserves Trump's 25% car tariff. Negotiator Akazawa fears electoral blowback if rice and autos are conceded. Trump has threatened fresh duties and rejected deadline extensions. Unnerving allies wary of US tariff volatility. And it's Apache scorecard. To date, the administration boasts only a limited UK accord, a Vietnam mini deal and a China truce. Canada, South Korea and the EU remains stalled amid tariff threats. And so we're still in that pause. But staying on Capitol Hill, Representative Brian Fitzpatrick says, Today I called on the unit on the president to address my serious concern regarding reports that the United States is withholding critical defense material pledged to Ukraine. This comes as Russia launches the largest aerial assault since the war began, firing over 500 weapons at civilian targets in a single week. Ukrainian forces are not only safeguarding their homeland. They are holding the front line of freedom itself. There can be no half measures in the defense of liberty. We must, as we always have, stand for peace through strength. I have formally requested an emergency briefing from the White House and the DOD to clarify these reports, review our nation's weapons and munitions stockpiles, and ensure the United States remains fully committed to providing Ukraine with the use, with the resources it urgently needs to defend its people and preserve the cause of freedom.

High level, what happened here, we basically opened a new front to this, like broader global conflict. Right. We had Ukraine, then we added Iran. We didn't and don't have the munitions StockPiles to sustain two proxy wars or in the case of Iran and actual.

I remember reading in the Journal, day this came like the day Ukraine, the Ukraine, Ukraine war broke out. People were saying like, okay, we are out of Javelins, like the shoulder mounts.

No, there's people like Dhruva Rohendra from Deterrence who we've had on the show at least I think a couple times, close friend and partner, and then Aaron slowed off, who have been talking on multiple occasions about how some of the weapon systems that are critical to, you know, the US's overall strategy will have like, yeah, we have like 500 of those. And then you use them and then they're out and then remaking them.

You're just not ready to be like, yeah, because like in peacetime, why would you be like, yes, I need to produce a million drones that are just gonna get blown up. Like the industrial capacity just isn't there. So Patriot missiles have paused deliveries of Patriot interceptors, Air to air missiles and artillery shells have been redirected to replenish U.S. stocks. Pentagon officials confirm the shift reflects a re ranking of threats from China and the Middle east above uk, Ukraine, despite earlier White House signals of Patriot transfers. And Ukrainian officials and US lawmakers say the delay could translate into more civilian casualties as Russia steps up aerial salts. They're talking about aerial.

Russia is taking advantage of this period where they know that Ukraine doesn't have the supplies to either get aggressive or defensive.

Yeah, and I mean, it makes sense that Russia's going back on the offensive because of Operation Spiderweb. There's a very aggressive Ukrainian offensive to destroy a bunch of Russian airborne assets. Now they're saying, you know, we're bringing the fight back to you. And so that continues. In other news, Tesla Global deliveries plunge 3.5% in Q2 2025. But I saw a lot of posts on X saying that this was not that big of a decline. So Amit says deliveries were 800 and 384,000 versus 389 expected. Analysts thought it would actually be 350. Where's the brand damage? Lowell? This guy's a fan of Tesla, obviously, but there are some, there is some data here. So for the Model 3, Model Y, they produced 396,000, almost 397,000 and delivered 373,000. And so the consensus was 394K marking the second straight double digit quarterly drop. And so the question is like, how much of this is about just EV cost or.

Yeah, Q1. Q1. How much, Tyler, how much were Q1 sales down? Because if you do two back to back.

Yep.

Double digit drops.

Yep.

That's pretty meaningful. That feels like brand damage. But it's also a very competitive market. Right. Every major manufacturer now has EVs in the market. And again, we've talked about this before. The biggest thing, it's possible, one of the biggest drivers here is that the lack of meaningful refreshes of the Model 3 and the Model Y and the Cybertruck is just selling seemingly well selling in beautiful Los Angeles. But it's the new very well.

But it's hard at that price point to break through to the real truck market. I'm pretty sure the, the ford Maverick, the F150, I mean the number one convertible in the, in the United States is Jeep Wrangler. Tesla doesn't have a convertible. They don't have anything that competes with the Wrangler. Right. Tyler, you got some numbers for us?

Yeah. So it looks like it's down 13.5% sales this quarter. Yes. Yeah. What about Q1?

Yeah. How was Q1? So year to date, Tesla's Europe sales are down 28% in May, and then in China they sank 7.86%. So there's a whole bunch of different narratives around like, like there's a lot of pressure internationally, but that's more about like trade. And then internally there's people that stop buying them for political reasons. But then there's people who started buying.

I think there's a big contingent in Europe that cares about the politics as well. The China problem is just the fact that you have Xiaomi and all these other players coming in with great cars, undercutting the market.

They got a Taycan knockoff, they got a Purosan way knockoff. They got knockoff. Yeah.

I mean, they're not exactly innovating on the silhouette.

Yep.

But they are innovating in the, you know, sort of internal.

Yeah, there's some crazy, crazy Chinese cars out there. Yeah.

So Q1 was down 2020.

Yeah. So I don't know.

So let me pull up the.

Where's the brand for Tesla and the rest of the Mag 7? I want to see how they're ranking. Here we go. We have our Mag 7 leaderboard and Tesla's at the bottom. $984 billion in market cap trading at 172 price to price to earnings ratio. So they're seventh. But if the Cybercap.

The most magnificent thing about the company is that the E ratio. Yeah, it's absolutely fantastic.

But if the Cyber Cab plays out, that is a significant, I mean you look at what is the market for all the different robo taxi drives probably in the tens of billions of dollars that Americans would spend. And so I think a lot of people are expecting that to hit and that to work. And that the fact that the Cyber Cab is at that 30k price point instead of the 300k price point for a Waymo being a durable advantage. The fact that they can produce, I mean the fact that they even produce make 400,000 cars in a quarter. It's 100,000 cars a month, 3,000 cars a day. That's a lot of industrial capacity. And so that has to have some, have some, some value and some leverage if totally the AI catches up and the Cyber cabin comes down.

Battle between Waymo and Tesla is just getting started.

Oh yeah. And I, and then a lot of people are buying the stock on the promise of humanoid robots is like it is, it does feel early. Optimus does not feel ready for production or ready to buy or ready to go out on the manufacturing line. But what other stock in the mag 7 would you buy to get exposure to humanoid robotics? If you think that's an important thesis. Like there's not really that much.

Yeah.

And even, and even in the, and even in the smaller like, like, like mid cap or small cap stocks, there really aren't that many pure play humanoid robotic companies that are out.

Yeah.

So there's not really anything that's sucking up that demand in the market. Anyway. Speaking of the market, Tesla Figma has dropped their S1 numbers are strong. $821 million in the last 12 months. Revenue 46% year over year growth 18% non GAAP operating margin. They turn profitable. $1.5 billion in cash, zero debt. 91% gross margins. Best in class for SaaS. $132% net dollar retention sticky. AF says Jason Lemkin 78% of the Forbes 2000 use Figma, 76% use Two plus products. Platform expansion from massive losses to 18% operating margins. This is how you prep for an IPO trading under Dollar Fig on New York Stock Exchange soon.

Absolutely. Absolutely incredible. Obviously Figma is a sponsor and partner of ours, but this story is really incredible. The other thing that Jason missed here that I think is interesting is Figma holds BTC on the balance sheet, which I think that retail is going to be excited by. There's also a really exciting AI story here.

Figma make almost $70 million in Bitcoin ETFs and was approved to buy 30 more million in BTC. Zach says this is the first company we've seen this cycle that's buying bitcoin and isn't bad. That's pretty notable. He used a different word. And Taney says Figma has 1.54 billion in cash on hand but only raised 749 million in primary capital in its history. It will go public with a negative net burn over its lifetime of 791 million thanks to their efficiency and the $1 billion breakup fee received from Adobe. Wow, that's remarkable.

Things are working out. Yeah, things are working out. I remember, I mean, you know, they obviously getting shot down by the Europeans on an acquisition that I think was exciting for everybody involved from Dylan to Scott to the cap table and the team and you know, Adobe broadly was I think disappointing for everyone. But I think this is going to be in hindsight I think people look back and say this was actually the right thing.

Totally.

And I think the markets are going to love another founder led company.

Christina had a good post here. The first thing I appreciated in Figma's S1, everyone said the S1 was the best designed S1 ever, which makes sense but was right at the start of Dylan's letter when Evan and I started Figma. Too often companies write departed founders out of a company story. Dylan did the opposite and that says a lot. I love it. Very classy. So you can go to figma.com, think bigger, build faster. Figma helps design and development teams build great products together.

Get started for free and you can check out Figma make. You can Vibe code whatever you want. Tyler is in the midst of vibe coding a new product that we're going to release a data product in the next couple weeks for free and I'm excited to share that. Otherwise I wanted to pull up, let's pull up the polymarket for the reconciliation bill. Zach Kukoff is going to be coming on the show later today to give us the update. Things are trending down across the board. The getting it done by July 4th is sitting at a 61% chance, down 6% over the last 24 hours. So we will follow this one. And I'm excited to get the update. It's funny, a big bill like this with this deadline of the fourth of July because there is this, like, overall incentive. I think people would like to enjoy their fourth of July versus, you know, angrily debating a bunch of small details.

I think the debt ceiling always comes up during the summer. I remember years ago it being a topic during summer. And it feels like a very weird thing that you have, like this ultra high pressure DC narrative during like, the slowest time of the year, the nicest.

Time of the year in France.

Yeah, yeah. It's very American. We grind it out here. We grind it out anyway. You don't need to be grinding on bill, payments, accounting, and a whole lot more. You can go to ramp.com Time is money. Save both. Save both. Easy use corporate cards, bill payments, accounting, and a whole lot more all in one place. Go to ramp.com we need to tell people about.

Sorry, I was gonna say our new ad that we shipped for wander. We shipped this 35 minutes ago. We should show this now live on location at a wander. And I would encourage people to just go watch it.

Can we pull it up on the screen?

Let's pull up a few seconds. Let's give a little.

A little. I think we can watch the whole thing.

We can watch the whole thing. Yeah. It's only 45 seconds. Let's pull it up.

Let's pull it up and watch it.

We went full golden retriever mode for this one, folks.

We go full screen on this. Is that possible? Go full screen. Well, we pull that up. Such a fun shoot. If you're a GQ fan, you probably noticed that we were heavily inspired. Maybe even you could call it a shot for shot remake of the. The GQ video with George Clooney and Brad Pitt. But I mean, fantastic inspiration from those folks.

It's pretty funny. We showed a preview of that to a few people and they'd go, is that a wander? We're like, yes, yes, obviously. But still, the property. This is a property in Malibu that you can rent this summer.

Yeah, I think we made it.

Absolutely fantastic.

Anyway, we got to tell folks about the TBPN derivative works clause. So if you're watching the show, if you're writing about the show, engaging with the show, you should know that we have a term of conditions, terms and conditions, terms of terms of use. We don't ask for much, but there is one rule that we take very, very seriously, our lawyers have been very serious about, which is that if you watch the show, if you engage with the show, if you write about the show, and then you go and create a derivative work. So a copycat, a live stream, a daily technology show, a clone. Yeah. We don't ask for financial compensation, nothing. We don't ask you to stop doing what you're doing. But you are legally required to drink one glass of municipal tap water on stream.

Just one, every time, Every time you do the show. If you were to copy us and do a daily show, you would have to drink one glass of tap water, municipal tap, daily.

Yes. And so a lot of people have been fighting this. They copy us and then they refuse to drink the tap water.

We.

But this is legally, it's only a matter of time until they tap out. They got a tap out.

Yeah, that's called tapping out.

Tapping out is when you, when you admit that, yes, you did copy tvpn. And so you're tapping out, you're drinking the tap water and it's legally binding. So, yeah, just if you're thinking about doing it, just be prepared and drink tap.

And we wanted, we wanted this policy to be fair. Right. It doesn't cost you anything.

No.

Maybe your health a little bit, maybe. But yeah, so, so feel free to copy us. As long as you tap out.

As long as you tap out.

Always tap out. And we'll be, you know, if you launch a derivative work, copycat, clone, whatever. We'll remind you.

Yeah, yeah. In the comments, everyone will remind you. Yeah. A lot of people say, like, oh, how will I know if what I'm doing is truly derivative? Like, is this or is there a definition? It's like, don't worry, you'll know. You will definitely know if you have to drink the tap water. Like we're going to make it apparent anyway. Let me tell you about Vanta Automate Compliance. Manage risk, Improve Trust Continuously. Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation, whether you're pursuing your first framework or managing a complex program.

I was talking with Buddy yesterday. He was walking by one of our billboards in New York. He sent it to me and he said that he signed up for Vanta on Friday because he signed a massive logo that I can't say for startups.

And they said Sock 2 compliance, you're.

Gonna need Sock 2 before we can dance.

Wow.

On Vanta. Now that's amazing and I'd love to see it. So shout out to Vant. Sorry. Shout out to Grant and Vanta for cooking.

Let's go a little bit deeper into Erebor. This is certainly the hottest story in our world today. It appears the Financial Times, I think what happened here is that they found the like a filing and then put the story together. So this is. There's a lot of no comments in here, but there's still some interesting facts that we'll go through. So Peter Thiel joins tech billionaires backing new lender Erebor to rival Silicon Valley Bank. And people have been talking about this new bank and the name and stuff. And so it had been out there.

But it had not been startup named after the Dragon's Mountain and Lord of the Rings.

So a group of tech billionaires led by Palmer Luckey, co founder of military contractor Anduril, is preparing to launch a US bank intended to fill the gap left by Silicon Valley bank serving startups including cryptocurrency businesses. To be named Erebor. The bank would be backed by high profile tech investors including Joe Lonsdale from ABC and he's a co founder of Palantir, according to people familiar with the matter. Thiel's venture capital firm Founders Fund would also be among the investors, according to two people like Anduril and Palantir. Erebor's name is a reference to J.R.R. tolkien's Lord of the Rings. Erebor is the Lonely Mountain whose treasures are reclaimed from the Dragon. Smoke, Lucky and Lonsdale, who were big donors to Donald Trump in the 2024 US president election, want the bit. Want the bank to take over the niche once occupied by SVB as the go to lender for riskier companies and cryptocurrency players that traditional banks might reject. We were talking a lot about like founder liquidity, but there really was a time when you raise a $10 million Series A and SV would almost always underwrite that if it was with a tier one to the tune of like a two billion dollar, two million dollar debt line. So you could just get an extra 20% as venture debt and use that to fund Capex or an inventory.

Yeah, venture debt, venture venture debt would always get a lot of heat.

Well that was because there were a.

Lot of used poorly.

Yeah. And there were a lot of aggressive lenders. But SVB was kind of one of the cleaner ones. Totally, yeah. So Erebor has applied for a national bank charter in the United States. A license that allows a financial institution to operate as a bank. The bank will be a national bank providing traditional banking products as well as virtual currency related products and services for businesses and individuals. According to the application, which was made public this week. Target market would be businesses that were part of the US innovation economy. In particular, tech companies focused on virtual currencies, artificial intelligence, defense and manufacturing. It would also serve individuals who work for or invest in these companies. It's also planned to work with non US companies seeking access to the US banking system. Erebor's co founders first discussed launching a bank after the collapse of SVB in 2023. According to a person familiar with the matter. SVB has been the main bank for US startups and their venture capital backers. Its assets were first told its First.

Citizens SVB is still operational in effectively name only. It's owned by First Citizens.

What's interesting is that it is like owned, like the brand is owned and the corporate structure changed hands. But different pieces of the firm got sold off, split off. I'm pretty sure like Morgan Stanley wound up with the venture debt group or something like that. Like different desks went to different places and they kind of parceled stuff up. Some of it was sold, some of it was just like hey, this person would rather be doing X, Y or Z at a different firm. And so maybe Goldman or Morgan Stanley is trying to get into a particular market and they're like hey, SVB has some great guys over there. It was kind of like a free for all for a little bit. So there's not too much in here. There's a little bit about who's running this company. So Lucky and Lonsdale were not expected to be involved in the day to day management of the bank. According to people familiar with their plans. It will be run by Co CEOs Jacob Hirschman, who previously worked as an advisor at crypto group Circle and Owen Rapaport, co founder and CEO of Digital Asset Software Air Compliance. Mike Hagadorn, former senior Executive Vice President at New Jersey based Valley National.

Let's give it up for senior executive.

Vice presidents for sure will be the bank's vice will be the bank's president. And so they're going to have co CEOs and a president on day one. What a stack lineup. Yeah, pt incredible. Palmer, Luckey, Lonsdale in the boardroom. Then you got co CEOs and a president. This is a stacked lineup. I think we're working on a graphic. I've been pitching it. The starting lineup for this new team is.

That's the beauty of business. It can feel like sports, but there's not the same kind of rules. Right. You know, you can put as many people on your team as you want.

There's no salary caps.

That's right. Most importantly, Juwan with Juwan was, was posting. He was like, we're gonna see salary caps. And I hope not because that would imply.

Well, Mamdani's trying to do that. Right. He wants the salary cap to be like $999 million. Like he wants to be a billionaire. Bernie too also wants salary caps in business and net worth caps. Net worth caps? Yeah, much like a salary cap. Imagine if there was a net worth cap in the NBA. It's like, yeah, like, okay, you've made enough money. Like you can't. You have to play for free now. Its head office will be in Columbus, Ohio, which we talked about. Part of the application was submitted confidentially and has not been made public, such as details related to its shareholders, equity structure and business plan. Lucky did not respond for requests for comment. Lonsdale confirmed he was a financial backer, but declined to comment further. And the other folks did not reply. Anyway, I'm sure they're going to be building some projects and you know what they're going to need? They're going to need linear. LINEAR is a purpose built tool for planning and building products. Meet the system for modern software development, Streamline issues, projects and product roadmaps. Xai Greg Yang. Great poster. Someone DJ Kaus says another great account says visiting the XA office really humbled me. And it's just tons of tents around a. Around desktops. Is this actually the XAI office? I'm so confused. And then Greg Yang, who actually works at xai, says, that's not our office. We have way more tents. So I have no idea where these original pictures came from, but it does look like a lot of tents in a. Inside of a.

The Internet's a tough place these days. You never know what's real. Right. They could have just made up tents in a tech office.

Yeah, it could have been a prompt.

Could have been a prompt. But I do know that the XAI team, I have it on good authority that they did. There are tents in the office. Sleep in the office.

Okay.

Not necessarily reporting every day. Someone else was giving them a little. Was chirping at them because they were saying if you stay up coding until 3, 4am Sleep in doesn't really count.

You're just nocturnal. What's impressive is actually working really hard for 12 hours, taking a break, having some rest time and then getting a full eight hours of sleep. And it doesn't really matter what cycle that happens on. I guess the benefit of the nocturnal programmer is that there's just less going on in the world so there's less distraction.

Knowing that you're grinding while your competitors are sleeping, that's priceless.

Yes. But I mean truly, truly there is like email is quieter during the night, right. And the news is quieter. So you know, it's like there's just less opportunity to get distracted by like the, the Elon Trump blow up that happened when they were sleeping apparently. So like yeah, they woke up and they were like okay, I guess something happened with our boss and the President. But you know, the story is kind of over. So let's get to work. Anyway, if you work at xai, you're going to have to sell this to big companies. You're going to need Adeo Customer Relationship Magic. Adeo is the AI native CRM that builds scales and grows your company to the next level. Get started for free. Anyway, we can go deeper into Microsoft. They are scaling back their ambitions for AI chips to overcome delays. The Microsoft is focusing on less ambitious chip designs through 2028. They hope by scaling back some of the designs and pushing out the schedule for other AI chips it's already working on, it can develop them more easily. The hope is that when these chips are released over the next three years, they will still remain competitive with Nvidia's AI chips. The decision follows delays in in house chip development. This is from a report in the Information. Microsoft wants to reduce reliance on Nvidia for AI computing. But if we pull up the MAG7 chart, Nvidia doesn't seem to be affected. They're still at the top of that baby at 3.8 trillion, but Microsoft's coming for them at 3.6 trillion. And so it's neck and neck. If Microsoft can fab their own chips, TSMC design their own chips, rip out Nvidia, that might be enough for flippening.

Yeah Matthew from Cloudflare yesterday said every time we've seen this type of scale up on the silicon side there's been and ended up being a glut. So we'll see. Praying that this time it's different.

Yep.

Would hate, would hate, love bubbles, Hate the crash.

Well we have a funny post here from Zach quote Tweeting someone says, you attract what you fear. And Zach says, ah, I'm so scared of $3 billion and drunk cigarettes with her. I love that you put this in here.

Zach's been on a roll.

Anyway, if you're selling anything, including cigarettes, you're gonna need to pay sales tax and you're gonna need numeral. Go to numeralhq.com also don't sell cigarettes. It's bad sales tax on autopilot. Spend less.

Says the nicotine. Says the nicotine.

Very different product.

No, it is very different.

Spend less than five minutes per month on sales tax compliance.

And we use to see it is the sales tax automation tool of choice for Ridge Grund Graza. The list goes on. Go check it out.

That's great.

We have a post here from someone.

Put us in the truth zone.

Andy. Yeah, 2 cents money 2 cents. We got to have him on some time to talk about his. Have you seen the screenshots from his app? Yeah, it's like Twitter but you're now just plaid.

Plaid and its consequences have been a disaster for the human race.

But anyways, he has a post here.

He says, can you imagine that this would result in a social network around your net worth?

He's like, no, it's the magic of building infrastructure.

I just wanted to make things a little bit smoother.

Anyways, he was putting Kylie in the Kylie Robinson from Wired in the true zone. Says claims to be wired wearing AirPods question mark.

Good question. Kylie, you gotta answer some stuff. This is big.

I am a firm believer in wired headphones. Try them out.

I thought you were gonna say wired the newspaper or the magazine?

If I it's terrible to say, but if I could only have one, I would take wired headphones.

Yep. Oh wow. Well, she should get wired headphones because that would be very on brand for wired. Anyway, we have announcement. We're using Fin AI now for customer support. We've gotten a lot of questions coming into our customer support line. People asking, you know, are you guys normal heights? And Fin can just immediately chime in. Of course these. Both of them are normal heights.

What do you think about. Yeah, we need it. No, we need to set up a support@tbpn.com for sure. Running on Fin.

Yes.

To answer these, press these pressing questions.

There aren't enough ads. Can I get a bespoke ad and it will just automatically generate an ad for you?

Yep. That would be what Jordy really 5 1.

Do you guys really. Would you really take a bullet for big tech? It's like absolutely. It's like Finn we were joking about.

Relax.

No. Finn is number one in performance benchmarks, number one in competitive bake offs, number one in ranking on G2. It's the number one AI agent for customer service. Go to Fin AI to check it out. In other news, Microsoft's laying off 9,000 workers. This is a short article, but we've already covered this. But it's interesting to track. What's interesting is like, is Microsoft the only one that's doing these big riffs right now?

It feels like normally this is what, this is what, this is what scares me.

Yeah. What is software?

There are so many, there's so many top signals right now, just everywhere you look. Top signal, top signal, top signal. I want to believe that this time it's different, but our industry goes through cycles naturally. Satya is a seasoned operator. He has incredible traction in generative AI. He's getting 20% off the top of whatever OpenAI makes. You would think that he would be going on a spending frenzy, yet he's conducting routine layoffs, he's giving up data centers. He's saying, yep, you guys, I don't need it. Even though yes, they're so I just think he's being, he's playing it. He's simultaneously being, he was being really aggressive like two years ago and now he's starting to play it a little bit safer. And I think in some ways, you know, some of these rifts are just practical.

Right.

He's realizing, yes, AI is making us more efficient so we should need less people to do more work. People were putting Benioff in the truth zone last week because he was saying AI is now doing 30 to 40% of the work at Salesforce and yet his headcount has just been ticking up steadily. And so, yeah, this Satya getting hyper practical scares me.

I mean he's. I'm pretty sure he joined Microsoft in the 90s.

Yeah. So he's been through his face.

It's not like he's just lived through the dot com crash and the global financial crisis. He worked at the same company and saw how Microsoft handled those two crises. I think Microsoft handled both of them pretty well. But he can totally pattern match. Maybe he's wrong, but if he's still. If he's wrong and there's no pullback, there's no glut of chips like the AI continues to progress. He's still set up pretty well for it. Right. He's still major capex, serious hyperscaler, huge bet on the Best horse in the race, basically the best independent horse hitched to OpenAI. Even as that relationship gets more and more complicated, it feels like he has a deeper connection to OpenAI than say, Google does with another startup in the foundation, Model Lab or Amazon or any of these companies. Because they've got all kind of like Anthropic has deals with Amazon and Google and they're starting to kind of warm up to being dance partners. But Google's not getting 20% of Anthropic's top line. It's just not that the deals aren't structured that way. So he has not only the leading lab, but the most dominant stake of any hyperscaler in a lab, period. So incredibly well set up. So we will continue to track it. But whether you think it's a top signal or a bottom signal, you got to head over public.com investing. For those who take it seriously, they got multi asset investing, industry leading yields, and they're trusted by millions.

And they have a Triple Ira match.

3%, 1%, 1%. Check it out. In other news, the talent wars continue to rage on. Sam Altman slams Meta AI's talent poaching spree. Missionaries will beat mercenaries, he says, quote, what Meta is doing will, in my opinion, lead to very deep cultural problems, said Sam Altman in leaked memo sent to OpenAI researchers. So big, big question. Paula says with all the AI labs and SF being in the mission, who called them members of technical staff and not missionaries? Oh, wow. This is an old post. Yeah, from. From over six months ago. And there's a bunch of other new details emerging from the new Meta Super Intelligence team. Didi, who's coming on the show right.

In like three minutes.

Okay. Wow. Yeah, we've highlighted a lot of his posts. Every single one of the 11 meta meta super Intelligence hires has an is an immigrant who did their undergrad abroad. Seven from China, one from India, one from Australia, one from UK, one from South Africa. Eight are PhDs or PhD dropouts in the US. Immigration is key to US AI innovation. Beth Jesos has a hilarious post here. The foreshadowing here was insane. And it's Sam Altman interviewing Mark Zuckerberg. And Mark Zuckerberg says, the thing that I think Facebook has done exceptionally well is hiring. This is an incredible series, the Startup School series that Sam did while he was at yc. He interviewed Zuck, Elon, a ton of really interesting folks and got these really definitive interviews from the top tech leaders of the next generation right as they were kind of on the biggest moment in their come up. Why are you laughing?

It's just I love, I love this business.

You're like tearing up looking at how happy. How happy.

No, it's just, it's watching this great power conflict play out in real time is wildly entertaining.

It's entertaining.

And the consumer is going to be the net winner.

I was about to say that. And also America. Yeah, it's extremely, this is a pressure cooker of the highest. It's, it's so competitive. But the good thing is that at the end don't like we aren't playing this game where someone goes to jail or someone gets, you know, like goes to the gulag if they don't perform or they get locked up. Like this is, this is, you know, the difference between being a, having $10 billion or $100 billion. It's like it's kind of all fake but it's all, all, it's all like, it's super high stakes but it's still capitalism and it's still like the consolation prize is still being able to build something cool even if you get beaten. But the end result is that we have this hyper competitive race to build the best thing with like kind of a safety cushion so that you're not like, it's not like the government's and it's global.

Right. It's greatest minds from all over the world coming to compete in this market.

Yep. Yeah.

In this talent pool. And many of them are getting maxed out contracts.

I'm sure. Both Mark Zuckerberg and Sam Altman are losing sleep right now over the talent wars. They got to get on Eight Sleep. Go to Eight Sleep.com TVPN. Get a new Pod 5 Ultra. Sam. Get a new Pod 5 Ultra. Mark. They got a five year warranty, 30 night risk, free trial, free returns and free shipping. So go check it out.

I'm back in the game today. I got an 89. I got massively dinged on my consistency which has just been all over the place. 90 minutes of deep sleep, two and a half hours of REM sleep.

I mean you won by default because I slept in Malibu. I'm moving the eight sleep after work so I don't have anything to react because I was size gong for me.

Size gone for me.

Congratulations, Jordy. You deserve it. You deserve it.

All right, without further ado, so let's bring in Dee Dee. Hey, how you doing? What's going on?

Hey guys, how's it going?

Great to have you on the show finally. I feel like we've reacted to almost every one of your posts. You're very, very helpful in terms of contextualizing and just like summarizing everything that's going on, giving breakdowns. It's really helped the show, honestly. So thanks for being here.

Totally. Thank you for being a core contributor.

I'm a big, big fan of the show.

I've seen all of your episodes, I think.

Wow.

So you guys are great. You guys are great. I'm so happy for what you guys are doing.

Thanks, man. I. I really appreciate that.

And did you join all in? What. What's this in the background? Are you.

Oh, yeah, this.

This message has really been overloaded since the podcast. But no, that's just our. Our Venture message. I mean, this is what Menlo. Menlo's slogan is.

You were using that slogan before 2020.

It was. It was no way.

They mogged you, they stole your line, they just steamrolled you. This has happened a few times. What was it? Oh, Gemini. I thought of it for years as the Winklevoss backed crypto trading app exchange. Now Google comes out and they're like, we'd like that brand actually over here. And so now Gemini, I think of as the Google product.

They went from Bard.

Yep, Barded up. They barred it up. No one was competing for Bard. Bard was clean space. But they had to go Gemini. And so good luck to the Winklevoss twins. Anyway, why don't you kick us off with like a little bit of background and kind of what you do just to kind of get everyone familiar?

Absolutely. Well, my background, I'm in Venture now, but my background has been somewhat non traditional for Venture. I spent most of my career as an engineer and a product person. So four years.

If you say non traditional, I expect warlord, Blackwater, arms dealer, kite surfer, or like someone who's like wing surfing, rock climbing.

I actually thought. I actually thought you were joking for a second because I thought you were going to say my non traditional background Adventure. I went to hbs. I was a product manager. I just found my way.

We're not letting you get away with that, man.

You would be hard pressed to find an actual person who's done engineering at a big company and a startup from the very beginning, but not like strategy and ops role for two years. I actually don't know that many people like that. I do know it's meme though.

You were in the trenches. You were in the trenches. You were in the technology trenches. Okay, okay, we'll give it to you.

Yeah. Okay. So yeah, how do you. How do you land at Menlo. Where you doing? Were you doing investing before? Angel investing? What did that look like?

No, I actually had no business being in venture. I never wanted to be in venture. I wanted to build stuff. But after Glean, I, I thought, you know, I was exploring, building something new. And then I was talking to a bunch of these venture firms just get the, get their thoughts on it and, and I was kind of roped in and I thought, hey, it's kind of strange that there's not as many people who are technical and really care about engineering stuff in venture. And it would be nice to have that because whenever I would speak to people, they'd be like, look man, like I don't really understand this technical stuff, but is this a big market? You know, like, is this gonna actually work as an idea? And that's kind of the ethos of why I wanted to join a venture firm in the first place. Like, like, I think it'd be valuable for technical people to talk to other people who really care and nerd out about the actual idea. So I've been doing that for a year and a half and it's been, it's been more fun than you think. You know, venture has a lot of connotations on what the job involves, whether it's a sexy job or not. And I don't know, it's just very different for very different people.

I like to believe.

What's been your strategy the last year and a half? Has it been trying to find and lead deals yourself? Support a mix of both. What is the.

Write a check, fire the CEO, take over?

Fortunately, no contrarian.

Everyone's founder friendly now. You got to differentiate by being unfriendly, very founder unfriendly.

If you're non technical, then you're going to have problems.

No, I'm kidding. No.

There's two things we do. I, I have led deals and, and also co led deals with some of my partners here. We've done about four. Three of them are announced. One is not in terms of series A's, but mostly like I got brought in to help lead the Anthology fund, which is our $100 million fund with anthropic. And that's been a lot of fun because it gives you a lot of leeway on the kind of deal structures that you can have. Not your traditional like hey, high ownership necessarily series A's and B's and C's. It's more early, more, hey, how do we help the ecosystem? How do we just get into ideas? Very, very early. And just meet founders so you don't have a two week process where you're trying to fall in love with the founder for the first time.

So interesting. Yeah.

The other thing, I'm sure it's been massively helpful because one of the challenges, if you go from operating and then you join a fund and you're only able to do a few deals a year, it's like, how do you build up that those the frameworks, how do you really learn the craft of investing and learn what not to do? And I think one of the there's plenty of stories in the valley of people that our friend Justin Mares, I think like his first two companies he ever angel invested in were unicorns. And there's a bunch of other stories like that. From my personal experience, the average deal that I've done has just gotten better and better across the 60 plus companies that I've invested in. And so that can be a really good big challenge. If you're thrown into a big firm, you're having to make big investments and yet you're only able to exercise the actual check writing a couple times a year. How's the anthology fund going? Are you guys.

Specifically, I want to talk about, you said like non traditional deal structures. And when I think of that, I think you can come to someone with say, hey, we'll do a $10 million Series A, but we'll also give you $10 million in credits. We've seen Nvidia do some of these deals. We've seen obviously the Microsoft or the Microsoft OpenAI deal was kind of structured as some sort of cloud credits and all this different stuff. And that can be really valuable. Is that what we're talking about here or is there something else or are cloud credits even relevant in like traditional deal structures these days? For AI companies that are building on top of foundation models, we do offer.

A bunch of cloud credits. So we go up to like 30 to 100k in cloud credits. I, I honestly don't think that's the main draw. It's like nice to have for these early companies. All these good companies can figure out how to get credits or how to spend money on this. I think what's more interesting is, you know, even when I was like angel investing prior to being in venture, that's a different art. Like it's much easier to get into deals as you guys probably know when you're angel investing because you just know the person, you're like, hey man, I'm gonna throw in 10k. Can I get a 10k allocation. They're like, yeah, you seem like a good guy, you can be helpful.

Here's, here's, the bar is just so much lower.

The bar is very low. And then you go to a venture firm in the same companies and you tell them, hey, I want to lead you around. And they're like, good luck buddy. Like here's a line, get to the back of it. So that's what becomes really challenging. But to the question of Anthology specifically, I think probably the biggest draw from what the founders tell us is hey, we just love to know what the labs are up to, how to work with them, how to plan for the future because we have no visibility into that. And for us it really helps because you can form a relationship with people and actually figure out, hey, you're an awesome founder. And I don't know that because you've more than just the three week process that you've pitched to me. I know that for a good six to eight months now. So we, you know, we've deployed about 30 odd checks to Anthology. I would say in the beginning it was a little bit of calibration needed to figure out, hey, what was the right kind of company. And we don't only do early, we do some like mid stage companies that are interesting to us for a future round as well. And the two of them have graduated into full series A leads. So that's, you know, what was our success metric going in, which is are these actually good deals that will graduate to real ownership checks or not? And so we had Good Fire and then Open Router sort of be those companies that graduated. Very different companies.

We had, we had Alex on from openrouter and neither of us, I guess realized that they were announcing his massive series B. Oh yeah, yeah, yeah. So we had them on for like a full 15 minutes on the day.

So I know, I know about opensea and I know how good Open Router is, but I didn't realize it was like somehow the news got lost that it was like, oh wow, it's like a massive, massive fundraising day. I'm interested to know more about corporations or large scale ups, startups partnering with venture firms. We've seen a bunch of different, I mean going back to like the Slack Fund, I remember Slack was investing in startups and there's this question of like, for a lot of founders it can be fun to have, you know, on balance sheet investment vehicle. A lot of founders have, are scouts at VC firms or they invest off their own balance sheet. Why would a big company come to menlo to do the fund and it seems like you're offering a lot of the work and is that the right model? Should we see more of that versus the fully in house effort that's staffed? I think Stripe's done some stuff in house. Perplexity has a FUD now and I'm wondering if you have an opinion on best practices or what works or what doesn't or when something would work on balance sheet or directly within the org versus outside.

I think it's just. I have opinions on that. Yeah, I think personal opinion is you need a fortress balance sheet and you probably should be like a public company to be making balance sheet investments purely because of the duration.

You mean like Google Ventures?

Basically if you lead or participate in a series A and then don't see a return, you're not going to see a return from that for maybe a decade.

I mean Bloomberg Beta has some other stuff like that. There's just so many fascinating different like paths to it. And then it also is this.

I like this because it's like hey, this is non core to what we're doing but it's important to the ecosystem we're building. Let's make it so that it's not our job to deploy this capital but we still get a lot of the benefits.

Totally. So yeah, I mean give us your take.

So I think of look, we did a lot of analysis on what a corporate fund structures look like and what kind of investments they do in the past and these are all gross generalizations. But generally if you run a corporate fund a you're not compensated on carry really. So the incentives are not really aligned to deliver returns. The incentives are pure ecosystem bets. So if you look at a bunch of the investments that Slack fund have made, look, yeah, there are slack apps, a lot of them and that's good for the ecosystem. But are they real companies? Are they real businesses? A lot of them aren't. And when the more you look at corporate venture like they don't spend a whole bunch of time out on the road sourcing really like you know, balls to the walls going to all of these people and trying to figure out what the good companies are. The reason we wanted this structure was a, so we didn't have that incentive problem. We could focus on real returns while also investing in companies that benefit the ecosystem. But not purely that. So a lot of the anthology companies, they don't even necessarily only use anthropic or use anthropic at all or necessarily even call LLMs. They are generally companies that are beneficial for the AI ecosystem. So that's one. And then number two, I think the bigger thing is most companies have no business running a venture firm. This is just not, like you said, it's not their core business, not their forte. It's not what they do. It's not something they care about. So it's really nice where you're like, okay, you guys care about it. We care about the ecosystem benefits. Why don't you help us run this fund? And then it works for everybody. And I think the one exception case to that is probably OpenAI. OpenAI startup fund has done very well, but, you know, not in small part due to Sam.

Yes, Sam's kind of a good player, potentially a generational vc if he was doing that full time. Yeah, yeah. What about the competitive dynamic? Like, like, would gv? I mean, Google Ventures is kind of spun out so much, but, like, it would be. It'd be crazy to think about, like, GV investing in Perplexity or Bloomberg Beta investing in Perplexity, even if it's not. Not actually that much of a threat. And they're. And the CEO is confident, like, hey, I'm not going to get disrupted by this. It's like you have a portfolio CEO who's out there tweeting every single day, like, I'm coming for you, Bloomberg. Let's look at how much money they make. I'm going to disrupt you. Like, that just might be like, awkward. And so has there been any dynamic around that where you. Where you have like either a mandate to like, hey, let's stay away from Preacher, from companies that do pre training because that's on Anthropic's roadmap. Or like, like, even with the clog code cursor thing, like, you want to stay away from coding agents because Anthropic's really, really, really pro that. Like, how do you think about, like, you want to be close enough that there's a benefit, but not too close where your competitors. Is that. Is that a real dynamic or how do you think about that?

It very rarely has happened. I mean, Dario is actually exceptionally missionary. I mean, I. I don't know how else to say it. Like, he really does care about what these companies are doing. I mean, look at Good Fire. Like, like they have an interpretability lab. They're like, look, another one is good for the world. We'll do it. So there have been certain cases where companies, you know, some of them haven't been announced yet, but one of these companies has alternate approaches. To doing a bunch of different model stuff. And we asked Arya, we're like, hey, is this bad? Like, are you okay with this? And he's like, completely cool with it. If it's not something that we have a very strong bet on that they're going extremely head on and have a bunch of capital to go head on right now, it's just not something we'd care about. So we do try to tread carefully. Obviously you don't want to bite that hand, but it's not been an issue at all so far. We've never run into a conflict.

How are you broadly advising companies that are working with labs, but also fear that they might end up competing with them at the product layer over time? And I know you were at Glean, the founding team there, and that feels like something that in the long run, obviously, you know, someone like an OpenAI might want to compete there. So I'm curious how you think about that dynamic.

I think the reality is, look like it's a free market and competition is going to happen one way or the other, whether or not Anthropic backs or Anthology backs you and then Anthropic decides to compete at a future point.

I don't think anybody, I'm not even talking, I'm not talking about Anthology as much as just generally portfolio companies. Let's say you're doing Codegen or something like that and you're working with a lab. And we had. Sarah Guo was on the show probably a month and a half ago at this point, talking about it's not impossible to figure out what's important to each of the labs and the things that they might do over time. And you should, as a founder, you should be well aware that what you're doing might be on the path at some point.

Yeah, I think that's a question we get all the time. And I think the reality is my guidance usually has been, and this was how we thought about it at Glean, too, is what is something that is a pie that is very valuable to you, but not at the scale that it's valuable to any of these big companies. Until it is, you know, like, you're really. The bet you're making is, hey, this pie is maybe $50 million in revenue and nobody, no big company will care about that. But once you capture that pie, it's actually way bigger than that. And if you do, then you can, you can reach breakout, escape velocity before anybody else is at, at your heels. And hopefully you've built off enough of a, you know, no I hesitate to say moat, but you've built enough tech that it's not easy to replicate overnight. And if you have a little bit of that, there are odds that you could have a breakout business. And I think today's climate is actually way better than a year or two before. Because the truth is labs have a huge revenue pressure. So they're not going to be investing in random crap. Like they're going to be investing in stuff where they're like, okay, that's a billion dollars in revenue, that's $500 million in revenue. I'm not going after this $10 million opportunity.

It's just not, yeah, anthropic one to $4 billion run rate in the first half of this year to justify a lot more resources. You have to be able to make a dent in that somehow or show that you can make a dent.

I mean, it's kind of like the Google problem with like, but at one order magnitude smaller where like, if you're running a business at Google that's making $500 million, it's like, okay, time to wrap it up, let's move on to something bigger. This is not really relevant. And you're like, this could be a public company. I don't wanna, Yeah, I don't know.

If you guys know this, but please started Google, had Google cloud search. And the reason we weren't worried about it is because we knew the team there and we're like, this doesn't make a dent on cloud. The fact is like Google cloud doesn't care about this. And now, you know, sort of, they sort of change their mind and they care about it again. But that's kind of what the sweet spot. You want to do something boring and seemingly small enough that nobody else cares.

Can you talk to me more about Glean? There's been some reporting about like the data wars. This question of what, sir, how sharp elbowed will Salesforce be or Google if I'm a business owner. We just started TVPN as a company and it's a greenfield. So you're like, this time I'm gonna get it right. This time I'll just have one suite of tools that'll all work together. And then three weeks later it's like, okay, I have Slack and Gmail and a bunch of other services and they're not really talking to each other the way I want. And as a, as a business owner, I'm very strong about like, that's my data. You have to let me have my data and not through this export button that takes two weeks to go. And I, I want an API that I can immediately send over to, to OpenAI or Anthropic or Glean. But they, the companies always find a way to kind of like have sharp elbows. And we saw this during the, during the rise of Google and Facebook. There was a pitch, John Patel, the founder of Wired, was arguing that Google, you should be able to go to Google and say, give my Google search results to Facebook. And Facebook you should say, give my social graph data to Google just so I get better ads. And both companies hated that idea and they would not let you do that. It's a little bit different in the B2B context, but how do you see the data wars playing out? How do you see the competitive pressure where a lot of the B2B SaaS providers also want to offer some sort of enterprise search experience?

Well, there's a couple of ways to look at this. I think number one is, yeah, the reliance on Glean's reliance on just Salesforce is a small percentage of the many, many connectors we offer and the many, many other things that we end up having to integrate with for especially Fortune 500 clients. They often don't even use a lot of the SaaS SaaS tools that are household names in, in the Valley. So the net impact is, is somewhat limited because of that. But I think the more important point here is the reason why, it seems like the reason why people like Salesforce want to cut off connector access is like what you said, to sell a competing product. In practice, I think it's actually a pretty synergistic relationship. Otherwise, that's why we work together so well for so long. Because you don't sell less seats because people use Glean. You actually sell more seats because people use Glean. Because now there are people who weren't on Slack, weren't on Salesforce, who are now discovering that information and going there and then reading the, reading the content. So your daily active users actually increase, not decrease when they use Glean. In fact, like if you don't have licenses on Salesforce to a certain set of users, they don't even see Salesforce results. So it's not like you can also sell 10 licenses, connect to Glean, and then have the thousand people in the company view those Salesforce data, that Salesforce data, we don't do that. We don't even offer that as an option because we want to stay true to the permissions of the source data source. So, you know, net Net. I think this is a situation that, you know, Glean will have to work out with all of these partners. But I think there's a good out. I mean, I don't think there's a reason other than them wanting to offer a competitive product that we, that they care about this access and the companies too. Like you said, like, if you were running TPPN and you want to connect to Glean, you should own your data. You should own your Slack and Salesforce data. They shouldn't own it. So you have every right to give away that API access. But you know, that's going to be a conversation that Glean's going to have to have with a bunch of the partners.

Where are you most excited about agents right now in terms of actual products in marketing, delivering value for customers outside of coding?

Okay, so this is the one I talk to all my friends about. It's not like a venture venture thing. It's just really a product I just have fallen in love with in the past week or so. You guys should all use it. It's called. And I'm not a paid shill. This is like just me loving a product. It's 19 pine AI. It's super elegantly designed, simple chat interface, your typical stuff. But what it does is it just goes and makes phone calls for you, but can coordinate very interesting things. So for example, just, just the other day I was like, man, my, my insurance on my car is just way too high. I need to get this down. Can you figure out how to get it down? And it goes and it like, these are note phone calls nobody wants to make. Goes and makes a phone call to my insurance, stays on hold for three hours and then gets them on the phone calls. GEICO calls State Farm, gets quotes, comes back and goes and negotiates. And then it brings my insurance down. I did nothing. And the payment model is so, so nice because you don't pay a subscription for this. You pay a percentage of what you save, which is agreeable to everybody.

Yeah, of course.

And so, yeah, you would do that.

Deal all day long with like, yeah, just call this person for me. If you save me like a thousand dollars, like, I'll give you like a few hundred bucks. Yeah, something like that. But yeah, I'm looking. I have the website pulled up. They got to get a better domain name. This is. This is absolutely brutal.

What's the domain name?

1919Pine AI. But they can call the IRS for you. Cancel your AT&T plan, talk to the DMV, get a refund from an airline.

I'm Gonna use this. I literally have an old Internet line in our old office that I need to cancel and I can only call during business hours and I just happen to be live on TV during business hours every day, so I just haven't done it. And I'm just getting billed for Spectrum every month and it's so annoying. So 19 pints coming for you. Spectrum, you're on notice. Anyway, I want to get your reaction.

But any other like this feels awesome. As a consumer agent that is solving real problems and they have the economic model that people have been excited about of like selling, selling results, any, anything else in the, in the enterprise, like sales agents, any, anything like that, that, that you're seeing that is actually working.

Look, there's a bunch of interesting companies and the revenue is inflecting, but you know, that's just not like it's good. Like these are good businesses that are working. I think the thing is things that, that excite me more at a more meta level is the, the technology powering what agents can do. And like, I think the elephant in the room is all of these, all of the new custom RL stacks that are helping agents use custom software is actually going to be a huge unlock. You know, like computer use hasn't really worked yet. It's not a thing that people use. But there is a world where, you know, just like 19 Pines, you could genuinely automate the work of people if you give computers access to your computer in a fast way. And I don't think we've seen that yet because the models aren't capable of doing that. We've seen early versions of that with Claude code, which is, hey, what happens if you take a model and then RL it to be very good at agentic work and call a bunch of tools. Well, O3 is similar. What if you apply that to very specific domains of like, hey man, like I have a gazillion enterprise SaaS apps that I don't even want to use and I don't even know how to use or touch figure out how to use them. So we can just throw these agents at this stuff that a human was doing. Because a lot of the glue work is still not solved by agents. Like a lot of the glue work is humans in the loop. Go click this button, go review this text. If we want to fully cross the chasm of like 95% of the work to 100% of the work, I think that's super exciting at a meta level. And all of these little agent apps and sometimes big agent Apps are going to benefit from things like that. So that's one of the broader themes I'm super super excited about.

I want to get your reaction to how the AI talent wars are affecting venture specifically. Will Menitis, friend of the show has a post here. He says Citadel's law any any industry with sufficiently high stakes will end up mirroring the culture of hedge funds funds massive cash comp for the top performers bid away dynamics where entire teams walk together extreme litigation of non competes hard o work 247 culture. But then he says he ends with this at the margin I think makes the capital environment for funding independent big AI things much worse. Not worth underwriting a team if you know they can do R and D on your dollar and then walk in the same way. My sense is fund staking terms got worse not better post pod Shop so he's talking about how in. In. Yeah, like, like if you are betting on a founder and then they get some massive, massive acquihire offer, is that going to distort the early stage venture market?

Oh, I mean I think like absolutely. I've read that post. Love that post. Great, great commentary. I think the thing I would say is this isn't new. I mean there were always startups never paid well right? Like you're still getting paid like 170 200k on average and you know Google and Facebook are right there to pay you like a 500k salary. So it's not like it's new that big companies can can, you know, throw out their wallets and give you more money Usually even at Glean we had this policy where we're like look, we're just not going to negotiate somebody who only cares about comp. If we lose that person, that's because they're a mercenary. And that's okay. We're not judgmental, just you do what you think is best for you. We really want people as missionary as you can be about like B2B SaaS but we want people who are, you know, more incentivized to do the thing that we're doing here.

So.

But that does $10 million a year paychecks affect that 100%. Thankfully it's just a few people who are privy to that. But if there's more of that happening, it becomes a risk. I mean absolutely. The one thing that I do the most when I do diligence is aside from the typical venture stuff is I really need to have confidence that this founder, no matter what, wants to work on this problem for the next five years. I mean, to the extent that you can get that confidence, that's what I'm looking for. But you can't solve for it. It's definitely worse. People are always on the spectrum for missionary and mercenary, and if the dollars go up, the spectrum shifts. So there's fewer people who can afford to be missionary if you're getting paid 10 million a year.

Yep. Last question from me, and then we'll let you go. Do you think the AI talent wars spill over into the rest of the Mag 7? I'm. You know, Mark Zuckerberg is founder, Mode CEO with a massive business throwing off billions of dollars in cash. He took. You know, he's spending $20 billion on the metaverse. Was able to slide some of that over, put some crazy offers on the table. We keep joking that some of these meta AI superintelligence offers are more than Tim Cook's annual salary. So that might be a culture clash. Even if Apple says, hey, we want our own superintelligence team, and we've earmarked $5 billion to do it because they could. Apple has produced almost a trillion dollars in dividends or like, free cash for their investors over the past decade or something like that. 13 years. And so. So the logical thing would be, okay, maybe we will see similar bidding wars, talent wars continue at Amazon, Microsoft, Apple. But maybe there's.

Or it just forces Apple to effectively outsource the R and D by paying OpenAI or anthropic.

That's totally possible too.

Billions of dollars.

So, yeah, how do you think that shakes out?

I don't think more firms, more Max 7 companies are going to compete like that. I think Zuck has a particular style and he always has. I mean, if you look at. At like you said, like Oculus.

Yeah.

He's very happy to pour billions and billions of dollars into something with the hope that it will one day work. Most other Max 7 companies don't usually operate like that, and they can get talent in many cases for, for, for way cheaper. I just don't think it's. It's also not. Apple might be able to do it on their balance sheet, but it's not just something that everybody can afford to do. It causes a lot of cultural rift within the company. If you talk to. The other day, I was scrolling metas blind for. Through a friend. I mean, people are upset. You know, like people get upset when you. You just hear that you can recruit somebody as an engineer and pay them $25 million a year. That's not a good thing for the culture of the company. They also check out mentally inside because they're like, what is the point? Why am I doing this? And so I think, yeah, there's one.

Frame where it's like, oh, wow, we hired a 100x engineer. And there's another one which is like, so, so I'm a 1/100 x engineer. Maybe I should do 1/100 x the work.

The right way to look at that. You have some top 25 AI researcher in the world joins your firm, you should be excited because you're already compensated really well and it will probably increase the value of the company in the fullness of time, which directly benefits you. But of course, if you're making, if you're only making 80k a month, you know, not even, not, not even hitting that, you know, seven figure mark. Yeah, that could be a little bit, a little bit frustrating anyway.

A month.

Yeah.

Well, thank you so much for talking about 3k more.

They could be in that.

The fact that you know the MRR numbers like the back of your hand is hilarious. This is fantastic. Thanks so much for stopping by.

This is super fun. Thank you for joining. Let's do it again soon.

Love it.

See you guys later. You might have seen our billboards are tearing up New York City. We have our New York City correspondent joining right now. But if you want to run a billboard in New York City, get on ad quick dot com. Out of home advertising made easy and measurable. Say goodbye to the headaches of out of home Advertising only combines technology. Out of home.

We went from the west coast to the east coast. Emily should should take over the west coast.

Oh, feed me billboards in la. We will ask but first we have to sing her a song.

Happy birthday to you. Happy birthday to you. Happy birthday dear Emily. We're hitting the size gong for you.

Congratulations.

It was yesterday.

It was yesterday. Yesterday. But July 1, she didn't come on the show yesterday, so welcome.

I would have come on yesterday if you asked.

We should have. We should have.

We could have had a cake. We could have had a cake.

What did you do for your birthday?

I had a few meetings. I went to the ymca. It was empty, which was nice. And then my husband took me to dinner.

Oh, that's nice.

YMCA underrated. You're a nice.

This New York weather, it's one of the only places blasting that air conditioning, so.

Oh, interesting. What is the gym tier list look like in New York today? You know, Equinox had a moment.

Lifetimes making moves.

Lifetimes making moves. I So I kind of do a high low thing. I live in Park Slope, so the. The YMCA makes a lot of sense for me. But I'm also a member at a club called Casa Cipriani with very small members.

Club.

Yep. It's fun.

The high low.

That.

The high low. John. John's in favor of that too. We just started going to a new gym, but John is like, actively wants to be kicked out so he can go back to Gold's Gym, which I'm like, no, we're not. We can't do this.

There aren't. There don't seem to be as many YMCAs on the west coast.

No, no, not that many. Lots of 24 hour fitness. Lots of Gold's gyms. I mean, Gold's Gym, it's the. It's the staple of Venice. It's where Arnold Schwarzenegger works out. You know, RFK's out there. It's a good cruise. True.

John. John thought Equinox was like a $75 a month gym while we were actively going there for a brief period. And I just loved how out of touch you are. You were.

It felt like premium, you know? Yeah, ultra premium. It's not the besties all in tequila level.

Yeah. The location matters a lot.

Yeah. What else is new in your world?

Yeah. Give us a. Give us a hands update.

People were assume no one's in the city right now.

Well, look, yeah, I got out here this morning. I don't want to brag. I recently inherited a car from my late grandfather to 2006 Camry.

Okay. Very nice.

Iconic era of really great vintage.

Very reliable.

Yeah. Awesome. So I drove out here today. It's crowded. It's really crowded. I'm hoping that this sort of level of crowds is due to the holiday weekend, but it might be a mistake that I rented a house out here. But we'll see what else is new since two weeks ago is the Hamptons.

Broadly, people a month ago were saying that rentals were down massively year over year. Does it feel that way?

Yeah. I don't know how much of that was because of weather or people's finances. I think some people were less motivated to make the move on booking a house because. Because we had really bad weather the past few months in New York. It was really rainy and gross and it was sort of a late start to summer. But I'm sure people's financial situation.

There was also a trade war that's still ongoing. And if your portfolio's down tremendously, you're not exactly saying yeah, I'd love to spend 100k for two weeks.

Yeah. I had a friend who was dating a Wall street hedge fund guy, and she was like, yeah, it's going so great. Like, it's amazing. And then like a couple weeks later, she's like, yeah, he's just like, really weird. Like, I don't think it's gonna work out. And I was like, showing her the stock market chart and being like, so his emotions perfectly mapped to the 10% sell off that we just saw. You're telling me that this guy, his entire personality is derived from the state of the market? The state of the market. And she was like, oh, this makes so much more sense now. Okay.

Yeah, I'm sure there's a lot of women who went through similar situations as well.

I'm sure. I'm sure the kangaroo market's not good for the dating scene. Is. Is that. Is that crazy white party going on? That's a Hamptons thing. Is that a. Is that a July 4th thing or is that canceled? I thought Michael Rubin was like, pulling back from that because it was like, maybe too flashy. Is there an update there?

Yeah, I will see. Is that supposed to be 4th of July?

I think that's like wearing. She's like, white party.

Like, what are you talking about?

Can't comment.

Can't.

No, not the Diddy one. Oh, yeah. I think that might have been a factor in pulling back from that. But, but, but I always see like everyone's like spying the watches, spying, like, you know what cars, people showed up in all this stuff.

There's a lot of paparazzi. There's a lot of tiktokers. See everything happening in real time.

Interesting.

Nobody can really like, leave with another man's wife anymore. Like, there's so much coverage of these events now. It really takes away the mystique.

The mystique. Somebody might.

What's up with. What's up this. With this place? Drugstore. You were covering it. It's trying.

Oh, yeah, this is awful. Bloomberg broke news yesterday that there's new movie smoothie pop up chain from like a Jay Z invested celebrity chef.

Okay.

And he's these. These like very Instagrammable smoothies, which obviously erewhon started, but now there's a lot of places that are kind of hopping on the bandwagon of treating like smoothies as a billboard. Like brands can collab with them. And then whoever's serving these smoothies is like making extra money from an advertising business, which is crazy. Like, I think to get an error on smoothie for your brand. It's like 200k for a month or something.

We have to do this.

Yeah, we.

You say that's.

Don't have to.

That's worth every penny for the TBP and Feed me Erewhon smoothie.

Yeah, we could do it.

We could do a collab out.

You could do a drink at like the US Open or something like. Don't do the smoothie.

Don't do the smoothie. Played out.

Yeah, it's played.

It's played out okay. Didn't Erlon have a collab with like Volkswagen or something or some really funny like car company? They did.

Yeah, you're right. They did do a car company. I think they, they also do like beauty brands like sunscreen and stuff. Like it's not.

Yeah, Chevrolet. They did a Chevrolet drink collab. And what the worst part is that I wouldn't have had a problem if they did the ZR1 like the really crazy Chevy sports car, but they did the 2025 Equinox EV.

So you're just movie.

According to the manufacturer, the collaboration between it combines Chevy's commitment to emission free vehicles and Erewhon's mission to promote sustainability as a merchant of organic foods and wellness products. The drink called the electric juice consists of Ch. Choo choo choo cho. I don't even know this. Which is the most protein rich plant source. And blue spirulina, the super food that matches the Equinox's EVs blue color. The drinks ingredients are designed to energize and recharge consumers.

I think the winner here was the ad agency that got paid half a million dollars.

I don't know, I feel like, I feel like everyone's getting the Hailey Bieber smoothie. I'm about to rock the Equinox EV smoothie. I think it has a certain badge value.

They're like 70 grams of sugar. It's like four Cokes.

I'm long sugar though. Four Cokes is a lot.

That is a lot.

But I'm going long sugar from here. I did an experiment my wife was very afraid of.

It's the worst experiment you ever. I continue.

Yeah, but anyways, people, people are very afraid of sugar. I. To prove a point, I drank a soda every single day. I drank a Coke every single day for six months straight. I know that's not.

And you work out every day and like there's like a million ways to offset that.

It wouldn't be good.

No.

A lot of, a lot of people.

Think No, I agree. It is overjoyed. People think like a single Coca Cola is terrible. Yeah, of course, of course. Everything's in the. The dose is the poison with all of this stuff.

Yeah, you drink a full gallon. What about. So. So this week on our side we've been covering the AI talent wars, how basically AI researchers are being comped almost like pro athletes and sort of like, you know, these sort of nine, nine figure massive offers. The reason I bring it up is because you obviously cover a lot of the media landscape. And we had Derek Thompson on who left Atlantic. I imagine that he's already like run rating well, well beyond. I don't know what his metrics are, but I would imagine he's already catching on substack.

He's catching up to me.

Oh really?

Yeah, go subscribe to. Subscribe to Feed Me right now and help. Help win the race. But he's not winning the race. But I was curious if you. It feels like if, if legacy media companies, I won't call anyone out, want to want real attention, not just the sort of attention that comes from their logo to some degree and the prestige, but like actually quality content. At some point they'll have to like kind of reevaluate their comp structures. I think the Atlantic was like 300k was being reported as like the kind of range. And when you can go on substack as a superstar writer and immediately be making somewhere in the seven figure range at some point or another, it just becomes really difficult.

I see traditional media companies more likely to start doing like handshakes with substack writers than starting to completely restructure how they're paying their staff.

I don't like syndication basically.

Syndication or like, I mean I've talked to so many editors from traditional magazines and papers who are like trying to figure out how they can work with Feed Me. Is it like a co publishing thing? Is it like creating a podcast together? And it's. It's been very interesting to experience firsthand to like talk to these people because I don't, I don't really need that. Like Derek doesn't need the Atlantic because the people who are reading him are reading him for him, not the whole surrounding.

Yeah, I mean, just to set the table for the folks who might be listening. When Emily comes on the show, we pay her a hundred thousand dollars for her appearance. So just, just to kind of set the bar just so.

And it ramps up.

Yeah, yeah, it obviously ramps up. But. But yes, I am interested in like the dynamic of like you can still go in public in traditional media? Is that just like freelance? What is the difference between. What do you mean?

Like, when I write for gq.

Yeah, exactly.

Are you a columnist that I'm doing that? Besides that. I adore my editor there, and I think that it's just, like, it's working, like, a different part of my brain. Like, I don't. It's like a gift to get to be edited by GQ's editors.

That's cool. Yeah. Do you. So are you a columnist over there or are you a contributor? Like, how you do is. What's the actual shape? Do you get assigned things? Like, I've had reporters reach out to me and be like, I was assigned a story on nicotine, so I need to talk to you as an expert or whatever.

So with gq, specifically, my editor Dan, when I wrote the Zen story and when I wrote a story about members clubs last year, both were. Because I had written about both of those topics in my newsletter. And he was like, I think it would be interesting for you to expand upon this in, like, a larger feature in a. In a print issue. And that's how that's happened. I don't really. But, like, I'm at the point now where if I have an idea of a great story that I want to write, like, it doesn't make sense for me to pitch it to the Times or to New York Magazine because I can get it done faster on Feed Me. I can outsource an editor if I need that. I can outsource legal services if I need that. And it's better for me to get to be, like, breaking those sorts of stories. You know, I just have, like, a good thing going with gq. It's kind of fun.

That's cool. I want to bounce an idea off you. There is this narrative that, like, when you go direct, when you're on substack, like, it has a different texture, a different. Different vibe, Different. Maybe more objective. Maybe more pro tech or pro, you know, creator or pro. Whatever the topic is being covered. I think that it's less about the personalities or the views of the people, and maybe just more about the economics that essentially, when you're in legacy media, you basically have a salary cap, and when you're outside of it, you have no salary cap. And that actually defines the economic terms shape the type of content more than anything else. Do you think that's reasonable? Do you disagree? How would you wrestle with that?

Can you. Can you keep, like, expanding on that? I want it.

So. So I think that there's, there's something where, if you, if you're working at a, at a particular outlet and there's some sort of like, salary cap. It's like, imagine if there were two NBA teams playing against each other and one had a salary cap of $300,000 per player and the other had no salary cap. Like, who would you expect to produce better basketball? Who would you expect to produce better content? When you have, when you have one that. Where a person can make 10 million or Joe Rogan can make 100 million, like, you're just going to attract the absolute top because that's where the extremely ambitious. Yeah. And they're extremely incentivized to just work extra hard. Because if they work, if they work a little bit harder and they compound a little bit more and get that extra guest and write that extra piece, it could not. It doesn't mean, oh, here you got a $5,000 Christmas bonus. It means Spotify signed you to $100 million contract. And it's the difference between, like, when LeBron James shows up at midnight to shoot more three free throws. Like, that could be the difference between like a $50 million contract and $100 million contract. So going that extra thing produces, like, better content.

Yeah. It's really interesting to see so many traditional journalists, you know, give substack a shot and like, they just plateau.

Yes.

But I'm like, I'm an animal. Like, when my eyes are open, I'm working. Like, this letter goes out every day. It's. I'm marketing it, I'm editing it, I'm pitching stories. But I've always had a really high tolerance for being told no. Like when I was getting paid 60k at a job and when I'm making like 10 times that now. Right. Like, so I think that substack and these sorts of platforms, even what you guys are doing also attract personality typ that had less boundaries and rules in the ways that they work. And then when people who don't think like that think that they can find the same success, like, they get disappointed or confused or think that it's rigged or something like that, when really they're.

Just getting out competition.

I mean, going above and beyond for a job where the cap might be. Like, what did that New York magazine story say about the Atlantic? 300k? That's what they're paying their journalists. It's like, okay, so like, you're going to go as hard as you can to make 300k and get those benefits and whatever. But the other Thing that I'll say, like, is you have to be willing to not have insurance. You have to be willing to, you know, have, have months where you're, you're not having like the support of a team. And it's just like a very specific type of personality. But I think you're totally onto something. Like, I think I'm playing a different game than my peers who are working in an office at a magazine or newspaper.

Yeah, yeah. Also that the insurance thing is real, but it's also a problem that is very quickly overcome. And you can quit your job and say, I'm going to pay out of pocket for insurance for two years and if this doesn't work out, then I'll get a rollback. I'm curious, you were commenting on Vogue partnering with Nutri Grain bars and I'm curious if you think as people are so used to influencer creator led advertising, like if you were to partner with a brand and people are like, wait, like what? Like why did she partner? That makes no sense. Like, that's clearly a cash grab that feels like it seems like legacy media brands could get away with that kind of thing for a long time. Even though we're saying, you know, why.

Is nutrigrain needs better marketing? Because the first thing I think about.

With nutrigrain shut down.

Yeah. I always think about the, I think that Nature Valley meme where it's like effing crumbs everywhere. I don't know if you've seen that one, but that's not Nutrigrain. And so my biggest association with these. They gotta work on that. Yeah, it's like it's. Is it a Fig Newton or is it the crumb one? I just know it's not good. So they gotta step it up. But yeah, the Vogue audience, I mean, do you think that there was ever a pitch where it's like, okay, we're gonna go after the moms that read Vogue and get them to buy it for their kids. And so it's some sort of like.

Bank shot strategy for me. What stopped me in my tracks is it wasn't a banner ad in a newsletter. It wasn't like, like one of those, it wasn't like somebody writing about their wellness routine sponsored by nutrigreen nutrigrain. It was a produced video of a stylist sort of like gallivanting through the streets of New York City. And I really like this girl Michelle.

Yeah.

When I saw her post it and I saw that it was between Vogue and Nutrigrain like, there was two. Who was in the room, who sold this ad and how much was it for, and how much did Michelle get and how much did Vogue get? Like, it was very confusing to me because there were. There are other ways that that could have been executed and made a lot more sense. And they probably didn't think that somebody would screenshot it and start that conversation.

But the conversation was started.

Yes.

Yeah. There's so many media reporters. If I didn't pick up on it, somebody else would have. But I think that Conde Nast overall right now is on life support. Like, we had this conversation when Vanity Farah is looking for their next editor. Unfortunately, like, it's happening immediately right after that with this new. This new Vogue editor. It's a lot of. It's a lot of pressure for, like, I don't know what, like, a few more years of running this magazine. It just doesn't really seem like how it doesn't. Vogue doesn't seem. I don't know what I would look to Vogue for right now.

Yeah. Yeah. It kind of goes back to that idea of, like, the. The independent media creator versus the legacy media creator. Like, we just released an ad for Wander that we said. They asked us. They barely even asked us to make it. And we just wrote it, shot it, directed it, like, edited it. We just did everything. We sent them the final thing, and they were like, yeah, cool. And so they had run through.

There was like a second that they asked for a tweak.

Yeah. They were like, actually, our font is this one instead of that one. And that was it. And so. And so, like, that is the type of, like, you know, entrepreneurial energy that I'm sure the particular partner that worked on, actually starred in the commercial would bring to a partnership and select and coordinate with a brand that was actually aligned. But there were just like, too many spreadsheet monkeys in the office that day. And so it just wound up being there. Yeah.

Like, I was like, why wouldn't you just hand these to the hungover girls at Barstool and, like, have them talk about, like, how this is like the bar, like, the hangover snack? Like, that I think would have moved the needle more totally. I don't know.

Are the. When I think of logical Vogue partnerships, I think of, like, luxury brands are luxury brands pulling back from that type of partnership. Like, have you seen anything there? Like, what are. What are the innovative things that you're seeing from, like, the really top tier brands out there? Have you seen anything cool?

You know what I Saw today. Do you know those David bars, like the Metallic.

Yeah, we had him on the show.

I love them. Yeah, I saw that Balenciaga at their show this week. Somebody posted a photo of like Balenciaga bars, like sort of metallic.

Okay.

Metallic label. I took a photo of it. I'll send it to you guys after this.

That's very cool.

So I mean, like, is that going to move their, their hoodies? I don't know. But like that's. Those sort of fun, weird moments are more exciting to me if we're talking about nutrition bars.

Yeah, I mean it certainly shows that like Balenciaga is still on the, you know, hip or just aware because like Peter Rahal, he's kind of, you know, known in consumer packaged goods world and a little bit in tech. And some people know the story of RX bar. And then David Barr with the revenue ramp. Like the business is doing well, but it's still like a complete insider industry story. Like the average person on the street know that. So Balenciaga is kind of signaling a little bit to that.

Yeah.

Hey, what's cool with the new business community?

And I'm sure that that took them like a couple thousand bucks like put together. Like, you know, it's not difficult as far as what, what luxury brands are doing. I think it's a lot of events now. Like they're putting a lot of money into events. Like I was just at a Chanel event. They do a lot of programming with Tribeca Film Festival, which is cool because everybody's in the city anyway, so they, they throw events and make sure that people show up to that. I think that a lot of my social calendar in New York right now is dominated by brand events. It just seems to be what's moving the needle because people go and they get photos of themselves and they post it and it's like the brand is the background. So you end up being like, New York just feels like a billboard every, every time you go out now because everything is a brand event. It's like dominating how people are organizing their social calendars.

I don't go to events, I don't go outside. I just stream what makes for a good brand event. Like, is this just a happy hour? Is that what we're talking about? Is there dance?

It's actually interesting that the thing. I'll say we were talking about svb. Oh, the club earlier they would throw dinners that like there's an entire founder dinner industrial complex that are just like financial or you know, companies that throw these dinners the Funny thing is that within tech, the company has historically failed to make them social moments. So it'd be like, it'll say the name, it'll say SVB on the menu. But people aren't like taking a picture of the menu as much as they're just like taking pictures. So I think what the luxury brands do well is you're saying, hey, we're gonna invite 200 people into this photographer and you're not gonna be able. If you, you post any photos at all. We're going to be getting exposure.

We're going to have a moment. Yeah, I threw a brand dinner or a feed me dinner with my friend Paul from the Infatuation last year and Andrew Ross Sorkin was sitting next to me and I gave him a feed me keychain. That was my brand moment.

That's a great moment.

So that's another thing. Like those dinners are awesome.

Yeah. How are brands being creative? Are we seeing like ice sculptures or step and repeats? Like, like how can brands stand out with one of those?

Like the worst possible brand event is a DJ and free drinks. Like just people standing in front of like a 22 year old DJ with like free drinks. And then more innovative things are like, I think when you give people an activity to do, like if you are doing like bowling or and making it like glamorous or like, you know, now brands are, if they're doing a collab with like an athletic brand, they'll, they'll take a bunch of women to like tennis courts or a golf club in Connecticut and like that's a photo op for everybody involved. So like these mini expeditions or like J. Crew last year took a bunch of people on a gorgeous sailboat in the Hudson river off of like the seaport.

It's basically like take your customers on a date date.

Right.

On an extravagant, give them adventurous date photo.

Yeah, yeah, exactly. And then if you level up then you're, you're gonna start talking about like brand trips. Like Hermes took people to Aspen last summer and yeah, I mean a ton.

Of VC firms do that all the time. Could come with us and we'll do a track day or movie night or we'll go out to, you know, the. Some, you know, beautiful place, some ranch and like shoot, shoot shotguns and have dinners and it's a whole weekend.

Where's my call? I want to do that.

They should get you on the calls. We'll introduce you some folks.

Post mortem on the Democratic primary, but just on the marketing, right. There was a lot of coverage of Zoran's basically film vertical video making abilities that seems to have played a key role in the campaign.

I'm curious, are you talking to any, like, tiktokers or content creators that are like, I want to build a business around this for politicians. That would be interesting.

Yeah. I had a friend who I used, actually used to work with at Conde Nast who worked in the White House for Biden, and I think that she helped out with Zoran's campaign. Like, there are people who know, I think, like, you have to be excellent at that job if you're working for a political camp political campaign, especially in the White House, because the turnaround times are insane and the levels of approvals that they need to go through are crazy, as they should be. So I think there's like a mini network. I mean, there are a lot of people that worked in the White House under Biden who used to work at Instagram. Like, they were recruited straight from Meta. So that's definitely a thing. But I think that organizing is a tremendous amount of work. And like, part of that is digitally now. Like, and, and Zorin did that really well.

Yeah. Yeah. It's interesting to think about how the presidential election felt like the podcast election, and then this primary felt like the short form election. And that sort of. It makes sense, especially as you're trying to just appeal to younger and younger voters.

I mean, his team invited me to start covering his campaign back in February. Like, I was, I was at a party, I think, in March for him. But he's. He also went on, like, throwing fits and a few other podcasts. And he did like a Q and A with me for Feed Me. So he, his team knew where to, where to act.

That's good. We've been talking a ton about the AI talent wars. There's a lot of people in Silicon Valley making $100 million for in signing bonuses. Silicon Valley, notoriously bad at spending money. New York famously good at spending money. What are some recommendations that you can give to these PhD researchers that have $100 million burning a hole in their pocket? They want to get in the game. They want to step up their fashion, their cars, their houses. What would you recommend?

It's funny that you go to fashion and cars because I think there's a real, a Louis Mangioni situation on our hands here with this kind of job. So I feel like you need a security guard.

Okay, sure.

Yep.

Plain clothes security guard, which means, like those tight Lacoste polos, you know?

Okay, maybe they should be carrying too. Maybe they should. You know, if you got the. You got to be everyday carry now.

If you're making 100 California, it's gotten more lenient.

Yeah.

Yeah. So I feel like you have to staff up in a position like this. Like you want a driver who's trustworthy and isn't going to like leak your information. A personal trainer who's trustworthy. A feed me subscription, of course. And then I think it's like removing friction from areas of your life with like staffing up like that. I recently learned that this neighborhood in the Hamptons called Sagaponic is a microclimate. And I feel like that's a good place to invest your money.

Wait, break that down a little bit better, A little bit more. It's just, it's just on an average day, it's nicer there.

I think it's necessarily nicer. I think it might be like wetter and more humid than other areas. Like, I am in a different neighborhood than that. And if you were there right now, it would feel different than here, even though it's not that far away.

Yeah, like. Like you can be like Long island as a whole is hot today. But Sagapona could, could have a different weather report because it's this micro subclimate.

Because of the dunes and the tides and the trees.

There's different parts of LA that have the same kind of dynamic.

Yeah. I mean, isn't San Francisco like that too?

Yes, famously. That's why in the summer it'll still be. It'll still be foggy and stuff, but then it's really warm later in the year.

Right? Yeah. So I feel like investing in a micro climate is a good way to spend your money. That just sounds.

Yeah. They're not making any more of them.

And then I, I think that you need to use your money to start taking some, some experts to dinner because you can't just start buying watches and clothes. There's a really high chance that you'll miss and go for like trendy items. So I would try to hang out with like. Do you guys know the watch dealer Mike Nouveau? Like somebody like him or, you know, like, take him out to him and, and get some advice. I bought a watch from him last year and he seems like he knows his stuff.

What'd you pick up?

I got my husband a nice watch.

Okay, let's go. Let's hear it for the husband.

Congratulations. Two size gongs, one interview. That's how you know it's a good one.

A birthday and a gift and maybe like befriend Chris Black or like another stylish guy and get him to advise, consult on the closet.

Can you get us Chris Black on the show? Because the only podcast merch I've ever bought was their merch that they released where it was a cease and desist from the New York Times because they, like, copied the COVID art and they just printed, made a shirt, sold it.

That's great.

It was a work of art.

Chris would definitely. Come on. I'll connect you guys. That'd be awesome. Yeah. I think that the biggest mistake that people who come into a lot of money quickly make is just, you know, doing that automatic Apple pay on while they're shopping and then you end up with stuff that doesn't fit you or looks horrible or something.

Yeah. So stylists, tech stylists, or paid friends.

Take people out to dinner, host parties. Like, keep them around, you know.

Do you think it's good to surround yourself with yesmin?

What do you. What's. How do you.

So if you surround yourself with yes men, then whatever ideas you have, all your yes men will just tell you, yes, that's the best idea ever. A lot of people say, don't do it, but maybe I would not do that. Everyone around me has been saying that it's a good idea.

No, I would not. That's not good.

Okay, so stay away from the yes men.

Yeah.

You want yes people on your team where you say, I want to do the impossible, and they just say, yes, we're gonna do it. But there's. There's a line, you know, you don't. You don't want to cross it.

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

Thanks, guys.

Send us the invoice for the 100k appearance fee.

And, yeah, send the next few. Yeah. Happy fourth of July. Fun. We'll talk soon.

Bye. Up next, we have Merchant from Helios here to talk about stablecoins, talk about crypto, talk about NFTs.

We're doing it all this show.

Doing it all this show. Very excited to talk to Mert. He put on an absolute clinic during the last show. He might be having some technical difficulties. So let's go back to the timeline. Daniel growing. Daniel says OpenAI is super nice for giving everyone the week off for interviews at Meta and Elon Musk. Musk is crying, laughing. Absolutely brutal. You saw the wander ad, but you can still find your happy place book a wander with inspiring views, hotel, great amenities, dreamy beds, top tier cleaning, 247 concierge service. It's a vacation home, but better. Vittorio says, I used to pray for the day nerds would be treated like athletes. The day is finally here and it's glorious. And it's our post put up by Tyler Cosgrove, the intern. Tyler got 25,000 likes. Vittorio quote tweets it get 50k. You love to see it. Anyway, we have Mert in the studio. Welcome to the stream. How are you doing?

I am having some technical difficulties.

Can you hear me? Yeah, yeah, you look good. The only thing is maybe rotate the camera if you're on your phone so we can go horizontal, but other than that, looks fantastic, looking good. Thanks so much for taking the time. I know you're really a jet setter at heart all over the world, but very excited to talk to you. There's a ton of stuff going on in the crypto world that we wanted to talk about. Should we start with the idea of private companies on chain? Is that a good place to start?

Yeah, sure. So there's a few versions of this that have just been announced. One by Republic and one by Robinhood. Yeah, do you want to just go over both or.

Yeah, yeah, let's start there. And then I basically want to cover public and privates, but we can start with privates.

Sure, yeah. I mean, so the details on these are a little fuzzy just because I think at least the name of the game for some of them seems to be regulatory arbitrage. But so for example, in the case of Robinhood, they just announced, announced that they're going to have it and they didn't really give any specific details other than it would be on their own optimized chain, blockchain for RWAs, for Republic. They basically said, and by the way, people thought this was for trading these private equity like these, these private shares, but they're actually just for buying them and then holding them. And then you actually get, in Republic's case, you're locked up for about a year and I believe the maximum size you can do is actually only $5,000. So I think they're basically just trying to test the waters and see how the regulators will react is my reading of it.

Yeah, the 5k limit is interesting. The lack of immediate liquidity is interesting too. I mean, it's clearly a product for retail. The other thing that I was thinking about is if you just release SpaceX shares on chain, there's almost going to be infinite demand for them. And so the limit could also just be, hey, we want as many people to be able to access These as possible or it just gets bid up and bid up, but it sounds like it'll be fixed price. Any reaction to Robinhood setting up their own chain to do this? And then it's also Europe only from what I could tell.

Right, yeah. And by the way, just on the Republic private shares thing, it's actually I looked into this, it's a mirrored asset, so it's purely mimicking the performance of the price of the shares and then it's a note on the upside in the event of a liquidity event. So it's not like you're actually owning the stock by any means, obviously for Robinhood. Yeah, I mean they. So it seems like they decided to launch on arbitrum the L2 to start with, with the eventual understanding that they will launch it on their own optimized chain. I think Robinhood has always been like a vertically integrated company that wants full control over the design, the user experience, the abstractions. And so I think building your own chain for this makes sense because I mean they obviously already have the liquidity, the users and the distribution. So I think it makes sense provided that they can hire Vlad, can go full zuck mode and hire some crypto talent to run the blockchain, because that's hard work.

Run the blockchain, Break down the. What's happened? I've seen enough screenshots this week so far. I think you were sharing some of the them of people swapping something, an esteemed asset like Fartcoin for some Tesla shares. What is actually happening there? Is it just are the X Tesla or whatever these are, are they tied to any real underlying asset? Is it entirely Vibes what's actually going on?

Yes. So as sophisticated investors, we've been swapping far coin into Tesla and Nvidia stock and even Meta now. And so these shares actually are backed one by fully backed by the custodian. It's quite fascinating how it works actually. And these are only available in Europe by the way, both in the case of Robinhood and Kraken and this company called Bakkt that's actually doing the regulatory stuff. So what happens is there's some KYC at the beginning of the process where you actually have to buy from the primary market, but then once it's withdrawn to the blockchain it's basically there's like this loophole where they're triaged certificates. And so you can as long as you say that these aren't supposed to be allowed for USA investors, as long as you basically just say that and you put the fine print, there's. Once you do that, they're just treated like digital assets. And so when you do trade Shitcoin 3000 into Nvidia stock, it actually is like you're getting something that's backed one to one. However, it's the same with the private stocks where there's a pretty big limitation on the liquidity, I think. You know, 100 million is like the max that can be issued at this point and they're doing like several tens of millions in volume. So nothing crazy yet.

But how are they trading? They're not stable, right?

No, no. Yeah, so there's a.

So they can trade at a bit premium.

Prices are. Yeah, no, they're not. This isn't. It's not like a microstrategy play. We can talk about the treasury companies as well, but the prices are supposed to be indexed one to one. And what's interesting is like you can now obviously compose this with other applications in defi. Right? So you can literally buy Tesla stock and then take out a loan against it and then maybe a fart coin at that point. Like you can do all sorts of different things and you can just like hold it in your wallet. It's pretty cool. But like, it's just, it's very, very early. So I wouldn't make any.

But how, but how it feels pretty monumental, right? Is this not what. What people have been excited about for the. For the last decade?

Two decades? Peter Thiel was saying that he wanted a cryptocurrency currency that was backed by a basket of every financial asset. And like we're on the cusp of it. It feels like that you could have like a token that owns a slice of the S and P, a slice of physical gold, a slice of real estate, and it'd be kind of like the most stable asset, even more stable than the dollar, potentially. That was like the original thesis, but.

Yeah, that was actually, by the way, that video. Yeah, yeah, you actually surfaced that video. That video is crazy. Predicted in 1999. But yeah, I mean it is huge, right? It's gotten a lot of attention and it's very exciting that you can just basically start to actually modernize the financial rails. Patrick Collison from Stripe had a really good way of describing stablecoins which was like this is like dollars on superconductor rails, right? Where there's now no friction. And you can just composed with any other financial instruments, obviously that's on the blockchain and it's just, it's got. It's taken so much work to actually get here and clear all these regulatory hurdles. But you know, it's, it's really just beginning. I think there's going to be. Once Robinhood gets in and everybody else starts getting. I think Coinbase will also start pretty soon. Right. They also have their own blockchain. So I think it's going to start getting very competitive.

Okay. I was talking to a retail investor who would be like so excited, like.

150 IQ.

Would be so excited to be able to buy private company shares on chain, be able to buy all sorts of stuff on chain. But he was particularly very upset and still held a grudge from the GameStop rally where there was no buy button, there was only a sell button and he felt like betrayed by the platform. Now when I dug into it, it seemed like the situation with Robinhood was much more nuanced than that. And there was actually some liquidity constraints and there were a lot of reasons why they chose the whole narrative of like, oh, they, Ken Griffin called them and told them to save some hedge fund. I didn't really buy into that. But anyway, I think the more important question is when you put something on chain, the idea is like, like the, the code is decentralized. No one has the button to turn something off. But the, the downside of that is that you could maybe get into some situation where the, like if there's, if there's a, you know, reinforcing feedback loop of sell pressure or buy pressure, you could really, really distort the market. And so I'm wondering, like, you know, I'm thinking about the Terra Luna situation or any of these situations where there was like this perfect perpetual motion machine. But then the fact that it was actually that someone couldn't step in and just put pump the brakes was actually a bug, not a feature at that moment in time. And I'm wondering if there are potential risks to having real world assets on chain that where everything is in code. Does that make any sense? Am I asking any sort of reasonable question here? Can you just on that a little bit?

Yeah, it's certainly a very valid point. And it's also why the genius act for stablecoins was pretty monumental because like, you know, Terra Luna also had some very funky concepts around their stable coins and they're not actually being backed by anything. Right?

Yeah.

And with the new stablecoin act, which like puts very or you know, much more clear framework on like how often should these things be audited? How should they be backed? Who can do this whatever. And with this new regulatory approach, consumers and users actually have way better benefits and way better guarantees. And it's also, it's obviously decentralized finance, but it's also programmable finance. Right. So in the case of an L2, for example, Robinhood can add again the same types of rules to enact some sort of control and maybe free something. USCC can already do this today by the way. Right. USDC can actually freeze the USCC on a per wallet basis. And even on Solana you can do this. The smart contracts have certain hooks depending on how the issuer of the token actually encodes these. So you can actually still program these to get like the compliance flavor that you want. The difference is that it's extremely transparent. Right. You can read the code and you can be like, okay, this is what can happen in this situation. And this is what they did. So you can actually audit it. And it's very verifiable.

I have a question about Erebor. This story kind of leaked out today in the Financial Times about Palmer Luckey, Peter Thiel and some other folks. Joe Lonsdale, working on something that sounds like a stablecoin bank. I would love to hear about the problems that crypto companies run into with banks. What might they solve if they, if you were imagining the list of to dos and then are there any. I mean there's been a few crypto friendly banks, but if I have it correct, like Robinhood isn't a bank, Coinbase isn't a bank, Circle's not a bank, Stripe isn't a bank. Or, or maybe you can correct me there, but just kind of take me through the problem that banks solve and then in the current structure and then what problems do stablecoin companies still face with the traditional banking system?

Right, yeah. So I think actually Mike from Pirate Wires and Nick Carter did a really good job on Operation Chokepoint where they kind of go in detail into the debanking of these industries. I have like a much more trivial example which is my co founder, he's Canadian, unfortunately, just like me. And he went to get a mortgage for his new house. And at first I think they approved him and then they looked at where he worked, which is Helios, the company that we run. And then they actually rescinded the mortgage and they said it was you're in crypto. Even though we don't have like a token, we're basically a cloud provider. And they said it's too high risk of an industry. And they actually, he couldn't Get a.

Mortgage.

And like this is Canada in terms of problems. There's a lot of. Maybe there's two dimensions to this. One is just the interoperability. Right. Like being able to off ramp, on ramp, compose. Like if you're earning money USDC on chain, how exactly are you transferring that? How are you doing payroll, all sorts of things, like just the tech side of it. And then just like being able to get access has been historically cult especially. I think this is changing, especially with the president of the free world and actually the first lady both being the founders of various shitcoins. I think it's certainly changing.

But yeah, we saw that. What was the announcement?

Are you talking about the social media founder? The founder of Truth Social. Because he has a coin, right.

Unicorn.

People don't give him enough credit. It's very rare that a founder of a social media company as big as Truth Social has a side hustle as demanding as being the President of the United States. Yeah, it's pretty remarkable.

Yeah.

There was also the news that I'm forgetting the governing body, but they were advising Fannie Mae and Freddie Mac to start looking at crypto assets in the mortgage process.

Federal housing.

Yeah. Yeah. So there's progress, right?

Yeah. So that one's super interesting where you can actually, I believe the exact rule is something like, like the assets that you have on a centralized exchange, like a Coinbase and soon to be, I guess, Robinhood, you can actually take out a mortgage using those assets as collateral. And I believe they actually. I don't know if this was real news, I didn't look too into it, but I believe they were actually looking at like if far coin is viable to use as mortgage collateral because like it obviously needs to have some certain properties. But yeah, that can be very interesting. I'm not sure how it'll go. You know, the devil's always in the details for these things, but there's a lot of progress. Right. Like two years ago with Gary Gensler at, let's say, being the Eye of Sauron, you could not even imagine, like, you could not even think this. You would. He would arrest you in your sleep.

Sleep.

And then everybody left the states as a result. And now they're coming back.

Yeah. Well, the Eye of Sauron is closed and the mountain of Erebor is born. So we'll have to have you back on soon to talk more. This is fantastic. Thanks so much for calling in. I'm sure it's late wherever you are and we really appreciate you taking the time.

Yeah, great to Catch up.

We'll talk to you soon. Have a good one.

So many zingers.

The crypto vernacular is just undefeated. But you heard Emily Sundberg tell it, mention it earlier. She bought her husband a watch. You can buy your husband or wife a watch on bezel. Go to getbezel.com, your bezel concierge is available now to source you any watch on the planet. Seriously, any watch. And our next guest has an absolute hitter. I saw him in D.C. wearing a JLC reverse out there. You go to dox your watch, but it's a fantastic dodge.

Great, great. Now he's got a higher security.

And by the way, look at this. Oh, he's got it on. Let's go.

The Uno Reverso.

Beautiful. It's a beautiful watch. It's fantastic. It has a great, beautiful watch.

It was actually. It was a gift from my wife for my 30th birthday. So. Beautiful watch my family for my 30th birthday. So I'm thrilled to have it.

Yeah. And if I remember the story correctly, it was different time zones for a while.

That's right.

That's right. No better watch than the Reverso to represent where you are physically and where your future wife is. Fantastic.

That's right. JLC sponsor the pod. We're ready.

Really? We are so ready.

So ready. Until then, getbezel.com was lovely having you on yesterday. A lot tends to happen in 24 hours these days. What? What's new?

Give us the update.

Yeah, It's a big 24 hours. First of all, has anyone ever done the two day in a row?

Delian did once, kind of by accident. He came on, he was scheduled every other Thursday, and he happened to stop by our YC demo day stream for a very different thing. We hit him with the confetti cannon. It was a lot of fun.

That's all right. All right. We love Delian. That's fine. Big 24 hours in Washington. In fact, it started right after I hung up with you guys. I don't know if you saw this in the background of my shot yesterday. You had this beautiful vista over downtown D.C. and as we were talking, there was a huge storm cloud starting to come in and completely raining on top of all of D.C. so you guys know this? I think people probably know in Congress there is no remote voting. You are either voting in person or if you are not there, you do not get a say in what happens. And right now, the big beautiful bill, which is in the hands of the House is on the precipice. It could go either way. It might pass, it may not pass. But every single vote is necessary. And I kid you not, there was a ground stop in the morning at dca and there were multiple members who were stopped, prevented from coming back into town to vote because of the huge storm in Washington last night. So just to give you a sense.

Of how quickly things move, it's a biblical metaphor, for sure. A flood. It is.

It is biblical. I don't always see divine providence in politics. In fact, I just about never see it. But it's pretty hard to argue with a full grandson.

Not in these scenes. So. So does this make JD Vance's vote more important? Less important? What. What's the read on Vance's role here?

So Vance is the president of the Senate. When there's a 50, 50 tie in the Senate, which is what happened yesterday, his vote is the most crucial vote. And that's why, by the way, if you are any Dem, Governor Shapiro, AOC, whomever, running in 2028, it is now a phenomenal gift to you that you can go and badger dance about his vote yesterday. But the bill is in the House today. So there's no Vance vote. There's no tiebreaker. There is no. In fact, the GOP basically has eight votes. Sorry, four votes they could lose. There were eight people who didn't show up this morning. There are four votes they could lose and still get the vote passed. And so all day, there's been this series starting in the morning, of procedural votes as a way of checking just who from the conference is even in town to begin with before they get to the actual vote itself. This is like, literally like a roll.

It's a roll call.

It is literally a roll call. My mom was a teacher for a while. My grandmother was a teacher her whole career. There is zero difference between my grandmother teaching a special ed class, doing Roll Call and what's happening today in Congress. Maybe a slight difference, but basically no difference. All right, so there's a couple of factions you guys talked about yesterday, the menu of different people within the gop. The big challenge today is picking up votes where the GOP can. Where the Republican Party can. In the menu of factions. The two big ones to keep an eye on are the House Republican moderates and the House Freedom Caucus, which are the sort of more extreme, further ideologically opposed. Folks, if you're a moderate, you don't like it for one simple reason, which is this has a bunch of Medicaid cuts. In fact, there are far more Medicaid cuts in the Senate version of the bill, which is what you're being asked to vote on in the House version of the bill that you passed a couple weeks ago. And you don't want to hang that around your neck when you go back home. Remember, if you're in the House, you are always in campaign mode. You're running every two years. That's why you were out of town to begin with. Because anytime there's a weekend, you're flying home to your district to go pick up a couple extra dollars. Right. So the moderates don't like it because of the Medicaid cuts. Huge Medicaid cuts. The conservatives don't like it. The far conservatives and the Freedom Caucus don't like it because it actually blows up a framework, a balanced budget or close to balanced budget framework that they passed in the House last time. So they got to a basically a compromise agreement with spending cuts and with tax cuts that the Senate completely ignored. So that's much harder for them. And part of Trump's success if this bill gets passed will be because he convinces people from the Freedom Caucus in particular to flip and become supporters of the big, beautiful bill.

So wrapping up on J.D. vance.

Yeah.

If the vote goes through in the House, then people will hang on him. What happens if the bill passes? Does the other bill automatically die? Is there some sort of trigger there?

So if the House passes the bill today, the House is looking at the Senate's version of the bill. Yeah, The House passes that bill today. It goes right to the president's desk for signature. That is the plan A scenario. If you're in the admin, what you're hoping happens is it gets passed, it goes through plan B or plan C even is the House says we want to reopen the bill and add a bunch of different amendments because we don't like the Medicaid cuts or we don't like the deficit expansion or whatever. And then it goes back to the Senate where, by the way, the Republican Party may or may not have the votes to get it done. And suddenly JD Vance gets into that spotlight again.

Got it. Okay. So the. So, yeah, so the House can't change anything in the Senate bill right now. They just have to decide yay or nay.

That's exactly right. If they reopen, if they make any changes, which, by the way, a lot of people in the House want them to do, then the Senate has to go through the entire procedure of passing it again, which is a huge risk factor for them.

Yep.

Is there any chance that The House tries to add back in that, that AI state level AI legislation, ban, or is that even.

I think it's. I think it's fairly unlikely. Right. Given the overwhelming. I mean, listen, given it was a 99 to 1 repudiation of a rebuke, rebuke of the AI moratorium, I think it's unlikely the House added back in because that's a signal from the Senate saying, look, if you put this back into the bill, all that's going to happen is we're going to lose even more procedural time getting it out of the bill again when it comes back to us the next number of days. But the only way this passes before the July 4 deadline, in my view, is if the House basically swallows the Senate version as is today.

Yeah. So you kind of define some of the, the attack lines on the right with the moderates going after the Medicaid cuts and the, and the farther right Freedom Caucus. Is that right, saying yeah, that they want more of a balanced bill. Have any particular attack lines been opened on the left that you've seen?

Oh, yeah, yeah. I mean, the left looks at this. This is Christmas. If you're on the left, right, you get basically to say, wow, the Republicans voted to take away your Medicaid, which is a huge bludgeon. You guys remember back in the OGT Party, all those signs, keep your government hands off my Medicare, Keep your government hands off my Medicaid. Right. This is a horseshoe issue where people on the left and on the right are equally impacted by it. And the net result of that is that it's a great campaign line no matter what side of the aisle you're on. So here's an example. North Carolina. We talked yesterday about Senator Tillis from North Carolina, who's not running for reelection and who ultimately voted against the big beautiful bill passage. Tillis has a group of folks in the House who are predominantly from his state delegation, who he has some influence, some relationship with. Right. Many of whom have signaled more. Many of them are moderates and many of whom have signaled I have some reluctance to vote for the bill. Part of the reluctance is because the Democratic governor of North Carolina wrote a letter where he basically said, look, if you vote for the bill, I am going to make sure that these Medicaid cuts become a big issue in the next campaign because they massively impact his state budget. Josh Shapiro in Pennsylvania did the same thing where he said, look, if you vote for the bill, you should know that you will be impacting the Medicaid of Pennsylvanians And I will use that politically against you. The last part is implied.

This has been really, really helpful. I've genuinely learned a lot. So thank you for stopping by. This is great.

Well, you may not have been the first person to go back to back, but I have a feeling you'll be.

The first to go back to back to back to back. This is very, very helpful. Thank you for breaking it down for us.

Thanks for having me on. I appreciate it, as always. We'll have you back on my Chiron. Yeah, exactly. I can't see the Chiron, but I'm hoping it says Washington correspondent.

Oh, yeah, we will definitely update it tomorrow. Thank you so much.

Vice President of Washington.

Yeah, great. I'll see you guys tomorrow, I'm sure.

Cheers. Cheers, Zach.

Cheers.

Thank you.

Up next, we have our lightning round, beginning with John from Paraform. Welcome to the stream, John. How are you doing? Good to meet you.

What's happening?

Good, guys, thanks for having me.

What's the latest? Can you kick us off with an introduction on yourself? Jordy's getting the hammer ready. What's the latest news? Do you have any good numbers for us? We love big numbers, hopefully with lots of zeros. Yeah.

Hey, everyone.

Yeah.

Great to meet you guys. I'm one of the founders and CEO of a company called Paraform. We just announced our 20 million Series A LED by Felicis and an incredible group of founders.

Congratulations. I like when Jordy does it because he goes in the camera feed. It's great.

Congratulations.

Congratulations. 20 million bucks. That's amazing. Tell us what the company does, the inciting concept, the idea, how big you are, the scale of the business right now, that type of stuff.

Let me guess, you're going to hire 1/5 of an AI researcher. Use of funds.

Yeah, yeah. We're getting one day a week.

Yeah. Fractional.

Yeah, fractional. Fractional AI researcher.

Oh, I saw that. And also like the founding engineer who, you know, for four different companies, I think his name was Soham.

Oh, yeah, yeah.

No, yeah. But for those of you don't know, Paraform is a marketplace where we help the world's most important companies hire top talent by connecting them with expert recruiters who actually act more like sports agents than talent sourcers. And, you know, I've actually seen your. Your post recently about the meta superintelligence team. Like, you guys framed it as, like, traded.

Yep.

Internally, we've been talking about this for over a year, and it's really cool to see that sort of trend sort of playing out live. But also it's happening a lot sooner than we thought. So really interesting to like dive deeper into like the trend of like just recruiting in general.

Yeah, yeah, yeah. I mean we've been, we've been talking about this a lot because it's good content, it's interesting. And, and I think the surprising thing that we learned this week, I think is that we put up that traded graphic for the meta AI researcher and it got millions of views, like millions of views. And it's funny and there's an element that can go broader, but it really does speak to the fact that tech has gone mainstream. The question that we've been mulling is where does this go next? Because when you look at the goat debate between LeBron James and Michael Jordan, you're looking at points per game, you're looking at lifetime career points, number of MVPs won, number of championships won. What metrics are you interested in Looking at? Is LinkedIn still relevant? Is GitHub still relevant? Are you quantifying these things? What are you doing to try and match these folks together? You imagine the, that AI can be very useful. There's a whole bunch of different ways that I could imagine this working, but you probably have to do some filtering on both sides to understand who's a good recruiter and who's a good, who's good talent for sure.

Yeah, I think like, you know, the, the trend that you guys picked up, it's like you're onto something and it's like very similar to what we think. It's like I don't think that important tech company hiring, let's say the best marketing person will be too dissimilar to the Lakers announcing that, that LeBron joined or the Dodgers getting Ohtani. I think how we think about it is as AI sort of rapidly advances and it automates a lot of the mundane work, I think humans sort of focus on more like higher order functions. Right. And one person will be able to do a lot more. Obviously the GDP will accelerate, so the average salaries will go up. We've always internally talked about how the age of like a seven figure salary is rapidly approaching. It already is for certain industries, but I think think it will be a lot more mainstream. So yeah, I think like there's going to be more scarcity towards good talent. It's going to get even more competitive. And when there's that scarcity, usually the value of a broker goes up and we actually see recruiters kind of turning into almost one of the most important roles in tech, where essentially they're like brokering this human capital between the most important companies and making sure they represent the best talent. In terms of metrics, I think very similar to maybe like the MBA or the mlb. Like, like obviously they'll measure the company's performance. I think a lot of people commented on how like it's ridiculous to pay $100 million or whatever. But really these companies, they're smart. Like they're doing that because the return on that is greater. Right. Like, I mean solving AGI I think is like a important enough problem where you do that. Right. So I think, you know, that's one thing that you're going to measure.

Right.

Like how well are companies going to perform and achieve their objectives? I think there's other like key KPIs, like, you know, I'd imagine it's just like the North Star metrics for different teams. I wonder if there's going to be universal things like points, assists and you know, minutes played. But yeah, I think it's something interesting to think about.

Yeah, Yeah, I mean we kind of already do this at a little at like a smaller level. When someone says like I was early at stripe, what that really means is that like you were there for the creation of like the first 50 billion of market cap or something. And so if you were one of the of 20 on that team, yeah, we might give you $50 million in points. Right. Essentially. So yeah. What else is going on with the top of Funnel? Where are candidates percolating up into the Paraform system these days?

Yeah, so one of the fastest growing roles we see are forward deployed engineers. Like obviously Palantir, they're the one of, of our customers too. They coined this term and really do a good job.

Shyam Sankar, the og, the first to ever do it. The first forward deployed engineer. He's been on the show.

Exactly.

Yeah. I think obviously you guys know this too, but I think the biggest bottleneck in AI is quickly becoming implementation.

Right.

Where we already have very good LLMs, we have already very good technology. It's just a matter of getting them into the hands of everyday people and the broader industry. And I think go to market motions are evolving to actually fit this forward deployed engineer role where basically you have to customize your AI solutions to everyday customers. So I think forward deployed engineers is where a lot of really good talent will converge. And at Paraform what we're seeing is not only companies like Palantir, but also companies, traditional companies like that. They're also like looking to hire for deployed engineers. So that's one trend we see emerging very quickly.

Yeah. Any commentary? You mentioned Soham Parak, the Indian engineer who works at three to four startups at the same time. He Suhail says he's been preying on YC companies. Multiple people are hitting the timeline now saying that they fired this guy. That feels like something where we need almost like an inverse first LinkedIn because clearly he can pull jobs off of his LinkedIn and say he just worked at one place at a time. But do we need some sort of reputation score? Could that be abused? How does that evolve?

Yeah, and I think it's also interesting to think about like the kind of remote work versus in person. Like this obviously wouldn't have happened if it was in person and you have to like validate things. But I think there's a really a lot of interesting things there. For example, you know I think with AI like I think AI can be incredibly helpful in the recruiting process making it more efficient. But there's also going to be a lot more spam. There's a of lot recruiting is always driven by this power law where unfortunately there's only a small amount of very good people and a large amount of it's a long tail. Right. And I think with AI I think that trend will even accelerate. So I actually think that having a human in the loop to vet these things out and not only just these things but think about culture, fit, interpersonal dynamics, do I like you? These kind of things are actually very much what recruiters are effective at. So I think paraform can help basically prevent these situations.

Yeah. One question just getting into, I'm sure this question came up in various pitch meetings and throughout building the whole company back in the early Internet days, the dot com era, every CEO said hey, we're going to need way less people. And people are saying that again right now. How would you answer that question? Your obviously building a talent platform is like, you know, and I'm sure people, you know, might say well what happens in a year if the hiring market gets gets wrecked? How is that going to impact your, your business? But I'm curious how you, you know, see that market evolving.

Yeah, I think fundament fundamentally in terms of value creation if you think about it from that perspective, like I remember seeing a story where when cursor wind surf these tools became so Mainstream, these Fortune 500s, they were asked hey are you going to hire less engineers because can one person do more? And they actually said no because of these tools we're going to hire even more engineers who can Use this to be even more effective. I think at the end of the day, there is a ton of value to be created. There's a lot more innovation to continue to happen. So I think that actually we're always going to need people, it's just that each person will be more effective. So I think we actually have a very optimistic view when it comes to the human and AI collaboration. And I think another thing is, yeah, like we talked about the sports agent analogy, I think each talent will become more and more valuable. So maybe the frequency of hire may decrease, but the value of each hire increases. And when it comes to sort of like a success based model like Paraform, I think the value we capture is sort of actually increasing even more with this trend. So, yeah, I guess not too worried there.

Yeah, talk to, yeah, talk to us about where the value is occurring and where, where it's landed historically and where it's going forward. Like when I've hired engineers through recruiting agencies, typically the company pays like one month or two months of salary. So it's like a percentage. And then there's some sort of clawback if they don't make it after three, three months. Sarah Guo was talking about how some folks are helping AI researchers negotiate and they're taking a cut of the pay. And it seems like they're taking it from the researcher. Because if I can get you 5 more, 5% more, that could be millions of dollars. You're fine paying me a million bucks for that negotiation. Then there's platforms. So do you see that economic equation shifting from the company paying to the employee paying to platform getting comped? How is that evolving?

Yeah, yeah, I mean, I think that's. Ergo, example is also like, exactly like what, that's what sports agents and, and even in the music industry, I think that's what they kind of do. But yeah, to answer your question, I think where we are creating value is actually while keeping quality consistent and maybe even better, we're actually reducing the time to hire significantly. So every customer cares. The number one thing they care about is the best candidates fast. Right. So the metric we look at is time to hire. And Paraform is making revenue by filling the most difficult roles in the world. And our average time to hire is under one month. And I think when it comes to efficiency, I think what, what happens is at every step of the procurement process or the recruiting funnel, we're making it really, really efficient. I think we're basically decreasing downtime. For example, if you want to work with a recruiting agency, you're relying on word of mouth. You never know whether they're good or not before you work with them. And there's a ton of these little things that waste your time and obviously cost. But at Paraform, based on what our customers are looking for, we're able to match them with the best possible recruiters based on their track record, their candidate network. We have so much data on them that we're able to do like an instant good match of like best recruiter for the best role. So that already saves like weeks. And most customers find the candidate that they want. They end up hiring within the first week or two of posting on Paraform. And these are like director of Engineering at High Touch. These are very, very difficult roles. So I think that's where we create value. It's reducing the time to hire while maintaining and improving quality.

Yeah, fantastic. Thank you so much for stopping by.

I think we had a. It's coming back to me now. I think we had a call maybe in 2023 early on, so it's amazing, amazing to see your progress. Congratulations on the new round. Excited to, to follow along.

We'll talk to you soon. Have a great rest of your day. Really quickly, let me tell you about graphite. Graphite.dev Code review for the Age of AI Graphite helps teams on GitHub Ship fat, ship higher quality software faster and get started for free. We got a $52 million Series B from none other than Anthropics Anthology Fund with Menlo Ventures.

Had breakfast with friend of the show Bailey Barrow this morning and he, he honestly wouldn't stop talking about graphite. He's absolutely obsessed.

He really was.

Did he say graphite fanboy?

He said he's running the TBPN stack. Yeah, he's got all of our linear graphite goes down the list, but it's all good stuff.

Harvey Asana Ramp Replit. They basically. They got every logo.

They got every logo.

They got every logo. But head over to graphics fight.dev and check it out yourself. Who do we got next?

Well, we have some more timeline posts. So Mike Noop, he's been on the show before talking about Arc AGI. Arc AGI 3 was just announced the AI benchmark from Francois Chollet and Mike Noop. And this came as unexpected, which is what?

Aren't people still.

They're still struggling with Arc AGI2. That's right. But the goal for Arc V2 was to. To endure 12 to 18 months. The goal for the B3 is over three years. And so he's talked about this before. When you design a benchmark, you have to assume that if the current models are scoring 50% right now on your benchmark, they are going to solve it in weeks. It's going to be days. People won't even bother with. We were talking about this with the imo with some folks from Cognition. It's like the IMO is so close to being solved that people aren't even really focused on it anymore because they're like, of course we can do that. Let's not bother with it. And so team was saying, when you're designing a benchmark, you need to target. Not only will it get 1%, it'll get 0%. You need to imagine a benchmark that in a year the AI will still be scoring, like, negative 100%. Like, it will get every question, like, so wrong. That. Yeah. That it's. That it's actually conducting even more points. Exactly. Like, you need to really think through the philosophy of what makes a really, really hard benchmark. And they've done it at ARC AGI and we're, we're really happy to support the team over there. So if you're building a foundation model, head over and give it your best shot. And if you're a human, go try and solve one of these prizes. They're pretty fun.

They're pretty fun. More breaking news. Quinton Farmer, one of the co founders of Tolands, has announced a Series A led by Raboy over at Coastal Adventures.

Congratulations.

They launched just a few months ago. They have 3 million downloads, $12 million run rate, and now adding this $20 million Series A.

Wait, is this that?

This is the one they're using?

Fun like, like Tamagotchi on it.

I think it's like purple, right? Purple dino things? No, purple aliens.

Yeah. Alien best friend.

Yep. Brilliant. Absolutely ripping 7.

76,000 reviews on the iOS app store. 4.8. Wow. An alien best friend. Very cool. And somehow less dystopian than the AI boyfriend. Yeah.

Wired has an article today that just came out. What could a healthy AI grow companion look like? So if you're looking for a healthy AI companion, love it.

Why don't you bring the gong for them and then we'll bring in our next guest.

Let's do it. Let's do it.

Hit that gong. Jordy for toen. Congratulations to the team over there. You're welcome to come on the show and, and you're welcome to send one of your alien friends over to the show to do the interview for you.

Send them on.

Send them over. Anyway, we have our next guest, Amit from Luma AI coming on the stream. How are you doing? Good to meet you.

Doing very well. Great to meet you as well and thanks for having me.

Of course. Welcome to the stream. Why don't you kick us off with an introduction on yourself, the company, any breaking news for us?

Awesome. So my name is Amit. I'm one of the co founders and CEO of Luma. At Luma, we are building basically what we consider to be the future of multimodal intelligence that includes what comes after LLMs, building multimodal models, models that learn from audio, video, language, image, altogether. Our first product is video generation models, models that are able to generate video, audio altogether. And currently we are working with Hollywood, we are working with advertising agencies, we are working with also a lot of individual creators and we have about multiple tens of millions of users. We just don't talk about the exact number of users.

Sure, sure, sure. What explain to me multimodal models like are you trying to tokenize everything into some sort of standard format so that an image, a video code, text, it all appears as tokens in the same kind of stream, or are you creating a system that can kind of shift gears between one and the next?

That's a great question. And this is basically the core differentiation. So right now most people are doing the latter. Most of the AI labs that have large language models, they're using those as backbones and trying to teach them how to read image and video as token. But the problem with that approach is that the output is very, very subpar. And even on the most basic tasks, second generation deep learning models, which are your convolutional networks, outperform on most VLM tasks. This is the first approach which is you create a joint latent space where all of this information is in the same representation and then when you learn on it, it is quite analogous to how brain learns a human brain. Audio, video, image, they're not really terribly different. For us these are all just signals. And when we think of an event like let's say dropping a glass, we can hear it in our head, we can see in our head, and we also can predict if it's a piece of glass, it might break. So currently this disparate approach of training a language model and a video model and an image model, things like that, this is going to get old very quickly because on language model side, we're kind of running out of data, as you're seeing with GPT5 or the previous thing that was called GPT5 10 times more compute 10 times larger model, but the performance isn't much better than GPT4. So it was retrogressively renamed to 4.0. On image and video model side, even the models we make today are kind of dumb. They're basically translators that take text or prompt and translate them into video, which is interesting, but that doesn't solve someone's problem. And the problem being hey, explain to me how something works or tell me the story or here's a script, make me the whole thing that needs a level of reasoning and understanding that these models don't have. So well, you combine them all together then one, you solve the data problem because just immense amount of image, audio, video data in the world and two, you are able to build intelligence that goes beyond what LLMs are able to do today.

Talk to me about images in ChatGPT. Do you think that there are multiple layers going on there? I was shocked by the quality of text and then the quality of the studio Ghibli moments. It feels like style transfer on steroids. But I was particularly surprised they could do both. And it felt like, it felt almost like two different layers running. And I was trying to do some tests where I would have generate some text and then have a snake weave through in front and behind the text. And it was kind of struggling with that and it was making me feel like maybe there was like a second pass here. Is that a reasonable theory? And also I'm not saying that as like a knock on anything because if I went to a human and asked how a human would, would design a, a birthday card with a, you know, a dinosaur in the background and then happy birthday text on top, they would draw the dinosaur and then they would use typesetting or they would draw the text on top. And so I don't, I don't necessarily think that's like the wrong path. I'm not knocking the strategy, I'm just curious about what is actually going on on in that model because it was remarkable but it raised a bunch of questions for me.

Absolutely. So to the best of our understanding, basically they're doing autoregressive image generation as the first pass that produces a blurry low resolution image or somewhat high resolution image, but you're missing high frequency details and then you do a diffusion pass on it to get to a decent quality image representation. Now so text is one thing, but I believe text actually text rendering in image models, especially like the latest image models. And then what we're seeing in our next Photon two model, that's actually getting very good. So that is not the problem. You know what's really interesting though, is structure. So when you do image rendering with just a diffusion model, like stable diffusion or name your pick, you rely on the text encoder of the model to kind of tell you about the semantics of what the person is asking about. And then you just rely on the diffusion model to do all the intelligence. Let's say you're asking it, hey, what is two plus two? If there's nothing in front of it like an LLM that is able to say, oh, user is asking 2 plus 2. So let me draw equals to 4. If you don't have that, then these models have ridiculously zero understanding of whatever you're trying to say. Right. It will just render two plus two equals two and stop doing it. Right. That's what I mean, zero understanding. But when you have an LLM generating it or a thing that has good understanding of language and it's unified, so they call it the four omni, then in language space, especially within the model, you can reason about it that like, oh, two plus two should be four. And let me render that. Now what I believe though, GPT4O is it's still a system rather than just a model.

Sure.

Right. So when you type something, it still goes through GPT. Right. They do all the prompt enhancement and reasoning about it in language space and then give it to the model and it does a very good job of rendering it. But when you ask for complex layouts, then I think that that strength starts to show. But of course, you gave an example of like, you know, the snake coming in and out. So this is again the question of prompt alignment. And Violet is good, it's not perfect yet.

Do you have any reaction to. In the Wall Street Journal today, there's. Hollywood wants AI protection. There's a lot of back and forth there. I was particularly shocked that there's been a number.

Save us, save us.

I mean, there's been a number of interesting moments in AI imagery and AI video. Obviously, like the narrative, the general narrative is like the technologists are like, this is amazing. We've created this amazing algorithm. And I agree with that. And then the backlash is like, we don't want all the, you know, the, the photographers to be put out of business. But the more interesting stories that have emerged are like, VO3 and Google seem to be able to generate Disney IP with perfect accuracy, no problems. And maybe that's because they have some deal with YouTube. And then when the Ghibli moment happened. I was like, wait, like, like, like GPT images and ChatGPT would reject different things for, for safety reasons, but also for intellectual property reasons. But not Studio Ghibli, which is a style, but it's also intellectual property and it's not exactly like the cleanest thing. So. But maybe there's a deal behind the scene. So how do you think, how do you think, like the, like the, the, the interplay between the startups and the technologists and the legacy media. Hollywood, Washington D.C. what stories are you tracking? How is that evolving?

Right, so there's two axes that you're talking about. One is basically the adoption and job losses and things like that, right? That is a narrative. And the second is the IP itself. And how do you think about that? So on the first one, it's actually at this point a foregone conclusion that unless Hollywood and the traditional way of producing media changes, they're on the path to extinction. There is really no two ways about it. And of course there are arguments like, oh, but it can't do this, it can't do this, it can't do this. The thing is, we spent last year, 2024 building out this infrastructure for training these really large, really capable models. And you know, it trains on petabytes and petabytes of data and much more complex than LLMs, because LLMs, of course, they're complicated to train, but you still have very little data. The amount of engineering that is required to train multimodal models is insane. Anyway, now that the infrastructure exists, the rate of progress is very, very fast. We saw the same thing in language, right? Initially everyone had to build from 0 to 1, but once you have 0 to 1, they're like turning out models takes a quarter, sometimes less, right? So anything that they think, oh, it can't do today yet, like, you know, it can't do 4K, well, tomorrow it's gonna like, we already do 4K. Oh, it is not able to let me control like, you know, exactly the pose I do. We launched this thing called modify video where you can give it your camera feed as a prompt and you can control exactly any camera movement, any pose, anything that you want. Right? So we work with some of our great partners in Hollywood and like, you know, some of them are very forward, some of them are very much not there today. But when you think about what a studio is, a studio is two parts. A financial institution that is responsible for green lighting things and managing, okay, this is the area we do go and this is the area we don't go and all these kind of things. And a production part, the production part largely works by contracting other smaller production houses, but generally that's the structure. AI especially generative media changes the economics so wildly from $100,000aminute and sometimes a million dollars a minute all the way down to $10aminute, $100aminute. When that is the level of economic change and you follow the money, there's just no scenario in which you don't come out thinking that, okay, the old way of doing is not going to work.

Yeah.

An interesting thing I've been thinking about is if the cost of production drops by an order of magnitude just by being able to shoot some, maybe you're shooting some scenes with, with regular actors and cameras and union crews, but then you're able to create certain scenes or moments. It seems like the Internet has this insatiable demand for content. And it feels like when you look at some parts of the sort of entertainment production stack, great writers might be more in demand than ever because like they'll have to write more like, but.

But sort of like stories would be more in demand.

Like you know, you can imagine a world where HBO is putting out, you know, an iconic, you know, potentially iconic new property or franchise or like shipping them monthly. Right. When historically maybe they only had one big, you know, one or two massive releases a year like a Game of Thrones. And so I actually think that the consumer is going to win in the industry. The people that adapt within the industry and say I'm going to lean into using these tools so I can write better, I can produce. Or even the idea that I'm curious if scripts are now. Certain scripts are now being encouraged. Bring your script with an idea of what this world looks and feels like so that it's not just this static text based concept. Right. Show me, give me the pilot. Basically because.

Exactly.

And it's top of mind because today we released an ad that we handmade farm to table content for one of our partners, Wander. And there's like a few scenes in there that you probably could have AI generated. We didn't, but they cost us real money. Right. Like that we could have potentially stripped out. And it feels like, feels like we're so close.

Yeah. So here's the over and under on that. Actually the demand for video is absolutely absurd. You're right. Like, you know, so average person is watching about three and a half hours of video on their phones every day.

I don't know if it's good, but It's a big number. We're certainly, you know, might be contributing to that.

But it's true. And the thinking about that is very simple, that people don't like to read. If they could watch an explanation, watch how to do something, watch a piece of news, whatever have you, they would do that over reading it or reading about it all day, all night. So that means for entertainment, yes, generally entertainment we think about, okay, you make it once and then millions of people will watch it again and again, all these kind of things. But when you think about music and you think about question and answer and you think about explanations and all these kind of things, that's ridiculous, right? Like, you know, you can't generate a video for every explanation for every scenario on every planet. It has to be done dynamically, it has to be done automatically, and it has to be done with AI. Now the number I track is, okay, for every person on the planet. Like, you know, out of that 3.5 hours that they watched. If we can generate at least an hour of video a day, that is what the success metric looks like for us. So that is looks like about 6 billion hours of video per day generated. One, there's no human capacity that can actually generate that kind of video, right? Editors can do that. And two, there's no computer in the universe at the moment to be able to actually do that. But that is what to track towards. Because video is now the substrate of information on the Internet. There's two things happening to the Internet, right? One, it's becoming zero click, which is like, you don't. You don't hunt around in websites and try to like, oh, this the answer, this answer. No, you just ask an LLM or now Google also has AI mode. If you're in the labs experiment, then you just get the damn answer. And second thing that is happening is the substrate of information is changing from text to video. This is what people do. So if you just extrapolate from this trend, video production has to change, entertainment has to change. And people who don't come along, like, you know, it's going to be very hard journey for that advertising. On the other, we can talk about advertising for about an hour. And like, you know, Cannes was very illuminating for me in that way about, like, how people are thinking about it.

Yeah, I want to get the full update from Khan. Is this your first time going Break it down? Like what? Yeah, this is the Cannes Film Festival, right?

I mean, one of the exciting things about what you're doing is when you look at the cost Structures of consumer brands which we both have exposure to so much. Oftentimes consumer brands are held back because they don't have net new ad creative. And so not only is it a cost, it just takes a long time to produce. And so you have these gaps where a brand will see lower performance, slowing growth, et cetera. And it's really just because they don't have access to net new ad creative that can drive new demand. And so that's an entire area that I don't think that's being fully priced in yet. How the great brands will actually be able to grow faster because of just being able to day by day generate net new ad creative.

Yeah, give us the full con recap. Sorry, answer that question and then I want the con recap.

Yeah, I think the answer to that question and can is almost the same. So. Yes, it was, it was my first time. Beautiful place to be honest with you. But, but really interesting. Like you know, one everybody, I mean it was, it's not an AI event, it's not an AI conference, but literally every conversation was about AI. Every conversation. But what's really interesting is that maybe 1% of the people knew what the hell they were talking about.

Nobody knows. But everybody's excited.

Yes, but I got blisters in my ears hearing the word agent. All these like, you know, agents.

Really?

Yeah, all these agency people pitching, right? Like oh, we have AI agents to do this and we have AI agents, we are building this to do this. This in fact like you know, Adobe was there and, and their salespeople were trying to pitch non existent AI agents to people and trying to log them into 3 year contracts, right? It was palpable, the level of desperation.

Shots fired. Shots fired.

But it's insane. And, and, and when you go and ask like, okay, what do you mean, right? You know, we, we talk to, we talk to everyone and, and, and, and what is it? It's a prompt box where you can select a bunch of models. That's it. That is not an AI agent. That doesn't do anything.

Right Senator, that's SaaS.

That is SaaS. Exactly. And the issue is that basically a lot of people are thinking, okay, well developing this is basically nothing and we'll just produce the models and it's going to happen. Developing models, developing agents in creative space, in visual space is very difficult work, right? You need new kinds of models, the kinds we are training. Because think about it, Gemini is the best VLM that is out there right now and it is barely able to meet about 3% of our benchmark of visual iteration. Let's say you're talking about brands. So we met quite a few, including our partners like Coca Cola and others. If you want to make an ad, and then let's not think about a Super bowl. If you just want to make creatives, say 100 ads for different markets, getting them to be right, getting them to be on brand, getting them to look exactly how they should look like, that's a hard job. Current image and video models, as I was telling you, are dumb. They will just make whatever they have no idea. And then when you use VLMs to critique them, the outputs are barely any good.

So.

You need the next thing after LLMs, the next kind of models that are able to understand visual information, instructions and what makes something engaging. That level of information is what you need. Like LMS are getting there, right? Like for stories, when you write text. LLMs are now making things that are very engaging. Now we need that next jump in models. So that was my overall impression in canon. The second thing, I felt like by next year this place is going to look very, very different. Fortunes are going to change significantly because. Because when you talk to some of the most senior folks in various agencies, whatever have you, the writing is on the wall for them. They employ hundreds of thousands of artists and people and they are asking us, oh, how do we actually do this more efficiently? And the thing is, that's incorrect way of thinking about it. So long story short, it was not an AI conference, but it ended up being an AI conference where nobody knew what the hell actually the AI is and consequently nobody knew what is coming.

What areas do you think are safe from disruption? We talked to a photographer who does stuff in CPG and he was saying there actually could be a situation where if you're McDonald's you can't generate AI images of your burger because it could be.

There are already FTC rules around. You have to use the real burger because there were all these like the false advertising lawsuits where they show the perfect cinematography burger that was perfectly made and looked amazing. And then you go expectation versus reality. And those, those memes went viral enough and the, the flat disgusting burger bubbled up and eventually I guess there was a lawsuit. And so now you know and even in if you're selling a phone, you have to say screen images simulated or if you're driving on a track, you have to say closed course professional driver. That's for safety reasons, obviously drink responsibly if it's an alcohol. So the FTC has A bunch of rules. Do we think there'll be anything there on the AI side?

I think those rules will change too. Yeah, basically. So currently, like. So the Burger example, right. The problem is you are over representing the product yourself.

Yes.

Right. It is not the problem of the tooling you used to make it with. You could have made it how the thing actually is. I mean, that would look very disappointing in the ads, but you could have made it closer to how the thing actually is, such that people aren't misleading thinking like, oh, I'm going to get this big giant thing which is very fluffy. And then when you get in, like, oof, this is the thing.

Right.

So it's really not about the tool you used, it is about the intention of what you had behind it. You clearly wanted to mislead. So I think the FTC rule is right, but the rule should not be, oh, they should not be created with CGI or with AI or whatever have you. The rules should be don't misrepresent your product. And for me, basically the singular thing to think about is follow the money. Right? So when you're actually in advertising, the point you were making is really, really salient, which is advertising is all about stats, right? And something that performs well, something doesn't perform well. It's very hard to predict how quickly can you replace the thing that is not performing well. And in the gradient of the things that are performing well, that will lead to the best output today, that cannot be done automatically at all in any visual form of advertising. That is done for text advertising. It has been algorithmic for a very long time now, and that results in very, very good outputs. So when companies actually are faced with the idea like, okay, well, on our other products where we can do this, it performs so well. And on these products, no, the regulations, they will lobby for the regulations to change all this kind of thing. And I think the regulation should be don't misrepresent your product, not don't use AI.

Yeah, last question for me. What about more narrow AI use cases in Hollywood? I'm interested in like, just, just like in the VFX pipeline. There are. One second, let's do. Sorry, I'm scheduling someone tomorrow.

We're doing it live, the show biz.

So my question now, we're gonna clip this so it doesn't matter. My question is more narrow VFX use cases. Like, is there a single company that's just doing amazing green screen rotoscoping? I've seen Runway do that in like kind of the Prosumer arena. But there's so many different pieces of the pipeline that I think artists would be much less jostled by if it's just, hey, I have to track a camera all the time in 3D. You know, there's services that do this using traditional like CPU based pipelines and solvers. There's imus that measure the movement of the camera. No one really cares about that particular role. It's kind of just a hassle that they have to deal with. And so all sorts of different, more minor vertical AI use cases, those feel like point solutions that could be deployed right now. People could go and get a foothold, maybe they expand to prompt to movie. But it is in the meantime, like how about we just speed up the rotoscoping?

Yeah, I mean, why not? Right? So for instance, like, you know, we released this thing, it's called reference frame. So what it does is you give it video in any aspect ratio and it's able to generate any aspect ratio out of it without cropping. So it will generate if you give it a 16,9 vertical video and now you want to share it on YouTube, for instance, where vertical videos don't perform well, you want something that is horizontal or 16, 9, ideally you can generate any of these variants, any number of these variants, all you want. And it's not a point solution. Doesn't matter how the original video was made. You could have totally created that clip using any traditional methods or whatever, have you? So I think this is, so here's something interesting, right? This is not what you asked me, but I promise I'll lead to the answer you're talking about. So at least in my life since I've been conscious about this, there's been two really big changes before AI. One was mobile and mobile was a very large consumer side change, right? People change their behaviors drastically. Not overnight obviously, but it looked like overnight, to be honest with you. And then the second one happened about a few years later, which was cloud and cloud consumers, it didn't affect them too much, right? Maybe enterprises. Yeah, but it affected companies. And companies were done inside out things they used to do on paper, things they used to do on prem, things they used to do like, you know, so these were two very radical transformations, but they happen to different segments of the world. AI is for the first time, at least in my memory, a transformation that is happening on consumer and enterprise levels. It's basically rewiring companies from the inside and it is changing consumer behavior like nothing else before. Right. So coming back to your Point. Now you can take the slow route. You can say like, okay, well, let's do rotoscoping. But the tiny studio down the street that used to probably just like, you know, create frames for you up until last year is now going to start releasing stuff that like, you know, looks kind of as good as yours and now, but they can do like, you know, at a fast clip. And they're releasing like, you know, weekly episodes and then iterating on it. And how long will it take before their IP is actually bigger than yours?

Yeah, right. Good question.

We have seen this happen on YouTube again and again and again and again. Coco Master, right? It started from nothing. And look at this now. So bigger than a lot of Disney IPs, right? And Disney, I think Disney bought them. Disney tried to buy them. Something happened there. So this is basically the difference. You can do like, you know, and there's nothing wrong with it. You should do the small things. But if you are in the business of producing content, if you are in the business of thinking about advertising, if you're in the business, business of touching pixels, the world is not the same as it was two years ago. And not in a small way, in a very, very deep way. Val, it might not solve the whole thing for you at the moment, but you have to start really thinking about what would the economics look like a year from now, two years from now, and five years from now, and what should the business be like? Exactly?

It's great.

Well, this has been great. We would love to have you back on to. There's so many different topics we can cover, so. And I appreciate how freely that you're willing to speak on. On so many different. On so many different topics.

So that's the only way.

That's the only way. You just got to be yourself. Awesome. Well, thank you for tuning in.

We'll talk to you soon.

Talk soon.

Thanks so much.

Cheers.

See you guys.

And we have some breaking news from tbpn. Let's go to the printer cam, which.

I believe the timeline.

Here we go. Printer cam's ready.

Terminal.

I see it spinning up. We got some breaking news. We got some breaking news. What is this?

He's timeline is in turmoil. Special segment starting now.

I love this because it's.

Soham Parikh just became famous in the last few hours. Not for the reasons that he would necessarily want to be famous.

No, no. Is this a real picture of him? Do we know?

Or is he swinging pictures?

But this is from Var Epsilon says breaking meta OpenAI anthropic and Google have signed Soham Parikh as fractional Chief AI Officer. Obviously a joke, but he's been on the timeline all day because apparently he's been working at multiple YC companies. Multiple companies at the same time. What's funny is that we were. We were working on a. During the show we were working on a post. We couldn't decide how to frame the trade deal of Soham and. But some fans just did it for us.

Happenstance. Fully sent it. So Matt Parkers from Anti Metal was actually. Soham was their first engineering hire in 2022. Really smart and likable. Enjoyed working with him. We realized pretty quickly that he was working at multiple companies and we let him go. I can't imagine the amount of equity he's left on the table. Matt says hiring Soham is a new rite of passage. Tbh. Any grand company should go through it. So Suhail was the first person to call this out. He said this morning, or actually it was last night. So this has been building psa. There's a guy named Soham Parikh in India who works at three to four startups at the same time. He's been preying on YC companies and more. Gary is not going to be happy about this. You do not. You don't want to put up the bat. You don't want Gary people to put up the Gary side.

No. You don't want to go up to gangster.

But apparently he's dorky companies. I fired this guy in his first week and told him to stop lying, scamming people.

Oh, this is fascinating. Did you see this from Pratika Meta? Tyler put this in that. You want to read this one first?

He's been all over the place. Roy said, I interviewed this guy yesterday.

No way. That's amazing. Tyler, break it down for us. What does Pratika have to say?

Okay, so apparently, I mean, this is still unconfirmed. I've actually, I've still. I sent him an email.

Okay. I obtained his email. I'm not sure if it's the real one.

We'll see. But she says the dude clears interviews. He's good with that. After clearing, he has junior people doing.

All the work and he has around.

20 employees and interns.

Basically a small dev shop. Wow. So he's just figured out the economics of like, hey, if I'm making 200k.

Or 300k, he's basically signing people to a 20k a month retainer. He knows he's going to have some churn. But it doesn't Matter. He's picking up some equity.

This is fascinating.

Equity comp as well.

Yes.

I wonder if he's vested I wonder.

If he's vested some shares on the.

Balance sheet at any of these companies. Ken Watana says Soham Parikh is the Bonnie Blue of yc. Very vulgar. Sar says all recruiting must stop until we get Soham on TVPN to figure out what is going on.

Open invite to Soham. We'll.

Santiago says Soham on TVPN when Daniel growing Daniel says Microsoft just laid out.

Here we go.

Laid off 9,000 workers, all of them. So if you.

I wonder if he doesn't have Soham Parikh in their inbox. It's time to start polishing your resume. Because he hasn't reached out, it means he's now bullish on you.

Austin Allred says, must suck being an unemployed software engineer and realizing that Soham Parikh has been hired 79 times in the past four years. And then Matthew Berman says, oh, you think AI is killing software jobs? Tell that to Soham Parikh. Signal says, not now, honey. There is a Soham gate happening on X. It is really funny. Is basically say, I'm going to build a dev shop, but it's going to be my name and I'm, you know, he probably, no, he probably could have had a just very successful dev shop.

Yeah, yeah, but, but, but there's a stigma around hiring dev shops. You want to hire founding engineers. Like, you know, the title matters and the idea like, okay, we have a full time employee, we got somebody who's delivering. They're a high performer. Our team is cracked. We don't have to outsource the whole thing with Uber. Remember Uber's initial build out was all outsourced and it was like, it was like a mark on the company. Obviously didn't affect them. They went on a generational run. It was fantastic. But there's this, there's this idea of like if you're building a startup, you need. The founders need to be coding, the employees need to be coding. Needs to be handcrafted, can't outsource. But you know, he's proven us all wrong, I guess.

True. So we're working on getting him live. We got another post here from Growing Daniel.

Has he gotten back to you? Tyler on email? It seems like he might not come on today, but maybe we can pitch him and get him on.

Hasn't got back. I sent it 20 minutes ago.

Okay.

I think he needs to come out and have an accident. Sorry. He needs to come out and say he's sorry.

Yeah.

And he needs to say this is the right time to launch his dev shop.

Yeah.

Yeah. And it should be called the dev shop of Soham Parikh.

Yes. I mean, I don't know, maybe he's.

Delivering, but Growing Daniel has a post here. Chief of staff brings list of applicants. Really likes this one Indian guy asks if it's a rune or Soho. She doesn't understand. Pull out illustrated diagram explaining the difference between runes and so she laughs and says it's a good engineer, sir. It's so hot.

Oh, that's a wild west. It's great. Oh, this is great. Anyway, anything else that we need to cover from the news today? Apparently Crumbl cookies have 62% of your daily calorie allowance.

Carter says staying on. So hum. Gabby Goldberg says I'm at the intersection of soham parikh and worker 17.

Oh yeah, worker 17.

Oh, we got another one. Growing Daniel is going on an absolute terror. We're just going to keep these coming. Pull this up in the chat.

Posting Soham Parikh new CS grads. Yes. Soham Parikh's really soaking up the job market.

He'd been sitting over this post being like the job market is actually fine. You guys are just bad at interviewing.

Yeah.

Jacob Postle says, introducing our newest product. Soham, the AI coding agent that can work for anyone. Any code base 24. 7. We've been testing in stealth for a while and results are looking excellent. Please reach out.

I wonder. Yeah, I wonder if he's a cursor user or what he's got. Maybe he's got Devin under the hood. Maybe he's got Claude code. Who knows? We gotta. We definitely need to know Soham's text stack for sure. That is interesting.

Ryan Peterson says Soham never even applied to Flexport. Sad face emoji. Ryan. Ryan would have caught him.

Delete that. Don't say that.

Wow. Adam over at Warp CTO has a few. He got to a round two. So hum. Got to a round two over at Warp. Sounds like he didn't make a hire. So it's fascinating. Daniel says Worker 17's name Soham Parikh. Antonio says he was Mark Safe from Soham Parikh today.

He did tear up the Valley today. Completely. Completely. Anything else we should cover in the timeline? We have someone in the House of Representatives is coming out against car headlights. They're saying they're too bright and I'm going to do something about this. Brightness has doubled in the last decade, but manufacturing standards haven't budged in 40 years. Drivers and oncoming traffic deserve visibility. It matters for seniors in rural roads. Are you anti bright lights?

I am. I think we could tone them down a bit.

Yeah. Really?

I think we tone them down a bit.

Interesting.

I don't drive at night much, but when I do, I'm reminded that lights kind of peaked probably in the early 2000s. They were solid.

Yeah.

We didn't have to keep improving.

They also got bluer because they went to led.

I think it will lead to tungsten. Widespread blindness.

I would say if you're driving a car, just, just light a candle and hold out a. Hold out a lantern. A lantern.

Like you're, like on a fishing rod almost your car.

Yeah. No, no, no. Just, just, just while you're driving like this. Yeah, just hold out the candle. The candle, lamp, lantern. That's great. We got it. We got a highlight signal before we head out. Signal Labs is unapologetically a wrapper company. We use state of the art models made by others and give credit where credit's due. Our first product is so simple, when you see it, you'll probably say, wait, that's it. That's really the point. That's our ideal reaction. I've been using it for two weeks and genuinely can't imagine life without it. At this point, we don't have any evals, no real qa other than our own intuition. But it surprises you and makes you feel something. Hopefully everyone enjoys as much as I do. So very excited for Signal Signals Launch.

Get us an invite.

Yeah, send it. Send the invite. We'd love to. We'd love to test drive it and we'd love to have you on the show when you launch.

Can't wait.

Anything else worth covering before we hop off?

I mean, there's so much news. There's so much news. This Flo actually hired Flo Crevelio? Yeah, he hired him a week ago. Fired him this morning after seeing, seeing Suhail's post. It's just funny that the guy wouldn't kind of mix up the name. It's a small industry. Stuff's going to get around. But.

Anyways, it is crazy. Yeah, you'd think different names, different GitHub profiles, different LinkedIns, that wouldn't be that hard. Actually, the best way to run this scheme would be to set up a number of different folks and Instead of having one person with 20 junior devs underneath, have five people each with three underneath. So you're still doing the subcontracting, but there's only One face per client, essentially. And this is basically, you don't want.

To give them too many ideas.

This is basically what marketing like, you go to like RGA or Wieden, they'll be like, here's your killer, you know, Don Draper, like, this is your account representative. They'll be coming up with the ideas and they come in once a quarter, give you a presentation on, you should do the super bowl ad and like, and investment banks do the same thing. Hey, we, you know, we'd love to do a private placement. We'd love to take you public. Here's this, you know, 200 page deck on why your business is, is amazing and how much money we can get for it. Who built that deck? A bunch of analysts working hundred hour weeks. That's right. For sure.

Well, the next 24 hours matter a lot for Sohan Parikh. We will be watching. We'll hopefully get him on the show tomorrow.

But do you think he's going to raise money?

He might. I definitely saw somebody joking around saying that he raised a series A from some friends of ours.

Tier one, tier one. Yeah, it's totally possible. But. But do you think the greatest irony.

Would be is he's like, sorry guys, what that guy was joking about a few minutes ago saying like, yeah, I've actually been building a coding agent. I named it after myself. This was the best way to test if it works. It's not as good as a 1x engineer right now, but it's good enough to keep a job for a few months. And so anyways, we will be watching.

His response violation on Soham. It is really dominating the timeline.

I'm Soham.

Oh yeah, you posted timeline and turmoil. Buy markets in turmoil, sell timeline and turmoil. Then I'm seeing pictures from our billboard. Oh yeah, it's funny, you just refresh and you just get more Soham. Everyone's having fun. I love, I love a good, I love a good current thing. You know, we started the show with the current thing. It was Erebor, the new crypto bank. We ended the show. We lived through it. We experienced the transition though. The old current thing is out. The new current thing is sohamperique. Fantastic. I wonder. We got to get his tech stack and I want to know how much he's actually pulling in. How much do you think this.

I, I would expect it, I would expect if, if people are triangulating now that he was maybe had three to four jobs at once. I'm gonna make the bold bet that he's actually has like I could see him having 20 30.

You think so?

Just imagining a really hardworking.

A lot of companies aren't posting and.

A lot of companies aren't public and he could be just finding some random SaaS company in the middle of nowhere. Not a lot of management oversight companies aren't in founder mode. You know, a few PRs a day and chugging along. The challenge for him.

Two million.

I'm going over, over, over. The challenge would be daily stand ups, right?

Yep. It's like you got to see him.

But then again but, but, but he's kind of on the press. He's kind of on the precipice of this video. AI models. He could have other versions of himself, voice modulation and AI generated images so you can have his voice.

Well the thing you can do is you can puppeteer the image so you have a different person. A real human's joining real human.

He just needs a video that he loops and it plays at the right time. It says nothing from my end. Thanks.

Yeah, somebody got to post that. That's definitely a banger. Like so Humparique when he. When he says nothing from my end. Thanks. This is the. This we're going to go through all of the current. All the current thing memes. The honey, how did you get. How did our family get so rich? Oh my. Your father worked at seven different jobs with a team of junior engineers.

This was kind of. I mean this was happening more in 2021 where somebody would have a big tech job but. And they were kind of bored. They were kind of just. They weren't working too hard and they would get. They would just start moonlighting at a startup.

That was a common thing during COVID.

And I expected a company to start that actually enabled anonymous work. Right. Didn't. Didn't. It was very much like it's typically a top signal when people start moonlighting too hard.

Yeah, probably. You've been on it. You've been close to calling the top. Top toying with it.

I've been toying with calling the top.

We'll see after the Soham saga. Pretty sure every few. Pretty sure very few YC companies will hire remote Indians. Classic case of one guy exploiting high trust society which leads to downfall of the others around him. Says Varunam Garesh from Warp. I think you already read this but Gary Tan chimed in. He said if this is true it would be sad. The Soham saga doesn't seem widespread to me. He seems like a five star standard deviation above the mean kind of fraud. Decent Startup idea in here though, there's probably a space for remote employee certification. Interesting. Yeah. I was thinking about it like the, you should be able to have an AI that, that looks over the GitHub. The GitHub. When I first heard about this, I thought he was directly moonlighting like it was like from, from, you know, you get a job on the east coast, you get a job in Chicago, you get a job in Denver and you get a job in, in San Francisco and then yeah, you're just going from one stand up to the next. And then you work three hours at one job, three hours the next job, three hours of the next job, three hours the next job. And I thought it was just him. I didn't realize that he had teams under him. But in that, in that scenario you would think something. Oh well like all the, all the pull requests come in at the same time every day and he's not doing like he's not sending slack messages randomly throughout the day. He sends them only in this two hour block. But that's not true if he's sending like no, there's actually two or three people that are dedicated to your account basically like all day long. Right. So I don't know how you do remote certification. You need to have like the camera on. We need worldcoin. This is a world Coin application.

Atlas says breaking Soham Parikh closes a $15 million round led by a 16Z 2000 likes.

This is great. Reid Hoffman says.

Reid Hoffman's chiming in.

Yeah. So do you think Soham Parikh's LinkedIn. What do you think Soham Parikh's LinkedIn editor is. This is hilarious. The founder of LinkedIn chiming in. It's great. Opens timeline. It's all so hum. Is Beth the timelines in terminal baby.

I. You would think that this would be something that book faced would have solved. Right. One of, one of the issues here is that when people make a poor hire, they don't exactly want to go tell a bunch of other people that they messed up.

Yeah.

And so you can imagine people just experiencing this and not wanting to talk about it and now it's becoming a meme to talk about it. There's a guy, I think this account was just created. Well it's created in February, he says. I think he's just joking around. He must have just changed the name.

I think he just changed his name.

I was fired from being a founding engineer, principal engineer, back end engineer and front end engineer today. If anyone is hiring for all four.

This is Crazy.

And this is clearly a fake account. He says, I make more than 99% of YC startups. This is the real reason they're mad.

This is very funny. Everyone asks, where is Sohampreek? Nobody asked, how is so Pamparikh? I'm glad we found this account. Whoever's ready is great. Elon can work at nine startups. Jamie Dimon can be on the board of multiple companies. But when you do it, it's scam. Oh, this is gonna be a good account to follow. Everyone should follow this. Phantomthreadd. This is definitely not him. But I mean if this is actually him. And he's a great poster too. That's a quadruple threat. Quintuple threat. We'd love to see it.

Aidan over at OpenAI says hey, sorry, was out this morning. Who's Soham Parikh? Must be pretty popular name because I think my accountant and front end contractor have the same name.

Wild.

I think, I think, I think the next move from here prediction clearly give Soham a maxed out contract for sure.

We've got another 24 hours. It's just too good of a meme.

Trade deal in the works.

Get him to America. Shoot a viral video. This is a no brainer. There's got to be. There's got to be a viral video in here somewhere. We are actively monitoring the situation. For sure. For sure.

We are. Yeah. Bobby says, looking back through our engineering applicants and Soham never applied. Chat. Am I cooked?

Yeah. Everyone search their email right now for sohelm. It's amazing. Okay.

Anyways, I think that's a good place to wrap. It has been a fun show today. Hopefully more breaking news on Soham. If we find him, send him our way.

We'll talk to you soon. Leave us 5 stars on Apple podcasts and Spotify and we will talk to you tomorrow. Have a great day. Bye.