Our Super Bowl Ad, Walmart hits $1T, Ken is Sick of Griftin | Dara Khosrowshahi, Mati Staniszewski & Andrew Reed, Gergely Orosz, Mitchell Green, Simon Hørup Eskildsen, KJ Dhaliwal, Nicolas Sharp

You're watching TVPN.

0:00

Wow, look at those flashes. Today's Wednesday, February 4, 2026. We are live from the TVPN Ultradome, the Temple of technology, the fortress of finance, the capital of capital. Let me tell you about ramp.com, time is money save. Both easy use, corporate cards, bill pay, accounting and a whole lot more all in one place. Ramp's doing a Super bowl ad. We're doing a Super bowl ad.

0:02

Get ready to get super bowl ad.

0:18

Ramp's in our super bowl ad. We're very excited for the super bowl just a couple days away. This was a very fun project doing. Do you want to take us through the thesis and sort of what we put together with this? Since I just got all the credit, but I had nothing to do with it, basically.

0:19

Let me pull up the ad week.

0:38

There was a very nice post from Blake Scholl over at Boom Supersonic and he said there's Clevver and then there's John Coogan and Jordy Hayes at tbpn. Clever. Behold, a masterclass in marketing. And I had nothing to do with it and I get the credit.

0:41

This was of course, Dylan on our team led the charge here. This is something he's wanted to do for a long time. We had done something similar back at Party Round. We ran a billboard with a bunch of different friends, companies, customers, et cetera. Did really well. And so doing this is the final stage in terms of advertising.

0:57

Fun fact. Dylan also spoke this week at the New York Stock Exchange. Want to change the world, Raise capital at the New York Stock Exchange. Very happy to be partnered with them.

1:16

Kendra Barnett over at adweek covered our ad. Yes, of course. And she says the 15 second spot isn't selling anything. It's simply what co host Jordy Hayes calls a love letter to our community. That's really what it is. TVPN is nothing without the community, the people that join the show. We made the 15 second spot in house. We featured a bunch of our guests and then if you've been on the show as a guest, whether it was for five minutes or five hours, your logo made it in here. So yeah, this will be a doing it a regional buy. We basically looked at where the majority of our audience was and we bought bought segments around that. So very excited for Sunday. And I said in ad week, it's completely unnecessary for a media company to buy advertising. The nature of media is that you're constantly putting out things that are promoting the business naturally just through the content. So why do it? And we basically did it. I said we believe in doing things purely for fun, so we're certainly having fun.

1:25

Some people screenshotted that and texted me to that. I thought that was a good quote. Let me tell you about FIN AI, the number one AI agent for customer service. If you want AI to handle your customer support, go to FIN AI. There was another reason beyond fun. I do think it's an important opportunity to, to introduce the football community to technology, to business. And that's really, that's my goal with this ad. Hopefully let people know if you're watching the Super Bowl. Technology.

2:29

We got to raise awareness for technology and business. Let's play the ad.

2:57

Let's play the ad. Let's see.

3:01

You're watching tvpn.

3:04

If you're watching this podcast, you've already passed the test. It's a great question.

3:06

I think it's super important and awesome.

3:11

This is going to be one of a hundred baggers.

3:12

You guys are the podcast in the world.

3:14

The gong hit. This is great.

3:18

Short, sweet. And we look forward to seeing it live on Sunday.

3:20

Yes.

3:25

Eleven Labs build intelligent, real time conversational agents. Reimagine human technology interaction with 11 labs. We have some massive news from 11 labs. Coming up on the show at 11:30. Let's pull up the linear lineup. I will tell you that linear is the system for modern software development. 70% of enterprise workspaces on linear are using agents. And we have a stacked show. Andrew Reed's back on the show. Matty from 11 Labs is coming on. We have Gerge from the pragmatic Engineer coming on.

3:26

Super pumped for that one.

3:54

Dara from Uber's coming on. Mitchell Green, one of our close friends, is coming on to talk about all sorts of stuff going on the market. And then we have an awesome lightning round lined up. So is there anything else we should cover about the super bowl or should we move on to other super bowl coverage? Because we're not the only ones running ads. Lots of tech.

3:55

Absolutely wild, wild morning.

4:15

Yes.

4:19

I was not expecting this out of Anthropic. They basically took the vibe War was like, effectively relegated to X. Yeah. It just felt like the same hundred thousand people saying, it's so over, we're so back. It's so over. We're so back. They're taking it to the main stage. Right. And it shouldn't be that surprising, right? Anytime Dario gets on a mic anywhere, he's taking shots at open air, but he's.

4:20

He's not taking direct shots.

4:43

Not direct shots.

4:44

He's always.

4:45

And this arguably is not a direct shot either. They just said ads are coming to AI. Yeah, Claude.

4:46

Yeah, yeah.

4:52

And anyway, so.

4:53

But truthfully, like a lot of the previous anthropic advertising has been very in their own lane. They've run a number of campaigns that we're just thinking above the credit, thinking where you're thought, partner. This one does feel like it's a response to what's happening in the industry.

4:54

So, yeah, we'll get into why they took this approach, clearly. Really? I think we're just excited to spike Sam's cortisol. And I think they probably.

5:08

They've definitely been watching some heights.

5:18

Yeah. Should we just play all four ads?

5:21

Let's play at least one of them. We got to play the height maxing one. That one's particularly funny. But are they here in the timeline while we pull those up, Let me tell you about Applovin. Profitable advertising made Easy with Axon AI. Get access to over 1 billion daily active users and grow your business.

5:23

And you were saying earlier off the show you were starting to like anthropic until they came out against that.

5:40

Yes, yes. I see this as an attack on me. I see an attack on one advertiser as an attack on all advertisers. I'm pro ads.

5:45

And this is not. This is not the right ad. We'll come back to this. That's another super bowl ad that we'll.

5:53

Have later in the show. The anthropic attack on advertising, it does cross the line for me. It goes too far. And I think that they should be supporting the advertising economy. Almost 10% of the American workers.

5:59

All right, so let's pull up this one. The first ad is, can I get a six pack quickly? And this one, I immediately thought, okay, definitely not just watching Clavicular. They're studying it, apparently.

6:10

Let's watch it.

6:23

I was not expecting this to be the looks maxing super bowl, but let's play it.

6:24

Do we have it?

6:31

No.

6:31

Almost. While we're doing that, let me tell you about public investing for those who take it seriously. Stocks, options, bonds, crypto, treasuries, and more with great customer service. Without further ado.

6:32

Let's play it.

6:44

Yeah. This is violation.

6:44

It starts out by just saying violation.

6:46

Yeah.

6:48

Perfect.

6:51

That is a clear and achievable goal.

6:52

Would you like me to tailor a personalized workout plan?

6:54

Yes.

6:57

Perfect.

6:59

Let me personalize this for you.

7:00

The lag delay.

7:02

This is a whole meme on Instagram reels. People will do impressions of ChatGPT voicemail like this £140.

7:03

Got it.

7:10

I'll create a plan that focuses on aesthetic strength.

7:11

All right, pause for a second. Like the. The slight delay.

7:14

It's.

7:18

It is iconic.

7:18

So good.

7:19

Yeah.

7:19

And clearly this was written with chat GPT. No, no, no, no. I'm saying.

7:20

What do you mean?

7:27

It's designed to sound like. It sounds exactly like. Got it. I'll create a personalized plan.

7:28

I don't think that's actually how ChatGPT sounds. I think, in fact, when I've watched the Instagram reels, they do the ChatGPT impression, and it's a little more like this is written for comedy. And so the fact that he says. What did he say? Absolutely. Twice in a row. That's something that isn't just built in the gym.

7:34

Tri Step BoostMax. The insoles that add one vertical inch of height and help short king stand tall.

7:54

What? Use code HEIGHTMAXING10 for big discounts. Ads are coming to AI hide not to. Claude is a great ad. What's the difference between me. That heightmaxing thing is so crazy.

8:03

So crazy to run an ad like this and not even do a call to action. Yeah, that was probably intentional.

8:17

Well, the call to action is like Google Claude and you land on that.

8:23

It's so funny. It's just truly the irony of being anti ad and then just spending. How much do you think they're running? They have four individual ads. You can imagine them. This will probably be one of the biggest buys of the Super Bowl.

8:26

Someone ran the numbers, and if they did all. If they aired all four, it would be something like $80 million or something. I don't think that they're going to run all four. That seems like a lot. Maybe there'll be some regional buys in there. I don't know. $80 million seems like a lot to spend. I mean, it's a big company. They raised tens of billions, so, you know, it's possible. But that feels like. That feels aggressive.

8:41

There's also bulk discounting.

9:03

Sure, sure.

9:04

And there's again, they could do. They could choose to do regional.

9:05

And isn't there something where if you buy a massive ad campaign like this, you'll also have to buy in the Olympics as well? NBC sells both, so we could see a rerun of these or another buy later.

9:10

Yeah, super bowl has an insane amount of demand. Olympics has way less.

9:22

But this is clearly like a particular.

9:26

Moment in time, and it's insane. I mean, it's incredibly clever. It's also incredibly dirty. I think they're like Trying to muddy the water around the ads rollout.

9:28

Oh, sure, sure.

9:41

As the ads, the ads that are coming to ChatGPT are effectively display ads. Right. Everybody in the industry by now should know this. Yet this campaign implies that the influence the response and that you cannot trust it.

9:42

Yeah, yeah.

9:58

And so it's fair game. But it's like dirty.

9:58

Yeah, it's sort of fake newsy. A little bit. It's a little bit. It's a little spin on top of what will wind up happening. And yeah, I mean, it wouldn't even make sense for the ads to really influence the content. They're probably just going to wind up doing what Instagram does and just showing you things that you're actually likely to purchase, no matter what you're looking at. Because that will just be what optimizes CAC and roas.

10:04

Yeah. I mean, at some point, I expect that the ads will be certainly targeted. Like if you search, I don't know, Yerba mate.

10:28

Yeah.

10:40

What yerba mate should I get? Yeah, it might show you, like, what it thinks you should get. And then separately it will have, like, here's also another ad where. Yeah, where.

10:40

Where you see a sponsored, you know, result and it's flagged and there's a little ad ad tag or slightly different background color and people are used to that. So I don't think anyone really expects, you know, ads as they roll out in Chatbot, to be some major violation of the social contract of what people understand ads to be. But what's interesting is that they don't actually say OpenAI. They don't say chatgpt. This is just what comes with the territory when you're the dominant player. Someone can just take a shot at the category and it feels like it's a shot at you. And it's sort of a champagne problem. You know, if. If ChatGPT wasn't in the position that it was in, they wouldn't. Everyone wouldn't be like, oh, this is a shot at ChatGPT. They'd be like, oh, this is a shot at the category. Because there's seven players that are all equally used.

10:49

It's like, no, let's pull up the next one.

11:35

It's a dominant app. Yeah, let's pull up the next one.

11:37

How do I communicate better with my mom?

11:45

They went off with this great question. Improved communication with your mom can bring you closer. Here are some techniques you can try. Start by listening.

11:47

Really hear what she's trying to say underneath her words. Build conversation from points of agreement.

11:59

Find a Connection through shared activity.

12:05

Perhaps a nature walk.

12:08

Or if the relationship can't be fixed, find emotional connection with other older women on Golden Encounters, the mature dating site that connects. Sensitive.

12:10

This feels so off brand.

12:20

Cougars for anthropic.

12:22

Anthropic.

12:24

Yeah. They haven't.

12:24

It's not classic. It's not classy.

12:26

Mm, yeah.

12:28

Like it's. It's. I'm not saying. I'm not saying it wasn't a good feels.

12:30

It feels appropriate for the Super Bowl. This feels like the level of humor that you would see in super bowl ads. Like if this was a Bud Light commercial, I'd be like, okay. But yes, I agree with you. It feels like totally unexplored territory for them. Yeah, it is a little classic.

12:34

Vulgar.

12:47

It's spicy.

12:47

It's. Yeah, maybe.

12:48

All right, pull up the next one.

12:50

Pull up Okta. Okta helps you assign every AI agent a trusted identity. So you get the power of AI without the risk. Secure every agent. Secure any agent.

12:50

Pull up the next one, please. Deception.

13:01

So what do you think? Absolutely.

13:10

That's such a fun and creative business idea.

13:15

You've got something really special here. Getting started can feel daunting. I can make a step by step mini business plan. Do you want me to do that?

13:17

Yeah.

13:25

Absolutely. Here's some steps you can take. One, research. Really get to know your audience. Two, think of a catchy name and start your social presence early. Three, new businesses often struggle with cash flow. So try Quick Dash payday loans because girl bosses need CEO money quick. Whoa. 400% APR rates may vary, possibly doubling or tripling without notice. What would you like to make a quick credit check.

13:28

Incredibly well played and incredibly dirty.

14:00

Yeah. Yeah.

14:03

Like this.

14:04

It's the worst possible outcome in an ads world. But so easily avoidable. And OpenAI has been messaging. I was trying to find ages. And they've stated this multiple times on podcasts and in blog posts and in essays. Like they've been beating the drum on this for a long time. That clearly they will gate who is allowed to advertise, what the context is, how these ads will be displayed. There's a million.

14:06

I was trying to find examples of these type of super bowl attack ads from the history of Pepsi and Coca Cola, Apple and IBM.

14:32

Yeah, yeah. I'm a Mac guy. Yeah.

14:42

And none of them were as direct or effectively as the 1984 ad was saying.

14:46

Like IBM is authoritarian, dictator.

14:53

It's like so aggressive but still less. It wasn't as like. The timing here is a big factor. Right. Right when the Ads are rolling out and just like really muddying the water. I don't know. Obviously OpenAI is going to do. Is gonna do fine, but this really makes their life a lot harder. Yeah, that one was also funny because, like, she's like, absolutely. And that's what, like, Claude usually says. You're absolutely right. That's like a claudism.

14:56

That's a claudism.

15:30

Yeah. That's not chatgpt.

15:31

What does chatgpt usually say?

15:32

I don't know if there's like specific phrases exactly like that.

15:35

Well, the specific phrase that I think of is like, it's not this, it's that. And I didn't hear that coming through, which is interesting. But I don't know. Do you think these are going to be effective, do you think?

15:38

Well, that's the question. So really bold campaign for a company that hasn't been able to consistently crack the top 25 of the iPhone in consumer. And so my question is like, does this mean they're going into consumer, or is this purely to piss off OpenAI and make their life harder?

15:48

Yeah, like there's a world where there's business leaders.

16:06

Like, is this like. Like somebody's walking by and you just stick your foot out and try to, like, trip them? That's kind of what it feels like.

16:10

It does feel like vibe.

16:16

It feels like consumers so far gone.

16:17

Yeah. I mean, we'll see. We'll see. Maybe at the end of the super bowl, we'll. We'll check the app store charts and Clyde will be up at the top. Because so many people have seen this, they thought it's funny. But it does feel like such inside baseball. Like, I bet a lot of that, a lot of average consumer consumers don't even know that ads are coming.

16:20

They're not even. They're not even pushing. They're not actually trying to push downloads with. Or they would put a call to action.

16:36

Yeah. Or QR code like, download the app now. Ad free. AI is here with Claude. Like, download this thing. Interesting. I mean, maybe the thing is that.

16:42

I don't even know all this to IPO first. I can see it being beneficial in just terms of preparation for the ipo.

16:52

Right.

17:02

There's a lot of people that are just retail investors that aren't aware that. That aren't really super aware of Claude. They might have heard of anthropic. They actually aren't really aware of their different products.

17:02

I just think if you're an enterprise that's deploying AI APIs and LLMs into your organization. You're not worried about the ads product in ChatGPT. You're like, oh yeah, OpenAI Codex is this quality. Gemini 3 is good for this. And Claude 4.5 is good for this. And my team will deploy and we'll look at costs and Pareto curves. It doesn't feel like any CIO or CTO of a big company is going to say like, oh, I couldn't possibly use OpenAI in a business context.

17:14

Gabe says ad free AI is here with Claude. Ask three questions before hitting limits.

17:52

Yeah, yeah.

17:58

Or use add GPT and be able to ask a bunch more.

17:58

Interesting. Here's a Super bowl ad. They should run mongodb. Choose a database built for flexibility and scale with best in class embedding models and re rankers. MongoDB has what you need to build. What's next? That would just be a clean ad read. Clean ad read for.

18:01

Does this ever come back to bite them? Because the idea that you're going to offer products to consumers and never monetize transactions, never monetize commerce, never run ads like it hasn't really been done in the history of the Internet. Even Spotify, right. Netflix, they always overlook at Apple. Right.

18:16

If the Claude app does go to the top of the App Store, if they do wind up eclipsing ChatGPT and become the most dominant force.

18:34

No, it is really funny. If you drive a bunch of people to an app that has pretty low usage limits, they try it for a little bit and they're like, they're not going to notice that the model might be better. They're just going to be like, I'm going back to ChatGPT.

18:41

That's funny. That's funny. I want to see what the battle between Mac, Apple and Windows looked like. This is maybe back in 2006, 2009, the I'm a Mac, I'm a PC campaign. It's a 10 minute compilation video, but we can just watch like the first one, it's on, it's on YouTube.

18:56

Hello, I'm a Mac and I'm a.

19:14

PC.

19:17

Because they get viruses.

19:19

Zoom tight.

19:20

You okay? No, I'm not okay because PC's got viruses. Macs didn't.

19:21

There it goes. Oh, yeah.

19:25

In fact, you better, you better stay back.

19:27

This one's a doozy.

19:28

That's okay. I'll be fine.

19:29

No, no.

19:30

Do not be a hero. Last year, there are 114,000 known viruses for PCs.

19:31

PCs, not Macs.

19:35

So just grab this.

19:37

I think I got a crash.

19:39

Hey, if you feel like that'll help.

19:40

Good. Is this that far?

19:42

Hello, I'm a Mac.

19:44

And I'm a PC.

19:45

And I'm a PC, too.

19:46

And I. What?

19:47

Yeah. See, now you can run Mac OS.

19:48

X or Windows on a Mac.

19:50

So in a way, I'm kind of like the only computer you'll ever need. Oh, touche.

19:52

I don't think you're using that right.

19:57

Touche. No, listen, you can only say touche.

19:58

If you make a point, that I make a counterpoint. You see?

20:01

So I said I run Windows, but you haven't made a point yet. Let's try it again.

20:03

You can get a Mac and still run all your Windows stuff.

20:07

Touche. Hello, I'm a Mac. Okay. What do you think? I mean, they're not taking a shot at Windows specifically or Microsoft or Dell. It's like all. It's the whole category, but everyone knows it's really, you know, Apple versus Microsoft. Talking about viruses, talking about this other stuff. These are features.

20:11

I don't know.

20:29

It's not that far off. It's certainly not edgy. It's certainly not edgy. Like, it definitely fits.

20:30

Yeah. So anthropic's edgy. It's also a timing thing where, again, they're actively trying to make their. And again, I'm not saying anything of. This is not fair play technically, but the gloves are off.

20:37

Yeah, it does feel like that.

20:53

And there is zero. I will go out now and say. I would say there's effectively a 0% chance that OpenAI can win the super bowl this year. They will be running an ad. It would be insane if they didn't run an ad. It'd be a huge ad.

20:54

Just goes way harder. Sam calling out Dario by name directly. Just like, Dario, you suck. Just a diss track.

21:08

Yeah.

21:20

I mean, yeah. Yeah. Just a bunch of sora slop.

21:21

Yeah.

21:25

Just going crazy.

21:25

Yeah.

21:26

No, I think, like, the other thing is. OpenAI. I really doubt can react to this now.

21:26

Oh, yeah, yeah.

21:34

Like, the.

21:35

Right.

21:35

Like, if this had dropped, like, a couple weeks ago, they would have been able to say, hey, we need to respond to this. But they can't now. At least. At least our experience buying an ad this year, we were. The content was locked last week, and there's a whole process. Right. They're subtitling all the stuff. Like, they're not. You can't just switch it up. Maybe OpenAI.

21:36

Well, maybe they win in the sense that, like, the vast majority of audiences have the reaction that you're having, which is like, actually, I don't Mind ads. And actually, this is sort of a weird ad. And I still like ChatGPT. And then ChatGPT comes out with an ad that's just like, very vanilla. Like, hey, use it. We'll help you cook. We have health now. You know, we have a bunch of features that seems helpful, like, no. Then they wind up winning.

21:57

Has the right word.

22:19

What is it?

22:20

He says it's propaganda, imo. It really. It really is. It's not.

22:20

Yeah.

22:26

They're not being. They're. They're, again, they're just kind of calling out the category, being general about it, but they're implying that ads are going to.

22:27

It's fear mongering. Yeah, it's fear mongering that the AI that you're used to being truthful and accurate and not trying to sell you some slop product.

22:34

Well, they took, like, Mark Cuban's, like, main.

22:43

Totally, totally.

22:46

The things that he had been like.

22:47

Yeah.

22:48

Going on and on and on about that. Everyone is like, hey, dude, you can calm down. Like, that's not how it's going to work.

22:49

Yeah, yeah.

22:53

And then they just, like, went with his point of view.

22:54

Yeah, yeah. Which is a little out of date. I'm going to tell you about Shopify. Shopify is the commerce platform that grows your business and lets you sell in seconds online, in store, on mobile, on social, on marketplaces, and now with agents. And then I want to go to Tyler.

22:56

Yeah. I was going to say, like, it feels like the vibe war is, like, really heating up.

23:07

Right.

23:10

Like, over the summer, there was talent war.

23:10

Yep.

23:12

It's kind of a cold war.

23:12

Right.

23:14

It's kind of like, okay, we got your guy.

23:14

But on Twitter, and it was all leaks behind the scenes. Like, Mark Zuckerberg didn't even really give an interview during that whole time.

23:16

Yeah.

23:21

And. And there was like, that leaked memo from Mark Chen saying, like, I feel like something has been stolen from us.

23:21

And I was saying they were not.

23:26

Coming out and saying. Yeah.

23:28

At Davos, like, Dario is like, basically.

23:29

Yeah.

23:31

He's still not saying OpenAI the words, I think.

23:31

Yep.

23:33

But he's saying, like, companies are doing ads. There's only one company that's doing ads.

23:34

Yeah, yeah, yeah, yeah. I mean, it's sort of like the anti. There's these. Ben Thompson has a great formulation of, like, the anti YouTube alliance between, like, Spotify and Netflix. So, like, once one company starts pulling away in a category, all the other companies sort of have an incentive to team up. And right now it does feel like we're getting A little bit of an anti OpenAI alliance because everyone's like, well, they pulled away and so, like, let's gang up on them. And they, you know, they did so much stuff on the supply side. They locked up all these different. All of these different supply chain partners. Like, we got to push back. We actually got to team up. And that's where you see at Davos with the buddy cop movie emerging between Daria and Demis. We will see.

23:38

Katie, the CMO of OpenAI fired back. She says ChatGPT has more free users in Texas than ODD has globally.

24:19

Boom.

24:28

Yeah, nothing like stats. They should run that in the super bowl potentially. Let's watch some of these other super bowl ads because I want to see this General Motors robot. What happened here?

24:30

This was. How'd you find this? Fourteen years ago, I was digging around.

24:41

Is this for the worst super bowl ads or the.

24:45

I would say this is the worst ad I've ever seen.

24:47

Okay, let's play. Let's play General Motors robot.

24:49

And they, they, they've tried to scrub this from the web, but we're bringing it back.

24:52

Let's play it. Let's.

24:56

Robot makes a mistake.

25:04

General Motors make cars. We make Cadillacs. Okay. Robot makes a mistake, leaves the factory.

25:05

That's fired.

25:15

Whoever produced this ad is not gonna do well in singularity. You did not consider Rocco's vasculous. Oh, sign holding Tadillac going by.

25:17

Well, now, now the robot's trying to get a new job.

25:31

Okay, Robot's working at McDonald's or something. Robot goes to the bridge.

25:34

This is like, who approved this?

25:41

Robot watches the Chevys drive by the General Motors cars and jumps into the water and then wakes up the GM 100,000 mile warranty. It's got everyone at GM obsessed.

25:43

Like, how do you run this ad?

25:56

Whoa. So it's saying, like, we care so much about the hundred thousand mile warranty that like, we will end it all if it doesn't. If we don't stand by it.

25:58

I think so.

26:06

I'm still.

26:07

Obviously, they got an insane amount of pushback from people saying, like, hey, you're effectively advertising, like the darkest thing ever.

26:09

No, really sad. Like, ridiculous. Well, the bar has been lowered, so don't worry. Let me tell you about Lambda Lambda is the super intelligence cloud building AI supercomputers for training and inference that scale from 1 to GPU to hundreds of thousands.

26:16

One more ad we'll play for you. This is a Bud Light commercial.

26:31

Okay? I like Bud Light commercials. They usually deliver perfectly, flawlessly. I feel like Bud Light is very, very steady. You can always count on them for pretty solid super bowl entertainment. My king. This corn syrup was just delivered.

26:34

That's not ours.

26:51

We don't brew Bud Light with corn syrup. Miller Light uses corn syrup.

26:51

Let us take it to them by name.

26:56

By name.

26:57

Okay.

26:58

Because that's in the ingredients.

26:58

Yeah.

27:00

So they can say it.

27:00

Yeah.

27:01

Pause. Pause for a second.

27:02

Yeah.

27:03

So. So if Claude had named Chat GBT and done this, that ChatGPT could have said, like, we are suing you for defamation. Defamation.

27:04

Yeah, yeah. Right. Yeah.

27:12

And potentially.

27:14

Sure, sure. Otherwise, you just got to go with the category broadly. Anyway, let's. Let's keep playing syrups.

27:14

We received your corn syrup syrup by mistake.

27:22

That's not our corn syrup.

27:25

We received our shipment this morning.

27:26

You're joking.

27:28

Try the Coors Light Castle.

27:29

They also use corn syrup. See, this is still, like, a little bit more playful and. And, like, classy. It doesn't go to, like, an edgy place.

27:30

It's also not misleading because you can see on the ingredient list.

27:43

Yes.

27:46

You can just look up and see they have corn syrup.

27:47

Yeah, yeah.

27:49

Like, my, my, my.

27:50

Yeah. It'd be very different if it was like, we're delivering your ads to the chatgpt Castle, and it's like, oh, we already got our shipment of ads. Like, take that over to Google. AI Search overviews. You know, like, take it over to Grof. They should have just done. They should have done a direct rip of that.

27:52

Yeah, no, I mean, incredibly, incredibly well played by Anthropic.

28:06

Yeah. So you think it'll be effective?

28:10

No, I mean.

28:12

I mean, to what degree will it be effective? That's my question.

28:13

Well, that's. I'm not even sure they care about it being effective.

28:15

Okay. It'll be. This is effective. Like, the fact that people are talking about it. People that right now on the timeline, people are dunking and being like, oh, good point from Anthropic.

28:17

Yeah.

28:26

Like, they're scoring points today.

28:26

Teapot Max. And they're able to score some points in the real world and just make like, senators are going to see this. Yeah, that's wrong.

28:28

Yeah.

28:38

Yeah. I mean, the whole, like, senator, we sell ads thing in Facebook, like, stealing your data. That whole thing is, like, very, like. It's still. It's still a meme.

28:38

It's.

28:46

It's still a thing in the general populace, which is unfortunate, because advertising is the greatest business model ever. And companies don't want your data. They want conversions. They want you to purchase. They want a black Box where they can put money and then get customers and get money out. That's it. I've run businesses that advertise many times and I don't want to know anything about these customers. I just want to know they're ready to buy and they're down and send them the link.

28:47

Yeah, the difference went between like a Bud Light going after some of their competitors is that they're actual competitors.

29:13

Yeah, no, no. It's a very, very pure oligopoly. But maybe, I mean doesn't that lend itself to like anthropic is punching up because they're coming from behind because they're a smaller company, they have less market share. In this particular market it would be like, I don't know, it'd be like if the Athletic Brewing guys ran an ad that was defaming all of the beer companies for all the bad things that beer can do to you. Right. Something I don't know. Anyway, Railway. Railway is the all in one intelligent cloud provider. Use your favorite agents to deploy web apps, servers, databases and more. While Railway takes care of scaling, monitoring and security.

29:22

Yeah, according to app figures, Claude got around a million downloads in February which was actually down for from the month prior.

30:01

Interesting.

30:12

So not a real player in consumer.

30:12

Yeah, we'll see what happens. I am going to be watching the app charts like a hawk. I want to know.

30:15

Yeah, the entire game. While we're at the game we're just going to be glued. We're just going to be in the back just probably. Anyways, I'm so excited for all the other tech ads in the super bowl. We got Mr. Beast and Salesforce. We'll have a bunch more. That'll be good and I'm looking forward to it.

30:20

Anyway, we got a our guests here yet? We have guests joining in just a few minutes but in the meantime we got to bring down the mallet from the heavens because we got to ring the gong for Walmart. They reached $1 trillion in market cap as E commerce booms. Let's bring it down the Lambda cloud.

30:41

Hit it.

31:05

Let's understand what's going on with you. What's going on with Walmart? They've been written off. Amazon was going to kill them. But seems like the retail behemoth is growing faster than ever. Let's see the achievement places Walmart among a small but growing club of companies that have a 13 figure valuation. Amazon, Nvidia, Meta, Microsoft. In trading Walmart stock passed $125 a share. Shares closed up 2.9% giving the company a market cap of 1.018 trillion. The stock has surged in recent months, fueled in part by Wall Street's enthusiasm for the growth of the company's online business, as well as investments in automation and AI technology aimed at improving efficiency. Sales have also ballooned as more shoppers have turned to Walmart for low prices, faster delivery and broad selection. The change at Walmart over the past decade, culminating with its trillion dollar valuation quote, has been seen as a profound shift at a retail company that we've ever seen, said a retail analyst at Morgan Stanley who studied Walmart since 2001. Wow. Simeon Gutman overnight success here 25 years studying Walmart Walmart's growth along with Amazon's creates challenges for competitors, he said. Meanwhile, Bentonville, Arkansas based company will have to navigate the ascension of a new chief as longtime executive John Furner took the helm this month. Most of the 11 companies that at any point reached 1 trillion are technology focused, even though Berkshire Hathaway is in there and Eli Lilly is in there as well. A decade ago, some Walmart investors didn't see the company's success as a sure thing. Rival Amazon was growing fast and the then new Walmart CEO Doug McMillan was was was investing billions to raise worker pay, clean up stores and grow online. Investors wanted to see whether the investments would pay off. Walmart's market value was $212 billion at the end of 2016. Warren Buffett's Berkshire Hathaway famously sold a large portion of its longtime stake that year and fully exited the position by 2018. Retail is changing so much, I don't think I understand it as well as I need to, buffett said during a call with the CEO about the Berkshire stock sale. He said, I don't understand the change. What will E Commerce do to your business? Buffett offered kind words about Walmart's potential to succeed, but the snub was motivating for company executives at the time. Walmart sales have since soared, propelled by E Commerce. Then the pandemic followed shop followed by shoppers More recent hunt for lower prices amid inflation Companies company's investments played a big part. Walmart made an effort to offer more items that appeal to higher income shoppers, such as trendy small appliances and store brand foods. The company accelerated home delivery capabilities and can now deliver orders the same day to 95% of U.S. households. So they've like fully responded to Amazon prime, which was the main differentiator for a long time.

31:11

Yeah. Yousef in the LinkedIn chat says Walmart is just super secretive about its capabilities. He previously was over at Walmart.

34:01

Thanks for leaking it to us.

34:09

Sucks on X says nobody talks about Walmart much. Age of tech giants. But if Sam Walton's fortune hadn't been split up, it would still be a fair bit greater than even Elon's and they're still the largest.

34:10

That was true back then. He posted this six months ago. I think Elon's way above this now, but it is interesting how successful.

34:19

But yeah, Jim Walton, Rob and Alice all over in the $100 billion.

34:25

I mean a lot of this growth has to be because of the rebrand. We got to talk about the Walmart rebrand. You've seen this. Let's pull up the graphic and see. Whoa, that's probably driving.

34:30

Stunning. Stunning and brave.

34:41

That probably adds 600 billion to the market cap, right?

34:42

I mean, easily, easily.

34:44

They really just did a slightly bluer background and they called it a day.

34:46

Well, without further ado, Figma.

34:53

Figma make isn't your average vibe coding tool. It lives in Figma so outputs look good, feel real and stay connected to how teams build, create code back prototypes and apps fast. And that's right, we have mati from 11 lives, the man of the hour. How are you doing?

34:54

What's going on?

35:08

John, great to see you again.

35:10

You look fantastic. This cover is iconic. I mean props to you, but also the photographer on this like perfect lighting. I've always been a fan of Forbes photography.

35:12

It's so nice to see somebody just beaming too, just having a good time. Usually people, people are on the COVID acting all serious, mysterious, you know, I like to see, I like to see.

35:22

We are having a good time at 11 laps.

35:32

It's a good, good, good, good.

35:34

And building at the current age and everything is changing with AI. We have a chance of building at the frontier of that change. But a little bit surreal to see the COVID in a zoomed in picture. So that's always a tip of the iceberg and reflection of the entire team. So great, great to see it.

35:35

Yeah, it's amazing. How did the negotiation go with this current round? I'm sure you had a sort of your interest was coming in at 11. I'm sure that gave some leverage.

35:52

Couldn't do 10.9.

36:04

Yeah.

36:06

This was in this round. The valuation was known far before the round even started with our aspirations from the early days to get that. And of course just surreal to be able to announce today that we've raised 500 million Series D at $11 billion valuation.

36:08

Congratulations.

36:26

Thank you guys, the next leg up will be tough getting to 11 trillion. But we're, we have it all planned.

36:28

It's going to be 11 squared next 121 billion as the next proxy of valuation. But this funding is great. We think this is what we need to get and transform how we interact with technology. Sequoia is leading the round. Andrew is joining the board. Asics TV is quadrupling down. Iconic is tripling down and alongside the round. As always, we are releasing incredible updates to our voice agents to make them more expressive, quicker, lower latency so people companies can build even better experiences and customer experience, internal enablement and sales.

36:36

So yeah, we were just, we were just laughing. We were just laughing. We were watching some of the new super bowl ads from Anthropic and they intentionally put a bit of lag in there when in the model and that's your opportunity and that's the opportunity. It's like just reducing that latency to. That's really fun or even superhuman.

37:07

Yeah. What has the.

37:26

A hundred percent.

37:28

What's been the biggest driver of growth? Just more growth on the core products, new products. Walk me through sort of how I would visualize the growth chart that went in the lead slide of this deck.

37:28

So we started in 2022. The first model was text to speech. And then we continued expanding with incredible researchers in the team. Still doing speech to text with the best transcription model 11 music with the highest quality music model productions with dubbing. But the real innovation that we brought in 2025 was agents and conversational models. So helping enterprises build that customer experience work, internal enablement. And in 2025 we got to 330 million in ARR at the end of the year. Powered by work with Deutsche Telekom. Amazing. Deutsche Telekom customer support Square Deliveral with training riders and importing them into the ecosystem better. Or calling restaurants to capture what's happening with the menu, what's happening with the opening hours all the way through to Revolut. Working with us across 4 million of their users in Europe to be able to bring that level of customer care that frequently wasn't possible. And we are seeing that. I think something that was impossible just a year prior. You have a real time voice agent interactions. Jordi, you were mentioning on Tropic. We think all companies in the future will want to optimize the latency, will want to optimize that reliability so they can interact with their audience, with their customers and whether that's in customer care, whether it has devices around us, hopefully some of the devices you have will be able to speak back to you and converse with you as the guests come in all the way through to the future robots. And that enterprise adoption over last year has been tremendous. We grew to 330 million and it took us 20 months to get to first hundred million in IR, 10 months to get to 205 months to get to 330. So hopefully that continues and it's a good proxy of the value that we can, we can deliver.

37:43

Incredible.

39:31

Well, we have a special guest. Yeah, we should bring him in.

39:32

Let's bring in Andrew Reed from Sequoia. New board member.

39:34

New board member. Alert.

39:37

How you doing?

39:38

What's going on?

39:38

Hey, guys.

39:40

Hey. Good to see you.

39:40

Good seeing you.

39:42

I was enjoying when you were bringing on Mati and I've realized that's very important for a Polish entrepreneur to become a one name entrepreneur because no one wants to go for the last name and we're not brave enough to do it either.

39:44

Yeah, much easier when you're true. So talk to us about the surface area of the business. Now, there was someone in the chat saying that they were just using 11 reader last night. Is that more of like a demo and a showcase of the product or do you see that as growing into its own sort of consumer AI application?

39:59

Yeah. And first of all, great to see you all here. Andrew, welcome to the board. Welcome to eleven Labs. After two and a half years, knowing each other is going to be a great fun after serving on Klowna Board together.

40:23

Oh, yeah.

40:32

Pleasure to have you with us. And it was the last time we.

40:33

Saw you guys, which is great.

40:36

So it's a full circle moment. But to your question, as you think about our business, we have the foundational research layer where we do everything across audio, whether it's text to speech, speech to text, whether that's the expressive narration, expressive voiceovers, the music work. And on top of that work, we have the agents work that we spoke about, but also the creative use cases. And really there is a great example of that, of how we can let our community create the narrations, create audiobooks, and then serve them in an easier way. You know what's crazy is that today Audible will block AI audiobooks. The only platform that allows you to serve their content on Audible is only the Amazon created AI content. So we need to find an alternative for the community to serve that. And similarly, happily, I have a partnership with Spotify to allow our users to bring that. And we have a lot of users. It's 5 million users doing doing that work.

40:37

There you go. Andrew, how are you processing the SaaS apocalypse, the turmoil in the markets? Is this like, hey, 11 labs needs a war chest batting down the hatches rip good times, or is this something else? That's. You don't really need to worry about who's getting beat up in the public markets because you're the reason why they're getting beat up.

41:32

Well, I go back to what Jason Lemkin was saying. He was bringing up 11 labs as an example last week because he said he has a lot of portfolio companies and they're like, we're AI native. We're AI native. Like, show me the numbers.

41:53

Are you growing?

42:04

11 Labs is AI native, and they're adding $100 million in new ARR every few months. So, like, that's what it looks like.

42:04

That's.

42:11

That's what real pull from.

42:11

But yeah, Andrew, what do you see in the market?

42:14

Yeah, I think the, you know, like the scorpion and the frog, the voting machine will be a voting machine. And I think in the public markets with software, there is a lot of baby being being thrown out with the bathwater. And I think as far as I can tell, the only prerequisite for being an AI native company is being a private company. And I'll leave that there. I do think what really matters is our businesses accelerating as these models are getting better and are they delivering value in a way that's like, differentiated and unique? And we are seeing companies like eleven that are growing faster with better economics, better customer references, and just faster deployment than anything else we've ever seen. Like, I've going back to when I first met Mahdi 2023, I remember when I first looked at 11, it was not obvious like, what kind of company this is. Is it a consumer company? Is it an enterprise company? Is it a foundation model? Lab is an application company? And it turns out it was all of those things. And I think for like, these truly, like the true AI winners, it's, you know, like the phrase going from strength, strength to strength. It's like everywhere 11 has been competing, they have just been dominating. And like, shame on me for not investing a zillion dollars three years ago. Fortunately, we were lucky to invest a little bit and I got to know Marty. But yeah, the new wave of these AI companies are just incredible.

42:16

Can you both talk about a little. Take me through more of that focus? Because you're saying 11 is a research lab and an enterprise company, and there's consumer product, but there's still a focus on audio still focus on voice. And we've seen the, a lot of the good vibes that are coming around Anthropic is like they're focused on code. You know, we've seen this like midjourney focused on images and you get to this opinionated thing. Is that where you're narrowing or do you see yourself kind of expanding out on all surface areas over time?

43:57

Right. And it is true. Like, as we started a company, the one connected issue from model through platform to application is audio and voice. We know that we can create the best models, we know we can serve them, build for the users, understand those use cases, and then bring them to production. But of course, to get most of voice frequently, we'll stretch from other sister fields to bring those integrations into default. With agents example, you need the knowledge base to really integrate. You need telephony systems to make sure the agents interact in WhatsApp or can call the phone numbers. You need the evaluation, testing and monitoring to really be able to understand that. So as we think about this work, voice audio is our superpower. That's the one connective tissue across all those domains. But if we can bring from other fields and elevate the entire voice experience, whether it's in voice agents, whether it's in a voice a creative space, we'll happily do that.

44:31

Where are you frustrated with slow adoption? Where is there a lot of potential where companies are not. They haven't fully processed that the models are advanced to the level that they have. And you guys have this feature set on a personal level. As a user, there's still a bunch of different, like media subscriptions that I have that have just like really poor like audio functionality. It's like if you publish an article, you should immediately have like, you know, a mini podcast equivalent that I can listen to in real time. It shouldn't be like, this doesn't seem like a hard problem anymore. But I'm sure you're seeing that kind of in other categories too, where you want industries to wake up and say like, hey, voice is here. You should be implementing it immediately.

45:27

Yeah, I can go quick one first on things that we believe in and Andrew is helping us push us to think about that space. The one audio will be default and all the content stories, knowledge that's available, it will be global. You will have all the languages, all the voices represented, and voice will be the interface to the technology around us. The big opportunity to your question Jordy will be in the media entertainment space that we haven't seen on the agent side Today, a lot of the use cases are reactive, supportive, but we'll see that proactive side where you have an AI concierge, AI assistant that can help you for the entirety of that journey. And as we work with Andrew, we frequently think of what's that connective tissue between research and product, how we can combine that experience together to deliver that. Andrew, I'm curious what you think. We spoke a little bit about this, but what's top of all your mind for the next year?

46:13

Yeah, well, I think at a high level it's very hard to find markets where voice is not taking off. And I think I'm old enough to have lived through the initial wave of chatbot voice agent companies in the kind of 2015-2019 era. And for those companies it was enterprise vendors foisting audio upon consumers and consumers, you know, asking for a real agent over and over and over again. Like that was sort of, I think.

47:03

Talk to a human. Talk to a human.

47:33

Like the customer research that like created, you know, like voice agents was people asking for agents over and over and over again. I think the difference now is if you look across like basically every sector, people like interfacing with technology through voice and audio. Like it's a, it's. There's a lot of end customer pull. I do think there's a big gap between what people expect the voice agents to be able to do and how they communicate with them versus what they're actually capable of. I think humans are very adaptable and I think Once everyone used ChatGPT and realized what the chat bots can do, people start interacting with text boxes on the Internet very differently then with voiced agents. Still people think they're in a phone tree and are trying to talk to it like a phone tree. In reality, it's this magical interface that overlaps all of a company's capabilities and it allows you to express yourself and receive information back in a brand new communication style and a brand new interface. I think once people realize what these agents are capable of, my guess is the way people interact with their institutions, with their governments, with their, the companies that they buy services and goods from is going to change very quickly. And I do think it's going to take like the leading, the leading companies in this space to show people what's possible and the whole thing's going to flip.

47:37

Very cool.

48:58

And you know, on the last one, just to add a quick example, we've seen that we work with the government of Ukraine on exactly that. And one incredible case was they are creating an effective voice agent for the citizen support where you can call and ask about the services and what's transpiring is that the moment you can call in and you frequently just don't know where to find information. If you can just speak through and let the agent navigate through to the right help, it just open up. So many more programs that were just not available. Whether that was in the how do I travel, what's happening around the country, how do I apply for certain help? Just opens up a completely new ways of you exploring that information altogether.

48:59

Yeah, I'm excited about seeing some new UI patterns. I feel like the walkie talkie is potentially a better even comp than the phone call. Because if I'm using a product and I have a question, I don't really want to like place a phone call necessarily, even if it's instant and quick. And I know like an agent, you know, an AI agent will like pick effectively pick up the call immediately. You can imagine like hey, how like where. What's the best document for like this part to understand this part of the organization, it can like pull it up and it's like having. Having these kind of new UI patterns I think will be very cool.

49:37

Yeah. Well, congratulations on the round. Thank you so much for taking the time to to stop by and chat with us.

50:11

Great stuff. Looking forward to the next one.

50:17

Have a good day.

50:19

Thank you so much for having us. Congrats on the new super blood.

50:20

Thank you. Incredible. Very excited.

50:22

Congrats to you, to you both. Thank you for supporting us.

50:24

We appreciate it. Thank you for having us.

50:26

Just make sure that. Make sure the ad's available in Polish too.

50:29

We will.

50:32

We'll do a translation.

50:32

Help us out. Help us out.

50:34

Have a good one.

50:35

Great to see you.

50:36

Goodbye.

50:36

Sentry. Sentry shows developers where it's broken and helps them fix it fast. That's why 150,000 organizations use use it to keep their apps working. Let's go over to Gastown. What I wrote about in today's newsletter and I want to pull up this very simple to understand graphic. As soon as you see this Jordi, you will understand how Gastown works. I get it Said Robert says Gastown is the modern day Temple Os. You have to be on the spectrum to design something this insane and it does have a lot of layers of abstraction. I'll try and take you through it a little bit. Basically my theory is like there's a lot of excitement about this project. You might already be familiar with Gastown, but the broader category is called Orchestration how you orchestrate a whole bunch of different agents. And Steve Yegi wrote a great breakdown of his new Mad Max themed Orchestrator. He says it's a new take on the IDE. It's called Gastown. He dropped it on January 1st on New Year's Day. So he has been locked in. He says that he's like, like having trouble sleeping because he's so obsessed with this thing. He's really pushing a ton of code and now he's getting one call per day from VCs asking to invest apparently. And so it's basically a continuation of the developer experience. And he maps out how this evolved. So you used to write code in a text file, save it, execute it in a terminal. Then we got basic IDEs with some code completion, links to file systems replacement, etc. LLM chat windows work their way into the IDE eventually with a coding agent asking you for permission to run tools, run code. Once models got better, developers trusted them more and the IDE sort of melts away and you're basically just interacting with the agent. So Karpathy has put it, he says your code writing skill atrophies, but your code reading skill improves because you're prompting and then you're just reading the review and you're saying, okay, yeah, this is going to do what IT think it's going to do. We'll test it. And so the most popular workflow currently is probably a single agent cli. So Claude Code, Codex, Gemini CLI are the most popular. They all have web and desktop front ends now, but it's sort of too soon to tell how fast those will get adopted and we can move on from this slide. So Gastown is like way, way more aggressive going fall into Vibe coding. So Steve in his post just all caps says you will die if you don't know what you're doing. It's like very risky. The code base is only weeks old, maybe it's over a month now. It's 100% vibe coded. He's never seen the code. And it's 225,000 lines of go. Like not a very simple.

50:37

Yeah, he said you should only be using it if you're already at the stage where you're using ten plus agents and otherwise like you will not be able to use it.

53:22

Yeah, yeah, so, but so there's tens of thousands of people using this. Some folks have dozens of accounts with the big labs because they're maxing out their subscriptions. They're like, I got the 200amonth plan over here. I ran out of rate limits, so I got another one, I got another one. Some of them get flagged for fraud, like it is boom time in Gastown and bills run into thousands of dollars per month. And you're getting close to having basically a full time software engineer salary on the line. Despite all the warnings, Steve has created a delightful little metaphorical taxonomy for how to explain how things work. The town is your hq. This allows you to work on multiple projects. The projects are called rigs. And then you sort of play, I guess that's the word as the overseer, you're the boss. But you also have a mayor who reports to you, like a chief of staff. That's an agent that you talk to. And the mayor kicks off work convoys to different agents. They're called like polecats. These are ephemeral agents that go and do like one little thing and then they write code. That code lands in a merge queue. But then you have another role, another agent called a witness that oversees all the polecats to help them get unstuck. And then there's a deacon that goes around, patrols the town, finds stuff that needs to be taken out or deleted. There's dogs that do maintenance, like cleaning up code branches that have gone stale. There's a crew that are specific to a particular project and those are longer lived than polecats. So if you have back and forth design work, you'll create a member of the crew who he says you'll love and you'll develop a relationship with and you'll be updating their agent workflow and their skills so that days later you can go back to the same agent about how you're architecting the app and it has all the context, whereas the polecats are just off doing one little implementation at a time. And so there's a bunch of different roles. It's a lot. It's very cool. And it feels like a glimpse of what's coming this year. You still have to have an idea of what to build. So we talked to, I've searched around for like, what are people actually building with this? And a lot of people are like, well, I re implemented this open source library in Rust. It's like, that's probably valuable, but that's not exactly like the breakthrough $1 billion individual solo developer project just yet. You still gotta have a genius idea, but then you can build it. And you still have to have a really good ability to manage agents and understand when things are going off the rails. But the rough edges are getting sanded. Down as we speak. Orchestrators. It feels like these are the next. It's the next easy unhobbling that will cause another doubling in the meter software engineering time horizon benchmark. If you remember that benchmark for how long a software engineering task can run without going crazy. It used to be a couple minutes, then it became a couple hours. Like, if you set up your Gas town appropriately, you could potentially do weeks of software engineering work autonomously. Do you have more context on the meter eval? I mean, meter only LLMs.

53:29

Yeah. It's not the actual model running for like four hours.

56:30

Exactly.

56:33

It can do a task that takes people four hours.

56:33

Exactly, Exactly. Exactly. So now I think you won't just be able to slot Gastown into the meter eval, because Meter has like, Opus 4.5, Codex 5.2, Gemini 3 Pro, and so you can't really just slot it in. They might need a new category or something for orchestrators, but you could imagine getting a week's worth of software engineering work done with sort of a single prompt or a single setup.

56:36

Yeah, it feels very similar to you. People talk about their custom vim configurations. You spend like hours working on these.

56:59

Totally.

57:05

I've seen a ton of people on Twitter, like, build their own orchestrator. I mean, this is very complex and cool.

57:06

Totally. Yeah. No, no. A lot of people are already building these orchestrators themselves. Gastown is just one that's open source, but a lot of people are building these various harnesses. And so orchestration and delegation, they make sense. Even in a world where models are improving in capability and declining in cost, a stock LLM just cannot spin up multiple instances, instances of itself if you give it a huge task. So imagine you're just like, categorize every receipt in my company has ever processed. Like, an LLM can do it, but it's just going to take a long time. You should use 50 instances of LLMs and you have to write some code for that. An agent should be able to do that. That's what you're sort of getting with this. But there's obviously still a bunch of kinks to work out in 2026. It's a very interesting paradigm to me. It's also so sort of, in hindsight, obvious that we'd get here. And I think we'll see all the major AI labs do something in this space. I don't know if that means launching new products, but they need to deliver on this experience in a more polished package soon. I also expect a number of startups to try and own the category or carve out sub niches. You know, Gastown4legal or something like that will probably be something we hear a pitch for at some point. And in the end the multi agent experience might be completely abstracted to certain end users. Like most people using chat apps don't care about the details of a mixture of experts model. They just want a good answer and they want the model to be good at math and poetry and writing and research and history and all these things. And MOE models succeed at that, but it's buried below the fold. And so even auto routing between different models, I don't think a lot of people care about that if they're just in the consumer world. And I think that this will be more and more abstracted. So all this will make 2026 more exciting than ever. You have anthropic and OpenAI fighting over the vibe wars at the Super Bowl. You got Elon and Sam under oath in an Oakland courthouse and another thousand startups just got funded to make something people want.

57:11

And so basically if you're working on AI agents and you just pivoted to building a harness, pivot to orchestration. Orchestration.

59:10

It does feel like it might be like the next hot keyword that we're seeing. Orchestration Market map. Orchestration also sorts of stuff. Anyway, Vanta Automate Compliance and Security. Vanta is the leading AI trust management platform. And yeah, people are having fun with Gastown. Will Brown got nerd sniped into finally reading the original Gastown post. And wow, it's beautiful and terrifying and hilarious and probably a glimpse of the future that will feel normal in six months. I couldn't agree more. Steve Yegi is a great writer, just very, very, very interesting to read all of what he writes. And he talks about his neighbor's squirrel in a very fun way and how chubby the squirrel.

59:17

Wait, his neighbor's squirrel?

59:56

His ad. He has an 82 year old neighbor who feeds his local squirrel against the city ordinance.

59:57

But.

1:00:04

But he's 82, so he's like what are you gonna do to me? And then he uses this like fat squirrel as an analogy for the Gas town that he's building. It's fascinating. And then people are having fun with this. Tetsuo says, last month I was generating 15,000 lines of codes per day with clog code. Once I discovered the Ralph Wiggum loop, my productivity shot up to 10x. This week I finally set up Gastown and I'm generating 1 million lines of code per day. At this rate, I'm mere days Away from completing my Minesweeper clone. And like, there is something here where people, you know, you got to know what to build. You got to have a good idea.

1:00:05

There's going to be a lot of people spending 15 grand to recreate a game that's 60 bucks.

1:00:40

Totally, totally. But if it has me in the game, maybe I'll play it. Who knows? Anyway, Gemini 3 Pro, Google's most intelligent model yet. State of the art reasoning, next level vibe coding and deep multimodal understanding. And without further ado, our first in person guest of the day, Gerge. Correct. Did I say that correctly? The Pragmatic Engineer. Welcome to the show. How you doing? Grab a seat, please. Sit down. And first time on the show. Please give us an introduction.

1:00:45

Yeah, good, good to see you in person. In person. This place is even, even better than on the show.

1:01:12

Yeah.

1:01:17

Yeah. So just flew in from Amsterdam and I got you guys some, some special stuff. This is stuff that they don't know about the Netherlands. It's called Crowd Noten.

1:01:18

Okay.

1:01:26

And it's best kept secret in Netherlands. Before Christmas. Little mini cookies, mini cinnamony cookies. Before Christmas. The whole city sells this all the time. But as per the Dutch, you can only have it before Christmas. So there's this one shop that sells it year round. Locals hate it. It's kind of unknown, but I think it'll be all right. And yeah, just some Pragmatic Engineer stickers for you guys.

1:01:26

Thank you. And how do you introduce the Pragmatic Engineer these days? Purely Substack Writer, a media company. What do you like?

1:01:47

I like to say that I'm a software engineer who has been a software engineer for like 15 or so years. I was a manager as well and then I just started writing for software engineers. I didn't think there would be any demand for this. I don't think anyone thought there would be any demand for this. And turns out there is. There was a ton of demand for people writing about just like for me, writing about in depth topics that I don't write for beginners, I write for experienced folks. Talking about, I started talking about stuff like how did Uber create platform teams, like, you know, tech depth, how to deal with that stuff. And people were paying attention, subscribing, paying. So it was. Surprised me the most.

1:01:55

That's amazing.

1:02:32

Well, and I feel like timing, timing wise, what better time to be writing about software engineering than right now?

1:02:32

It's a crazy time. Like I, yeah, I've been in the industry for about, you know, like 20ish years depending on how you count it. And I'm now luckily been able to talk with people who've been, who are still alive, like Grady Booch, who is a legend who had been around since closer to the beginning of it. And I don't think there's ever been a time like I remember when I was working, you know, we were a software engineer. Like we loved our jobs and I think we still do, but what we were doing is kind of automating other people's jobs like customer support. We were saying how many savings we had and we kind of took this for granted. And there's a bit of existential crisis now with devs because this is the first time in history where the stuff that we build could potentially automate our work.

1:02:39

Maybe more. Software engineers used to have the luxury of like, I'm not going to, I'm just going to do my work. I don't need to really pay attention to what's going on and the big changes that are happening. And now it's like, oh, I should really pay attention. Just so I stay on the edge. Because if you're not getting efficiency gains right now, I think you will be left behind. That feels obvious.

1:03:23

How wide is the spread of experiences that folks in your audience share with you? Because I imagine that there are some folks out there saying like, like I'm using Gastown and agents and I'm vibe coding, I'm doing everything and my job has completely changed. And then I imagine that there's some people that are like, it's kind of business as usual at my company.

1:03:43

So interesting enough you would think that. So first of all, like the people who are using the agents and doing all this cool stuff and being on the cutting edge and you guys shut off a gas out on Stevie. I was just hanging out with him yesterday.

1:04:02

No way.

1:04:13

Yeah, we were together at this place called the Future of Software engineering summit. So 25 years ago there was the Agile Manifesto that back in 2001 it kind of shook the industry because it said like talk to customers iterate fast. It was like four simple things.

1:04:13

Anti waterfall.

1:04:27

Anti waterfall. And there were a bunch of people who gathered at this resort in Utah and they came up with it. And 25 later, Martin Fowler, one of the Adjoin Maniflosta founders who organized a retreat with a bunch of thinkers of today. And we went there and we gathered and about half the people were more traditional companies. Think about like John Deere, 200-year-old companies, enterprises, Cisco, et cetera, and then some of them for startups and the crazy thing was that like I was thinking, you know, there were businesses, you'll know, like this is the first time I'm seeing a technology where even in the kind of most old school companies, they're using it, they're trying it, they know that they need to catch up. And funny enough, some of those old school companies are a little bit ahead of like, I don't know, some mid level tech companies because they kind of have processes to like approve vendors and all that they've been used to. So like, I don't see anyone who's, who's not everyone's changed, impacted by it. But there are some people who are on the cutting edge and you know, using Gast Town and playing with Claude, et cetera. I think they're still the minority, but the gap used to be massive. It used to be like years or something, even a decade behind Steve.

1:04:28

Steve, you put him in the same kind of camp as Peter from openclaw and that like just loves the craft and just loves experimenting. And it's like non commercial because it's just like we have this interesting dynamic right now where you have these hyper, hyper, hyper commercial labs, like anthropics, which is idealistic but very focused on let's make as much money as we possibly can from generating code. And then you have the Peters of the world and maybe the Steves of the world that are just in it for the love of the game.

1:05:34

I was hanging out with Peter two weeks ago, right when Claude was starting to blow up. And then my podcast went out when it was at peak popularity and people thought I was talking with him that same day. But I would say Peter is way more obsessed than Steve. Like when I met him for two weeks, I was the first human that he met. That guy is pulling it, but he's very clear that it's obsessive. But the thing that both him and Steve share. So first of all, both of them are, I think they're a little bit like on the kind of like FU category in terms of like they're doing what they want to do and they want to build. Yeah, pretty much. However, one thing that they both share, and Steve was telling me this, I was asking, how are you, man? Because I saw him about a year ago last time and, and he looked a little bit more pale and he said like, dude, we need to talk about something that is really getting to us, like early adopters. This thing is like a vampire. It drains you out, you have trouble sleeping. Like a lot of people who are in this like multiple agents mode, they're napping during the day. You know, there's an email list, the kind of secret AI email list with these folks and they share this stuff. And so both of them are seeing it and Peter was telling me the same thing. It just really is draining. This might be, maybe this will pass. But back to your question. I think they're both, they're just building for the fun of it. Both of them are being chased by crazy investors, like crazy amounts. And they're at least for now saying no. Same with Steve. And I love it. I feel we're finally back.

1:06:05

It's so refreshing.

1:07:28

The last time I remember it was like 90s or 2000 when there was a hacker culture and this is it.

1:07:30

Yeah, yeah. I mean, yeah, you go back to like the Linux folks and like there's so many open source folks who did say no to the business community and wound up building. Great. Yeah, please. This one's empty, but we can add that one. But it does feel like a return because while the mobile platforms were so closed down, there was less opportunity and now we're sort of migrating back to desktop and open source.

1:07:35

And I think both what Peter and Steve are doing, the reason it resonates so much, they're doing the stuff that the big guys will not do because it's too risky. Peter, like, Claude took off because he himself connected to all these things and of course there's risky and security risk. He knows that. But like, and I think one of his.

1:08:00

Read the disclosures, read the disclosure, but.

1:08:17

No one would do this because it's not ready.

1:08:19

Steve also, the other labs would build it, but they wouldn't make it. You wouldn't be able to use multiple models. It would be pretty kind of clamped down.

1:08:21

And Steve was telling me that what he did with Gastown. So both Steve and Peter are amazing software engineers. They have built incredible systems. Steve's had built a lot of Google's internal systems. He's been at grac, at Amazon. Amazon. Peter, you know, built. If you see a PDF on a web or online, it's probably his business. Seriously, it's amazing. It's so big. And Steve said that the reason he built Gastown is not because he believes that Gastown will work. It's crazy. You guys have deployed the mayor polecats, all that. It's a mess. The reason he did it is he wants to push the boundaries and kind of like wake people up that something will be coming. It's not going to be Gastown. This is just A start. And he said to me that he feels that he succeeded because so like Steve is. I think Steve is a lot more provocative. Peter is a lot more like I'm just gonna build whatever I want to build.

1:08:29

How have you been reacting to Steve's position on like the death of the junior programmer, the revenge of the junior programmer? Like what advice do you have for young people that are joining the software engineering community today?

1:09:16

My position changes on changes so fast. I used to be really, really worried folks like, you know, because you look at it like it's just, it's just been like such a bad like five years. Covid started. No one started. Like I was at Uber and Covid started. We just stopped hiring. Well, we still hired, we were paying people a lot more after we rebounded, but we didn't hire junior people because you didn't want to onboard them remote. And it's difficult and we didn't know how to. And then when remote was returning, AI started. And of course now there's a thing like why would you hire a junior when a senior can like pair with an AI? Now on the other hand, what I'm seeing. So it's just really hard to get a job. But when I think back of when I started, like my first job was around the financial crisis. It was just really hard to get a job. I didn't realize at the time but.

1:09:31

And there's so much else. There's so many other things going on. Like there's Ukraine and like there's, there's all this geopolitical stuff and the economy, what's the dollar and the gold doing? And business leaders are not. We think of them as only focused on AI in terms of hiring. But there's a million other things going on in the economy.

1:10:12

Yeah, but when I started out, like I remember like a lot of people in my, in my class, university class, they just dropped out. They never became software engineers. But I never thought of it that I would be given a job. I knew I needed to earn it. So what I did is I did a bunch of projects on the side. I entered competition, I built a bunch of things. And honestly I was probably like a more impressive. And I think like first of all, like we don't need to. I'm not gonna, like when I remember this, I'm not gonna feel sorry for this generation. Cause I did a study with like Gen Z, I talked with a bunch of like young Gen Z young folks and some of them are like frigging amazing. You know Andrew. So the kids are gonna be fine. That's one. The other thing is the turns. There's this. I'm gonna spill it, I guess to everyone who's watching tbpn There is is a massive advantage of hiring young people and Shopify is the first one to figure it out. Farhan Tawar, head of engineering at Shopify told me that he saw something interesting years back. Shopify was so early to AI they got copilot licenses everyone and no cost limit and all that. But they didn't see many teams using it. But there was this one team that was using it a lot more attracting tokens and Farhan, look what happened. Oh, there was an intern on the team and so it's like, oh, what happened?

1:10:26

So.

1:11:35

So they onboarded the intern and gave the intern his two week task and the intern was done in a day. And he goes like okay, what next? Right? And they're like, how did you do that? It's like, oh, well I just used this AI and did this. What next? They're worried they want to get a return offer. And suddenly the people on the team felt stupid, right? So they started to learn from the intern and use AI, but the intern was not threatening them. The intra was never going to take a position. So next thing, Farhaan did hire an intern. Every single Shopify team. So this is why Shopify had a thousand interns. He's in some CTO groups. So why do you think CloudSphere is hiring 1,1100, 11,111 interns? Why is GitHub hiring more insurers? Farhan told them this is the biggest hack in actually getting your team more productive. Get an intern, a good intern from one of the things.

1:11:36

Yeah. The other thing is I've heard stories of people saying like I've submitted a thousand resumes or a thousand applications. I'm like, if you want a job as a software engineer, build a piece of software without asking.

1:12:26

Yes.

1:12:38

Just you don't need to be one job.

1:12:39

One job a week you should be applying to, not 150 a day.

1:12:41

Well, not even that.

1:12:43

Or one job a day. Like we hired somebody last summer. We hired somebody last summer, built us a guest directory that like in real time would take. We had had an episode. It would scrape it.

1:12:44

Yeah.

1:12:54

It would put it into a little guest.

1:12:54

If you want to drive at Shopify, build plugin.

1:12:56

We hired the guy.

1:12:58

There's so many different ways ways to plug in.

1:12:59

But yeah, you can every single company out there, if you want an internship or a junior job, just go Build a product for the company, make it look, make it match the brand kit exactly. Make it look like something that you could ship. Even if it's just rebuilding something that's existing. That's how you're going to get attention for sure.

1:13:00

Yeah, completely agree, totally agree.

1:13:18

And also like, you know, one thing that I think will, will happen with software engineering, which has happened like for 20 or something years, software engineering is not too different to like electrical engineering, mechanical engineering. When I graduated a lot of my friends were in that domain. In those domains. To get a job as a mechanical engineer you need to go to university and you go to one where you get a placement. So I think honestly this will probably return that. The most straightforward way to get a job is go to a reputable university that has placement programs and has industry connections and get in there. If you go to a lower rank one, you can still get it, but it'll be hard. And if you're self studied, you can also do it. But now you need to do amazing work, you know, like, like build something like OpenCloud or something like that. But the lower prestige, like it or not, tech was this really special place where pedigree did not matter for years or decades. This might be ending, unfortunately.

1:13:22

Are you interested?

1:14:11

Yeah.

1:14:13

The other thing is you don't need to, I think telling somebody they need to. If somebody can go out and create something like openclaw and go viral, like amazing. But you can also just look at a company, look at a feature, find a feature that their competitor has and just build the feature and say here, you don't need to create something totally novel, just build the thing, share it and that's how you get attention. I guarantee if you cold email four people in the org with a link to the product, you will get an interview.

1:14:13

Yeah. And don't forget I think the intern disclosure and all of these how to get things. A lot of people, when you do the thing that most people do, you no longer stand out. So there's a lot of advice. There used to be how to get into faang of like, all right, do lead coding. And what happened is everyone is really good at lead code, the algorithm problems and those companies don't really look at that signal anymore because everyone is so good. So I think at any point in time you want to do stuff that others don't do and it'll be a lot more effort, it will be more risky, but that's the way you're going to stand out. So like, you know, if you're young Just do crazy stuff, especially with AI.

1:14:39

Help me understand more about standing out. I remember there used to be job titles for Django developer. It's like not just backend, not just Python, the Django framework, right. Rails developer, front end in a specific framework. Now it feels like every developer can use every tool front end, back end because of vibe coding and models. Would you recommend someone say, okay, yes, you're a great developer, you can use all the tools, you can write any language whenever you want. But in order to differentiate yourself, why don't you add design or add finance or understand a little bit of legal? So is there something else that you can should pull in so you can be a dual threat?

1:15:12

Yes. And this is. So this is what's happening. When I looked at for example, the jobless thing, workos. Workos employs what they call product engineers. They don't have a product, they have one product manager for like 80 engineers. And when I looked at the job description, like I remember this evolution where it started, you know, in the 2000s, like a specific language, then framework, then a backend or something. And the job description actually had things like, you know, like soft skills, like emotional intelligence, like good communication. So first of all, hiring is turned to like the stereotype of the Guilfoyle developer. It's just not going to get hired in a lot of companies these days. And most places are expecting business sense, product sense. And the thing that almost every modern company, from linear to Work OS to all of the other similar ones are saying is taste. For software engineers now, this is pretty, pretty difficult. Interesting enough, inside Microsoft there's an internal core that they're trying to teach taste to people. Yeah, crazy Scott Hansel almost telling me this, but it's kind of hard to define. But it's a mix of, I guess, do your own projects, try to do real work that actually either makes money or helps people, or does things keep trying and use latest tools. It's hard to give universal advice because the thing has changed so much. And one other thing that's happening, some companies like the big change we've had over the winter is everyone has realized who has paying attention and paying side projects, that us software engineers will no longer write the code. The agent will write the code, we will prompt it. Which as some people are grieving, I'm kind of over that. But a lot of companies have paused hiring for now because they need to figure out who are we going to hire. Because there's no point in coding interviews or much less point when the agent is doing it so there might be like a cool of like 6 months to 12 months where there's a lot less hiring as companies figure out what is this new person. So in the meantime, people looking for the job is like first of all, if you get a job offer, appreciate it, probably just take it, get your door in the industry, it doesn't matter where. And then, yeah, keep pushing.

1:15:52

Yeah, I wonder if I feel like a big opportunity right now is software engineers finding the companies. Of course, the companies that are really beat up in the public markets right now, the software, there's tons of companies that people are just saying it's over. Of course many of them will recover and, and a few years from now be worth multiples of what they are today. But doing the work to figure out which one of those companies have durable customer relationships aren't just going to be fully ripped out. And then you could have a situation not that dissimilar from people that ended up joining companies in late 2022, 2023, when things were really corrected, where then the Stock just trades up 3x over the next two years and you look like, like an absolute genius.

1:17:48

Yeah. And one thing that companies are starting to hire for, but they're struggling. I talked with a startup in Amsterdam. They were looking for an AI native junior engineer. Let's hire junior engineers. And their expectation was like, knows how to use all these modern tools, has built some, doesn't have to be worked with some project with LLMs, you know, like built a bot or something, they couldn't find anyone. So like go and use these tools, build some stuff, put it on GitHub, launch it on the app store, et cetera. And the reason they want to bring it in again, they want like some of these folks will use it a lot better. So there's a lot of opportunity. I think we might be overthinking this a little bit. It's a massive tech change. A lot of existing folks inside the industry with like 5 plus years of experience are paralyzed because that skill coding, it takes friggin hard effort to get good at coding. It took me three years to get okay and then 10 years to get really good. And that's kind of down the drain on one end. So a lot of people are grieving that. So there's an opportunity for people who are kind of like stepping over this and like, all right, how am I going to be effective with these?

1:18:29

Yeah, this is great. Last question for me. How do you think about the like it used to be that software engineers would join a company because they were obsessed with the technology. They joined 37signals because they wanted to work on Rails or, you know, Shopify. And there were a whole bunch of like rail shops and if you enjoyed Rails, you would go there. Now it feels like there's more opportunity than ever to find an industry or a category. Like if you're obsessed with travel or you're obsessed with gaming or music, you can, the tools are maybe less important, but the willingness to be excited about the product that you're making is going to be the differentiator as opposed to just, yeah, I work at this company, but I don't really listen to music. So why am I at Spotify? It's like, you know, if you love music and you love advancing that industry, you should go there more than anywhere else. Does that resonate with you?

1:19:28

It resonates 100%. At my brother's company, he's building a startup craft docs over the winter he realized like these agents are really good. And he just mandated his team everyone using cloud code. Everyone needs to do prompting and you need to generate your code. And he said about 50% of his team got really demotivated and one engineer quit because he told him, I do not want to be a prompter. I love the craft. That engineer was really code was his kind of identity. He was the go to guy. He understood the code base better than anyone. And what happened in this first week is people used to go up to this person asking for like, hey, how does this work? Et cetera. No one walk up to him because the AI could answer it. And on this retreat in Utah, I talked with a bunch of people who've been on larger enterprises, like big companies, and they were telling me the same thing, that the people who are struggling, whose identity is the code and the craft and the people who are thriving, their identity is less of the code. It's the impact. Let's build cool stuff, let's help the business or let's. I love this industry and I want to advance it. So I think that that's going to happen. Like builders who don't care about the tools that much are going to thrive. People who grab at the craft, there might be niches for them and it's not going to go away, but it'll be harder.

1:20:21

Yeah, you always see that in the video industry.

1:21:31

Or I'm thinking of the Dwight. Dwight.

1:21:34

Every programmer loves video games. And so, I mean not everyone, but a lot of them. So go there and they'll be like, oh yeah, pay me half as Much as if I was doing ads because I. I want to work on League of Legends or whatever.

1:21:38

Yeah.

1:21:48

I'm reminded of the scene Dwight Schrute trying to out outsell selling against the computer. Right. Like immediately he's like doing fine and the. The website just steamrolls it.

1:21:49

Yeah, it's wild times. Well, we really appreciate you coming by the.

1:21:59

Come on. Anytime, anytime.

1:22:02

We'd love to have you more.

1:22:04

We'll do so.

1:22:05

It was great.

1:22:06

Thank you for the. Thank you for the cookies.

1:22:07

Let me tell you about Turbo Puffer, serverless vector and full text search built from first principles on object storage. Fast 10x cheaper and extremely scalable. And I'm also going to tell you about console consul builds AI agents that automate 70% of IT, HR finance and support giving employees instant resolution for access requests and password resets. And without further ado, I believe we have Dhara Khosrowshahi from Uber in the Restream waiting room. So let's bring him into TPP and Ultradam. Dharak, good to meet you. How are you doing?

1:22:09

Nice to meet you. I'm doing great, thank you. We just had earnings. They were really strong. Although the market is going crazy as we speak. Who knows?

1:22:38

Let's start there. Break it down for anybody that wasn't paying attention to your guys. Fantastic quarter.

1:22:47

Yeah. What happened and maybe why did you have this particular quarter?

1:22:53

So the quarter ended a great year for us in the quarter. We announced trip growth of 22%. We're now at a run rate of 15 billion trips. Well, there you go. 15 billion trips for the year. That's 40 million per day. So the business is really scaling. Gross bookings, growth actually accelerated. Usually as companies get bigger, they slow down. We're actually accelerating. Profits is defined by EBITDA. We're up 35% on a year, on year basis. And then we threw up off almost, almost $10 billion of free cash flow. 9.8 billion of 40% year on year. So the business is really scaling. And at the same time you've got the mobility business that's growing at really healthy rates. 19% and then Uber eats that. Really we built like entirely organically. Now is over $100 billion run rate, solidly profitable. And that grew 26% in the last quarter, really accelerating nicely since the beginning of the year. So. So the business actually looks really, really great. And then obviously we've got the onset of autonomous and autonomous vehicles. And we kind of. In our investor presentation, we had a lot to say about autonomous, which we could definitely get into if you guys want to. Yeah.

1:22:58

Before we get into that, like, what are the key drivers outside of obviously the businesses maturing? You've been in the driver's seat for long enough now to really start having clearly a tremendous impact. But like, what's really, what's driving this acceleration.

1:24:16

Yeah. So I tell you one, in terms of the businesses, our business is really supply led. So the more cars we get on the road, the more restaurants we have on the platform, our product gets better. There's the conversion of the product. You know, when you open up your app, your ETA is lower, you have more restaurants available. So the basically we are supply led business and we continue to add supply onto the platform. We've got almost 10 million drivers and couriers on the platform, well over a million merchants. And first it starts with supply and our selection and supply just keeps getting better and better and better. I'd say that's number one. The second is we're the only platform out there that has both rides and eats and is global as well. And that allows us to more deeply embed with our customers. About 30% of our eats first trips come from riders, so to speak. And we can promote, cross promote from the rides platform to the eats platform. If you open up your Uber Uber app now, you see eats being offered, you'll see grocery being offered. If you're looking to go to a restaurant for dinner, we'll also offer that restaurant for delivery as a reminder that it's available for delivery as well. So this cross platform kind of promo that we can do, it's just no one else is doing it. It's taken us years to perfect it because any pixel that is promoting, let's say Uber eats on the Uber app, it can get in the way of your experience on Uber. So you have to be very careful to have an optimal experience for your base business while at the same time nudging you over to the other products, you know?

1:24:31

Yeah.

1:26:15

Was there a period where it just felt like you were running two very distinct businesses and now it feels like they're actually coming together, working?

1:26:15

Totally.

1:26:23

And you know, my philosophy is that when you're building businesses that are younger, you want kind of full stack teams. So for example, our technology team, we had a dedicated technology team for eats. As that business got more mature and larger, we combined our tech teams, we combined our product team so that you can build one core infrastructure, one matching stack, for example, bring the customer data together and kind of create more commonality between the businesses. Now, for example, the my I promoted our head of global mobility, Andrew McDonald. He's been here for 12 years. He's now president and COO. So he's running one P&L. And essentially we can trade off between benefits on one business for the other business to kind of grow the whole. And there's no one else who again, has the breadth and depth of offerings that we have have. And then on top of that, We've got the Uber 1 membership program. We got 46 million members. Close to 50% of our gross bookings come from members. And that membership provides a real lock in Members spend three times more than non members. Customers who shop on both eats and rides spend three times than the ones who don't. And so it's that combination of the whole platform coming together that lets us grow fast, that lets us accelerate, it lets us gain category position versus the others while at the same time having a lot of profitability.

1:26:23

What's the biggest lesson from Expedia that you've been able to apply to Uber?

1:27:52

Yeah, it's a great question, I would say. And it was a mistake at Expedia was not recognizing that the business is supply led. So with Expedia we were very much focused on the customer and bringing in more audience, so to speak, and then following with supply. And I think one of the regrets, and you have to kind of learn your lessons as a CEO. It's like we're humans too, we make mistakes too. We learn as well is that booking.com was really focused on more supply in Europe, for example, and because they had more supply, they were able to build a more compelling platform in Europe than Expedia was. And so I took some of those learnings and then of course, you know, some tips from the OG Uber folk, including Travis as well, that this is a real supply led business. And so it was to some extent learning from my past mistakes and then making sure I don't make those mistakes again in the new job, so to speak.

1:27:57

Yeah, I remember when Uber came to la, the push for supply on the platform was so aggressive that I was getting directly targeted. Hey, do you want to drive for Uber? I was employed at the time. I wasn't signing up, but I haven't seen those ads. Is that a function of better targeting or do you just have a more natural driver acquisition flywheel, like what's working on the supply side, increasing the supply of drivers these days?

1:28:53

I think supply is definitely cyclical. So in the early days, obviously it was kind of a new thing. No one knew what gig Work was, was so Travis and team had to introduce it to drivers, had to pull them off of, you know, driving for a carry limousines or some of the traditional players.

1:29:21

Yeah.

1:29:37

Onto the Uber platform.

1:29:38

It was very broad.

1:29:39

Yeah, totally. And then we had to kind of grow the, the overall marketplace as well because there are way more Ubers on the road now than taxis etc. But we have had kind of supply shocks in the past. So for example, Covid. One of the great things about COVID as it related to Uber, it was terrible event obviously was that we were able to move our drivers who are driving for Uber to make money on Uber Eats. But then after Covid, we thought when the world would open up, everyone would get back to driving because why the heck not? It turned out that a lot of our drivers, rightly so, they were worried about their health. They didn't want to catch COVID I remember I was talking to a driver and I'll never forget this line. I'm like, well why don't you go back to driving people because you can make more money that way. And she said, well, big A Big Mac isn't going to give me Covid. You know, it's, it's. So safety was a big issue so we had to reinvest in supply very aggressively post Covid. We did so faster than our competitors, which helped us. Now we have a pretty finely tuned machine in that the retention of drivers is really high.

1:29:41

Yeah.

1:30:48

On average, drivers are driving more supply hours. They're spending more time on our platform because they can drive, they can deliver, they can shop, they can even do, they can judge algorithms. Now we have a group called Uber AI Solutions where they can look at algorithmic answers and decide which is better, etc. There's much more to do on the platform. So the position that we have in supply is, is pretty strong. But believe me, we're investing. I guess we figured out that you guys are more consumer customers than drivers and good for algorithms.

1:30:48

Good work on the marketing side. Everyone's been able to make the bear case for traditional enterprise software over the last year. As models have gotten better, there's been some chatter about kind of risk to marketplaces with agents. I'm not necessarily bought into some of those theories yet. You know, you're operating this, you know, massive network of drivers and supply and it's just a lot more difficult to sort of no model is going to one shot that anytime soon.

1:31:25

I'm expecting to go to my terminal and it will call people randomly in my neighborhood and ask them if they want to give me a ride, that's.

1:31:56

The future I want to live in anyway. So we're not really quite seeing that yet. But how are you thinking about kind of risks and opportunities with. With a more agentic kind of Internet and consumer product?

1:32:03

Yeah, I think that if you have unique fragmented supply and the global scope that we have, then you are safer than, let's say, others who might not have unique supply or they don't have the fragmentation of supply that we do. Like we got over 9 million couriers and drivers on our platform. That's very difficult for anyone, whether human or agent, to replicate as well. So we are working with agentic platforms. We've got a great relationship, for example, with OpenAI and so you will be able to call an Uber using ChatGPT. One of the things that we want to do is we want to get our brand in front of you as well. So once you call it, for example, A and Uber Maps so you can track your car, do all that good stuff. So I think, like, companies like ours who do have, you know, we're in over 70 countries, we've set everything up. The membership program is very, very deep and embedded. We have a fragmented supply base that's going to be very difficult for anyone to recreate. I think we can rest easier. And for me, I want to make sure that my brand gets out to the customer and at the same time, I'm building those experiences within the Uber app. So a driver can interact with an app using voice. They can. They have an agent who can help them out if they don't know why they're not making more money. We've got agents for ordering, we've got agents for ordering your rides, etc. So I think as long as the experience on app doesn't trade trail the experience off app, I think off app actually can ultimately, for players like us, expand the marketplace and. And that's definitely something that we're looking at.

1:32:15

How are you thinking about the next iteration of Apple Intelligence rolling out? We were talking to Mark Gurman and he was saying that we've been able to call Ubers from Siri for, I think he said a decade or something like that. It's been there, obviously, I'm sure you have the numbers and there's probably a lot of people in nominal terms that are using that functionality, but it feels like it hasn't broken through, but it feels like it's about to. With the next iteration of Siri and Apple Intelligence, Intelligence is that that's lower friction that feels good for you. At the same time, if I don't open the app, maybe I don't go and do other things and I lose a little bit of connection. What, how are you seeing that trade off?

1:33:52

You know, my, my view is don't think too hard and build great consumer products. And so if someone wants to, if someone wants to get us through Apple Intelligence, ChatGPT or they want to come to the app, I don't care, you know, as long as, as long as we're getting used and we're getting used used often now, yes, there are considerations. We have a $2 billion advertising business.

1:34:30

Yeah, yeah.

1:34:55

If that is subsumed underneath an Apple Intelligence, that could hurt our advertising business. But for me, like if, if the North Star is build out consumer experiences that are absolutely first rate and then we have this, the multi product category and the membership underneath, I think the economics can kind of take care of themselves over the long term. And if we make a great app, you know, like people will keep coming to that app. And I remember when Google Maps was at some point amalgamating Lyft and Uber content, people were worried about that. She talked about Apple Intel. Our app is pretty damn good. And my job is to make sure that it remains, you know, leading and keeps attracting audience because ultimately that is the most engaged audience that you can have. That direct audience.

1:34:56

Yeah, that makes a lot.

1:35:49

Yeah. I mean with food, the benefit with, with the Uber Eats business is like I don't really want plain text experience ordering food and chat. And then yes, you know, chat can generate images and stuff like that, but.

1:35:50

It'S several jumps in ui.

1:36:03

Eventually you're like, well, I can just pop over to the app. Yep, we got to talk about autonomy. I'm sure, I'm sure you spent a bunch of the earnings call on that note, but what were the highlights? What should people be paying attention to?

1:36:05

I think the biggest highlight, and we have a big section in our supplemental materials that I encourage folks to look at is that what we're seeing with autonomy is that it's a net, net positive in terms of demand into the ecosystem. So there's one view of looking at it. Hey, autonomous is just going to replace humans and you know, robot cars don't get tired, they don't get distracted, they can work 20 hours a day, they do need to get recharged and clean, etc. So they will replace the workforce, so to speak that I think a lot of people are talking about in other markets as well. What we're Seeing in markets in which we have launched autonomous is that actually those markets are growing faster. There's a new audience segment that comes in our new customer acquisition. You know, overall now we have 200 million consumers coming to our front door every single month, growing 18%, growing pretty quickly. But in the markets in which we have av, our audience growth is actually accelerated and we're gaining new customers because it's a really frickin cool product. So one is that we've always had the hypothesis, hey, mobility is a trillion dollar marketplace, but autonomous can add more to that market. It can be trillion dollar plus or another trillion dollars. And the early signal is that, yeah, it is actually additive, it's not replacement only the second, I would say big finding, if you want to call that for us, is we've always had the hypothesis that autonomous on our platform is going to take advantage of the platform to be much more utilized than off platform. You know, like to some extent a car is a box with wheels. The autonomous car is super expensive. So you want to keep it running and you want to keep it earning money for as long as possible, possible. And we're finding that the utilization of those cars on our platform is 30 plus percent higher than cars that are not that kind of are building their own platforms. That utilization bonus is hugely important as it relates to the value of our platform. And so we're seeing that come through too. So one is it's additive, which is great. And the platform is doing what it's designed to do, which is, is bring more business to a driver where that driver happens to be human or happens to be a robot.

1:36:18

Yeah, so you're saying so the unit economics of AV provider can actually be better if they're adding it to the Uber platform.

1:38:44

Yeah, I mean that's how we have to earn our take rate. Right. If we're not driving higher utilization, we won't earn our take rate. And right now the utilization premium that we're seeing suggests, suggests that our take rate is, you could argue, too low. But we have a very strong case to be made for the take rate of the platform.

1:38:52

Is there any role for Uber in the rest of the self driving car stack? From cleaning to service to storage? There's a lot of these things that could be handled and they're currently handled by the sort of OEMs and the big autonomy companies. Other you could imagine in the future it being much more distributed. But does it go full individual maintenance and whatnot?

1:39:13

Yeah, yeah, totally. So if you think about the AV stack, there's the distribution. You know, that could be Uber, it could be Waymo, it could be Tesla, right? Then there's a driver and the driver, we're partnering up with a driver, then there's the operator. And the operator has depots, is cleaning the cars, recharging the cars cars. We have a network of fleet partners now on a global basis because about 15% of our supply is provided by fleets today. And you know, they rent out cars to drivers. We're taking those fleet operators and we're saying, hey, we want you now to operate AVs so that those fleet operators in partnership with us, we have investments in a bunch of them. They are the ones who operate the fleet and they can do it in a really low cost manner. They're doing it with hundreds of thousands of cars. And so AV is different, but not that different. So you got the operator, then you have someone who owns the car. That could be us, it could be Waymo. But ultimately I think it's going to be third parties, like private equity players and all this stuff who are going to want to own these big assets just like they kind of do rental car fleets. And then you will have kind of debt out there. So, so there will be the financialization of the marketplace as well. But we are definitely playing in the demand side. We're partnering on the driving side and then we can and are providing through our fleet partners all of the operations in a city including depots, cleaning, etc.

1:39:37

What kind of conversations have you had or what should people expect from other OEMs outside of Tesla? On the Autonomy side, it's an interesting dynamic now where Tesla's are getting FSD is getting really, really good. It's a meaningful differentiator. If you're a consumer and you either want to drive less on your way to work or one day be able to add your car to a platform like Uber. What's kind of coming down the pipeline in terms of just overall competitiveness from other OEMs?

1:41:10

So what's interesting is that when I was having discussions regarding autonomous with OEMs, it was more, they were more interested in L2, which is where FSD is now, than they were L4. The prospect of every single car being sold with a $5,000 priced piece of software that has a cost of goods of zero. It's like OEM's eyes went like this. They were super, super interested. As it turns out. I think that L2 is proving to be less interesting than people thought because you still kind of have to watch a road et Cetera, you can't really turn off. I think the great, great customer product is just I can relax, I can do whatever I want and I can get my time back the way that you get your time back in an Uber, sit in the backseat, do whatever you want, make a call, etc. So the conversations with OEMs actually are getting a lot more constructive for us, us because they see L3 and L4 and we want L4 obviously as the big prize now. And that was not true three years ago. So we're having actually really great conversations with OEMs. We have a deal, for example, with Lucid who builds awesome EVs and one of our other partners, Neuro and Lucid coming together. And there are many other discussions that we're having with OEM partners. And then of course now Nvidia is in the game, right? They were building the reference compute architecture, now they're bringing the sensor stack and now they are building self driving as well, the model as well. So Nvidia will provide a full stack software and hardware solution to any OEM provider. And as you can imagine, Nvidia brings a lot of heft, a lot of know how and then a lot of credibility to the ecosystem as well. So, so the direction of traveling, you're.

1:41:41

Sitting there as like a supply driven business, just being like, let's go, I want every car outfitted with this when it's ready. And you can imagine a world where the Uber wait time is 20 seconds every time. It's just like.

1:43:36

The dream is 10 years from now, every single new car sold is L4 capable. Because that is obviously very strong prospect for a supply. What I started with, with. And hopefully you'll get that 22nd ETA or you'll be pissed off if you have to wait for more than a minute.

1:43:54

Yeah.

1:44:10

If you step back from maybe the specifics of Uber and just think about how autonomy might change the American landscape. Like, how do you agree with the thesis of like the suburbs will become more popular, people will be commuting longer. Like what knock on effects do you think are potentially underrated or under discussed of, of moving in a more autonomous direction?

1:44:11

Yeah, it's a great question. I'd say generally in the US suburbs are growing faster than cities. So this is a trend that started with COVID it continues. And our growth in the suburbs and less dense areas is about twice our growth rate in the cities. So it's been a really attractive growth driver for us. I do think that it's going to create a lot of Space in cities. You know, I think 30% of space basin cities is for parking. A huge amount of it goes unutilized. Some of that will be taken up by depots, etc. But I do think that it is going to increase the breadth of where people get to live and I think it's going to make transportation available for many, many more people. As the cost, it's going to take a while and it's probably going to take two or three generation of cars. But the hardware cost for these sensors for the computer is absolutely coming down. And this will result in lower cost mobility being available to many more people in the world, whether they live in the city or they live outside of the city. And I think it's going to have a great impact on society having more transportation available more affordably to a higher percentage of the population. Yes, it's good for a business, but it will undoubtedly be great for society.

1:44:37

Last question. What advice do you give to someone who wants to be a public company CEO?

1:45:55

One is, I would tell you, like, don't hunt for it, pick. I'd say that the advice that I give folks is people get way too hung up on what their title is or how much they're going to get paid, etc. I have always, you know, what's worked for me is I have been very, very specific about who I want to work for. And when you pick the right people to work for one, you learn a lot from them. But the one thing that I've seen in every environment, like environments are unpredictable, but people usually stay successful. Like you look at Elon, right? He succeeded in so many different areas. And that's true for really exceptional people. So exceptional people stay exceptional regardless of the environment. You know, if you pick a company or you pick a position that can change, that can go the wrong way. If you pick the right person, you pick the right person, not only do you get to learn from them, but they usually move up in the world. And so as they get promoted, as they build a bigger company, you kind of get to free ride off of them. So you get double the benefit, which is you learn from someone who's awesome and you get to move kind of in their way awake until you're ready to strike out on your own. I'd say that's number one. And then the second thing I'd say is like, you know, advice for young, young people in terms of learning, just learn how to work hard. Like it's again, your skill sets may change, your profession may change. People are like Discussing is coding the right thing or. Or, you know, learning the liberal arts. The most important skill in life is a skill of working freaking hard.

1:46:03

Hard.

1:47:41

And if you work for the right person and you learn how to work hard, you're going to do just fine.

1:47:42

I love that.

1:47:47

I love both those answers.

1:47:48

That's a fantastic answer. Thank you so much.

1:47:49

So great to finally have you on the show. Congrats to the whole team. Fantastic quarter and we'll hope to have you back on very soon.

1:47:51

We'll talk to you.

1:47:57

By the way, I don't know if our IT people have been watching you guys, but we have ramp coming on to Uber, so.

1:47:58

No way.

1:48:03

Something happening. My IT people are watching way too much pdp.

1:48:04

Adam, that's amazing.

1:48:10

Love it.

1:48:11

That's great. Well, congratulations and we'll talk to you soon. Thanks so much. Goodbye. Let me tell you about Restream 1 livestream. 30 plus destinations. If you want to multi stream, go to restream.com and let me also tell you about Plaid Plaid Powers, the apps you use to spend, save, borrow and invest securely. Connecting bank accounts to move money, fight fraud and improve lending. Now with AI public company CEO, we.

1:48:11

Have our dear friend Mitchell, Mitchell Green, Lead Edge Capital coming into the show. Let's bring him in.

1:48:39

Oh, he's here. Mitchell, how you doing?

1:48:45

You should have had Dara and I on together. We're buddies.

1:48:47

Oh, you are? That's awesome.

1:48:49

Yeah, we should. If we plan better, we'll add you both next time.

1:48:52

Yeah.

1:48:56

How's life?

1:48:57

It's good, it's good. Yeah. We're going all over the place today. We're talking about the market collapse, the end of SaaS, but the birth of agentic coding and all these crazy things the developers doing. Yeah.

1:48:58

What's up?

1:49:11

I got a question for you. If everybody, my partner actually came up with this idea, he's like, so if all the SaaS company of all the software companies are all going to disappear because just like Vibe code your way to building in one of these companies, well then can't the next AI company Vibe code that AI company?

1:49:12

Yes, yes, yes. And I mean there is seriously a push in the open source software community now to loosely Vibe code everything you need. Like you build all your own tools, you build all your own software. This used to be the domain of programmers that would stay up all night for months on end to build a little to do list app and now it's much more robust. But, but we haven't really seen where all of this goes. It's still pretty early and most people are just building tools that build more tools at this point. The iron law of the universe, I don't know if you agree is if it goes up quickly, it comes down quickly. You see these amazing growth charts and you get a little nervous because if it. Easy come, easy go.

1:49:29

Yes, it's funny. I think people are going to be in for a rude awakening. I think some people.

1:50:10

I mean you were calling this the last time, the last time you were on the show you were kind of more just kind of skeptical. A lot of these valuations.

1:50:17

Yeah. What's funny is it's like public market software is now cheap. We're buying it. But we think companies like Workday are incredible businesses. Company like Transferwise in the payment space remitly Toast has been crushed. We're not in a company called Applovin but huge gaming company. I have a good buddy who's made billions on it and is now rebuying it. Like these things are just like.

1:50:27

Yeah, the crazy thing, I was like.

1:50:51

All sent out at the same time. So.

1:50:53

Yeah.

1:50:55

And yeah, I was looking at, looking at PayPal. You have a. It's trading at seven times earnings now. When I looked this morning. They have half a billion users. It's a global payments network. It doesn't seem like anything. You can just vibe. You're not going to vibe.

1:50:55

We're going to vibe code it, we're going to build it, we're just going to make money. Same with workday as well. We're going to vibe code our way. You know, only took work day 20 years to get companies like Nike and Procter and Gavel to give them all their ERP data. Of course, you know, Mitchell and Jamie, John and Jamie. And we're going to like. We're vibe code our way.

1:51:12

Yeah, I think, I think people are, people are undervaluing trust right now.

1:51:27

Totally.

1:51:31

People forget how software companies like building code has never been the actual like hard thing about a software company. There's not like building semiconductor chips. It is like distribution. People forget. You build a piece of software, then you have to sell the software, then you have to maintain the software, then you have to add features to the software, then you have to connect it with other systems. Those systems change. You have to have user permissioning. This stuff is really complex. I'm sorry, the dude sitting on Silicon Valley in a shed on Sandhill or in San Francisco, who's going to vibe code their way? Oh, I'm sorry. How about the other 10,000 software engineers? Salesforce has or the 3,000 software engineers at Workday has. I mean are they just like sitting around on their thumb? No, of course they're going to try to innovate and what's going to happen is no different than what happened in 99 and 2000. If you look at the top 50 sellers on the Internet today of like the largest e commerce companies like yes, Amazon is one but, but Walmart, Home Depot, you know they made it through Macy's. Yeah, it's a good E commerce businesses.

1:51:32

Like one way really quickly.

1:52:40

You can actually test that thesis empirically by looking at R and D spend of tech companies and how much they actually spend of the money they raise before they IPO on software development specifically. And it's shockingly low. Way lower than you'd think. You feel like software company goes publicly.

1:52:41

I literally ask all these like pod shops and all these hedge funds, all these hedge fund guys and by the way the trade, there's a great chart I could push a pull it up which is like, you know, it's like software exposure for hedge funds has gone like this. Yeah. Semiconductor exposure has gone like this.

1:52:57

Sure.

1:53:13

And you know usually if you do the opposite of what everybody's doing, you'll make a lot of money.

1:53:14

Okay.

1:53:18

One, you know, one, one kind of test you could run is like if you're thinking about a company like Workday is if you went to all of Workday's customers right now and you said hey I can, I can, I can build this for and I'll sell it to you at 20% of the price. Would anybody actually change switch over? Because it would be like okay, it's gonna be cheaper but what if it's like way less reliable? What, what if I get, what if it's so unreliable or has so many issues that I get fired over?

1:53:21

Yeah.

1:53:49

Does anybody want.

1:53:50

There are like multi billion dollar revenue businesses that are literally built to sit on top of companies like Salesforce and Workday to help them manage the companies to manage the software. It's now again so then you think about like I think companies that are like selling to big enterprises are actually going to end up great. Now again you might have the Sears and Kmarts of the world that didn't innovate or were over levered. I do worry about some of these private equities software companies that get way over levered and if they try to drive too high EBITDA margins and, and sacrifice R and D. Now there are other firms like a Vista like what Robert Smith does and he's actively like they're trying to be like, no, we need to like double down on AI and like double down in R and D and these things. But I think there will be some people that are just trying to maximize EBITDA margins that'll be hurt and guess what? Somebody will go build the next competitor of that. So then you're like, look, we think about like an area that's probably more right for AI disruption is companies selling into like SMBs because the software isn't nearly as complex. But again there are companies. It's not like companies like HubSpot which is a great business that sells in the S and B there. They have probably a thousand software engineers.

1:53:50

Yeah.

1:55:03

They're sitting there trying to disrupt themselves constantly. So it's not, I don't know, I think it's.

1:55:03

Yeah, it's definitely market by market, sector by sector.

1:55:08

Look, I think, I think people don't truly understand how like software companies actually sell.

1:55:12

Yeah.

1:55:19

That just say that people are going to buy code it away now. What is incredible is how fast some of these AI businesses have growing that are built off the tailwinds of some of these companies. Even like a company like a Clickhouse which just raised at a big round. We were the early investors in it. Luckily, totally by dumb luck.

1:55:20

We have.

1:55:41

Four investors in Grafana Labs, this software company, but embracing AI and just had an incredible quarter and plan. And so you are seeing there are going to be a bunch of companies in software that continue to innovate.

1:55:41

Could you vibe code equipment share? That was one of your.

1:55:56

It's funny, I am vibe code. I was going to vibe code car collection software but then I decided since I actually have a job for a living, I'm going to hire somebody else to vibe code it for me.

1:56:01

Total software engineer. Victory.

1:56:12

Still.

1:56:15

Still creating another job.

1:56:15

Actually some, some 18 year old senior at Santa Barbara High School is benefiting from this.

1:56:18

That's fantastic.

1:56:22

Build me software.

1:56:23

But they're gonna have, they're gonna have a job forever because they're gonna be maintaining. Stop buying cars.

1:56:24

They're gonna be maintaining this thing for years.

1:56:30

I'm gonna get you guys in on it as well.

1:56:32

Oh, I like that. I'm down.

1:56:34

Equipment share. Look, I mean it's up almost 30% from the IPO. It's with my partners Tim and Zach. My partners Tim and Zach led that deal. I take no credit for it. They did a phenomenal job. It's look, it's benefiting from, you know, the ongoing. There's a massive if you were to isolate you know, CAPEX and infrastructure spend outside of like GDP know driving GDP growth right now. I think it's like a very meaningful part. Highways, energy products, data centers. I think another big thing that's going to drive a lot of CapEx is this, this, you know, accumulated depreciation, this accelerated depreciation and the big beautiful bill that will benefit people like Equipment Share. And to put it in perspective, I think companies like Equipment share in the S1 because they haven't filed results yet. I think last quarter, you know, in the as one result in the last quarter that's one and the as one was like growing nearly 30% a year. Yet you have companies like United, Hertz and Sunbelt that grow like single digits and yet these things trade at almost similarly but our margins. Now the reality is is people that tend to buy companies like Equipment Share Hertz Equipment are not used to businesses that grow 30% so they need to just like continue to put up results. And I think they'll look at the, look at the multiple expansion over time but they are, it is like a great second derivative play on AI totally.

1:56:37

In the public markets. When you look at a software company it feels like a lot of people are looking at, you know, is top line, accelerating. That should be happening. If you're a true AI company, you're also looking at stuff like are you over leveraged? Then you can also just talk to the management team and if it doesn't sound like they know anything about AI, you could be very, you know, skeptical of the company. How do you think about when you find a new company? What research? What's your process for understanding how they'll be positioned over the next couple years?

1:58:00

Yeah, that's a great question. I think what we should do is just count the number of interns they have frankly because 24 year olds and 22 year olds are going to know more about this than 50 and 60 year olds and 40 year olds like myself. All kidding aside, the way we think about it, it is, we just think of how is this like squarely in an area that could be like very easily disrupted by AI. We just sit and like how deeply is this product integrated? How many people are using the product? Is it as a result of stuff coming out of. Do they own their own data? I think we just, we think about how complex the product is to build. You know like we talk to a company that's growing like a rocket ship and that makes like software to help like voices in call centers. And for us it's more like a rocket ship. But we're just like, but they sell into call centers, but the call center software companies just build this themselves and then it's like a function of, okay, well what valuation then do they want? And then you dig in deep and it's like, and I'm not saying this one way or another, do they have like one or two customers that are 30, 40% of sales concentration? Know we were, we were looking at a company recently where like 40% of their sales is to a company that, you know, we're pretty skeptical on whether it should just in 10 years. And it's like 30, 40% of their sales.

1:58:29

Yeah, that makes sense. It's kind of like a dying gasp from that bigger company. And they're sneaking the money out the back through this new startup. Talk about the IPO window. Feels like it's been open. There's a lot of rumors about the big company SpaceX OpenAI anthropic. But what does it mean? I mean you saw equipment share, like what, what's your take on the IPL market and where it might evolve over the next year?

1:59:56

Good companies can go public at any time. Except like, you know, equipment share is up 30%.

2:00:18

Yeah.

2:00:24

Like, boy, it was a great, it was a good deal.

2:00:25

Yeah.

2:00:28

I think, I think what's going to be very interesting about some of these giant IP equipment shows, like a billion dollar deal.

2:00:29

It is not sucking liquidity out of the markets like going out and raising $200 billion.

2:00:38

Yeah. A lot of people seem super confident that the market will fully support a SpaceX anthropic and OpenAI David Goggins market.

2:00:43

We shall see. By the way, the money has to come from somewhere.

2:00:53

Yeah.

2:00:56

So you have to say, was it going to come out of cash? Is it going to come out of Treasuries? Is it going to come out of Google, Amazon, Facebook, Microsoft? Yeah, probably. The reality is, is a company like a Space X or an open air and anthropic or a stripe or let's see, let's use like ByteDance. So ByteDance trades as massive amounts of earnings. You know the negative is it's China, right?

2:00:56

Yeah.

2:01:21

But the positive error is massive amount of earnings. And you know, you can get to a giant number like valued like SpaceX of a fundamental earnings number number. So you're like, where is $150 billion going to come from? And so in that one you probably say, well it would probably attract new money back into China, but then might people pull out some, pull some money out of Tencent and pull some money out of Alibaba or Google or Microsoft to get exposure to this thing. I'm sure in the case of a SpaceX it would take money out of a, you know, it would take money out of Tesla because it's like Elon, you know, it's like die hard forgivers. But maybe he's just going to merge, you know, SpaceX and Tesla together then change it to the Elon Musk company.

2:01:23

Yeah, yeah, I think, I mean I think Tesla shareholders would have been very frustrated to get XAI merged in, but getting SpaceX merged in and maybe you got a little X. I know nothing.

2:02:09

But it shocked me if it happened.

2:02:21

Someone in the chat just said that the ticker X just got reserved at Nice. So we'll see what that means.

2:02:24

That's exciting. What do you think of the secondary market? When I, when I moved to Silicon Valley it was like seen as a total bear signal if a founder CEO was selling stock at the early stage. Now companies stay private so long there's massive, massive secondary sales. Elon's been a master of, you know, liquidity, keeping SpaceX private, private for 20 years. How is the secondary market evolving? What are you seeing?

2:02:30

It's definitely evolved a lot. When I started working at Bessemer in 2000 4005, you know, it was a very, you know, it was very, you know, world class venture funds were like, oh my, absolutely not. You should not let the executive areas. And I got into my first investment ever in 2009 in a company called Bizarre Voice doing a secondary. Now I will tell you, most deals have a secondary component. It feels a little bit frenzy ish right now. Like the Facebook, Twitter back then. I was just at an investor conference yesterday speaking in Arizona and it shocked. I had multiple people come up to me. Institutional investors, like random companies, pension funds that were like, oh, what do you think of Anthropic? We're like buying into it here. Like, oh, this doesn't end well. Not just for like, I didn't mean to drop it.

2:02:55

No, no, it is one of those signals when like the person who's not a professional investor is talking to you about things that feel like professional investments.

2:03:55

Yeah, I got even a better one. One of my big institutional investors, fantastic guy name will be said but he's like, you need to start a micro secondary fund that will buy secondary interests in these tender 30 ACT funds. You have like, you have like the coat, the new code to fund and all these and it's a huge wave coming. Yeah, and basically they offer like quarterly redemptions. Like you can take out like 5 or 10% a year. It's what like bereavement was. Right. And then. But occasionally if there's too many people that want to redeem, you have to gate them. But you never want to gate. But when do you think everybody's going to gate? Like everybody's going to. All the dentists and doctors are all going to pull out at the exact same time. And my buddy was like, oh, you should start a second a microsecondaries fund to buy out their interests in all these interval funds.

2:04:03

When they all.

2:04:58

When they put it. When the gate comes up, up. And all of them want out at the same time. And they're like, wait a second, we can't get out. We're locked up. What do you mean? I thought this was like liquid. Yeah, it's coming.

2:04:59

Don't worry.

2:05:08

It's.

2:05:08

What do you. What do you think? What do you think Ferrari the business is worth?

2:05:09

Oh boy. I thought you're gonna ask me what I thought of the new car. Lewis was Hamilton was fast and tested at preseason. Dustin. It's an amazing business. The stupidest thing I ever did was buy the cars and not the stock. I actually have several friends that I raced for our challenge with that bought the stock at the ipo and now they slowly dribble the stock out to buy cars.

2:05:16

Wow.

2:05:38

It's an amazing brand. I don't know what it is.

2:05:40

Yeah, it only IPO'd in 2015.

2:05:41

Yeah.

2:05:43

What is the stock like? I mean, what's Ferrari stock right now?

2:05:44

Stock's down 10% in the last month. It's a $63 billion.

2:05:46

Built me a Ferrari factory.

2:05:53

The nice thing is you can't. Yes, it's Doc.

2:05:55

You can vibe code but it's up.

2:05:58

500% since the IPO. Right. When you got into cars you should have been getting into.

2:06:00

I should have just bought the cars.

2:06:05

What advice would you give to someone just getting into venture private equity, raising a fund, looking at opportunities. We're seeing this like K shape right now with the mega funds scaling up. Up feels harder than ever to start. Sort of a mid sized fund. What are you seeing that's interesting that new managers are doing?

2:06:08

That's a good question. I'm going to steal a word. I'm a steal a line from Jeremy Levine who's a senior partner investment venture partners and probably one of the best venture capitalists over the last 25 years. You guys can ever get a bond. He's incredible. Like, an amazing investor was like that very early investor in Pinterest and LinkedIn and Yelp than like, Shopify that with Alex Ferrara. He said to me, I would always complain about what they were paying me at the time. And he's like, but, Mitchell, you're learning a lot. And so the most important thing is actually to go find if you're trying to get into this industry. And I think there's too many people that are young starting funds and they're going to actually like or entrepreneurs that have actually never really professionally invested. I would actually tell you you're better off going to find the apprenticeship and like, and literally go find an amazing venture capitalist or venture capital firm that you can go work for. Convince Sequoia or Benchmark or Mike Maples or Josh Koppelman at first Charlie, guys that have been doing this. Gals have been doing this Theresa gal who've been doing this for a long time and go shadow them for two years. That's how you're going to really. That is honestly the best.

2:06:29

Or 10.

2:07:36

Or 10.

2:07:37

Or 10. Yeah.

2:07:38

Like. Like so many. Yeah.

2:07:39

A lot of people get in these funds. Two years later, I'm not. I'm raising my own fund.

2:07:41

It's like, like you haven't seen a full cycle going. Like, maybe you're a good investor. Maybe you've made some good angel bets. Maybe you did a couple years at a fund already. But going and saying, I'm going to do the, the, the path of doing the 30 to 50 to 100 million to $200 million fund, like that is, like, it's tricky, especially where prices are.

2:07:44

I mean, I, I Give Harry at 2020 BC a lot of credit. Like, he started off the podcast, he, like, hustles his way. Like, I'm sure you guys are going to get to start to see, like, really interesting deals from doing this. And it doesn't mean in like 18 months now you're gonna go start a fund. No.

2:08:02

Yeah.

2:08:15

Yeah. I think it's people confusing angel investing and being successful at angel investing with all of a sudden, if you get 100 million bucks, you'll be really successful. It's really, really hard. And I think people need to find mentors and go work for people that have. That have invested as you through cycles. 100%.

2:08:18

What. What does it take to actually be contrarian? Because it, it's a. It's like whenever, you know, it's such a popular word in this industry now, everybody says they're contrarian. And yet, of course, Only that's consensus. Consensus. Yeah, exactly. And so I feel. I feel like it's something like human. Humans have an innate sort of instinct to kind of go with the herd and, and 100% and. And so you have to like fight that. But how do you fight?

2:08:38

Steve Cohen has always told me, if you want to know where the stock market goes, but he's like, look, you got a lot of smart hedge fund guys in your fund, himself included. Ask us all where we think the market's going. When eight of us tell you one thing and two say the other, please call me and say what the two said. That's where it's going. The market causes the greatest amount of pain to the greatest number of people.

2:09:07

Well, yeah.

2:09:26

And to be honest, right now, if you looked at X, you would say the consensus view is that AI is in a bubble and that we're going to have this massive correction. And the contrarians are saying are the runes. Obviously he's very Dylan. Dylan Patel saying people aren't prepared.

2:09:27

Anthropics numbers are alone.

2:09:46

Yeah, look, my view on all that is good company and good Investment can be two very fundamental things. You may wake up in five years in anthropology is worth 300. $300 billion. It's $350 billion. I want to go back to the contrarian thing in a second. People are underestimating over the next 10 to 15 years what is going to do myself included. Every one of your guests included. In 2000 I could have sat on your show if it was around and nobody would have Talked about the $3 trillion of social media that's been created value. Yet people always overestimate the near term. The bubble will probably. Whatever bubble we're in will be different than 20, 20 and 21. It'll be different than 99. There will be some correction. Nobody will be able to predict when it happens and the correction will just happen. But then that is probably the best time to then buy a bunch of these names and hold on to blunt. How do you be a contrarian? I don't know. You'd be like Lindsey Vaughn and Terry Rac Allen, try to get out, go win an Olympic medal. You know, I think people that have. I think contrarians tend to be people that understand how to take very calculated risks and are not afraid of doing that. So like what we look for in our analysts that come and work here, you know, cold call companies all day long. You know, by the way, the good CEO doesn't call you back. It's a CEO you call every 10, 10 times and calls you back after 10 times times like that's. It's like grit, persistence, not being afraid to be different. So like you look for like athletes. I'm sorry, but if you drop the ball at the Rose bowl, like you have faced adversity. I'm sorry, getting a D on a test is not adversity. Like, you know, but like, so I think like a lot, we like a lot of like athletes.

2:09:48

Yeah.

2:11:37

Or like people. Like we're really big in like arts or music or just, just doing something different. Being a straight A student is not different. Especially with grade inflation.

2:11:37

Talk about the differences culturally between the hedge fund guys, you know, and the venture capitalists. I was reading Dan Wong was talking about how in the hedge fund world you're wrong five times before breakfast because everyone has a different thesis and then it's immediately proved right or wrong. So you get more used to being contrarian and everyone wants to find that edge. Whereas in Silicon Valley there's a little bit more incentive to go with the herd because you can just ride the wave. What else have you seen and do you agree with that take?

2:11:47

Yeah. So hedge funds, it's got to get very, very generalist view. Extremely generalist view. Broad view. What I would say is I think if you were to take the best hedge fund managers I know and the best venture capitalist I know.

2:12:18

Mm.

2:12:35

We've been fortunate to like partner with a bunch of people, work with people, have people back us. I would tell you that in three month or six month increments, I think the smart, the hedge fund guys are smarter and like have a better sense of where the world is going. But I think over like a five to ten year period, the venture capitalists have a better view just in general, like how they think makes sense.

2:12:36

Yeah. Kind of aligned with the business. Business contrarian takes. I saw one here. The center of gravity is shifting. More innovation is happening outside of Silicon Valley with Countries outside the U.S. producing category leaders earlier in their life cycle. I read that, I felt like I disagreed with it. Defend it.

2:13:01

In a world if you believe AI is coming and it's easier to build software or build whatever. Well, who says you have to be in Silicon Valley? Like where it's really expensive to live, it's really expensive to get talent if you can build a. Build it anywhere. People have built now public companies that are completely, you know, Grafana, by the way, is a big business.

2:13:21

Yeah.

2:13:42

And it's completely distributed. There are other big companies, like if you talk to people like Dell And Microsoft, they will tell you that they've. They shut down a lot of, like, small, small regional offices and just have people work remote. So I think in actually, the more you believe in AI, the more actually stuff could come out of application layer. I think at the models. At the model level, it becomes about recruiting the best, like, you know, PhD tech level talent and stuff like that. But just to build application companies. I think you can build them anywhere.

2:13:42

Yeah. Yeah, that makes a lot of sense. Jordy. Anything else?

2:14:11

This is great. What's the update on the racing side?

2:14:15

Did you watch the Rolex 24?

2:14:19

Yeah. Any reactions?

2:14:21

Watched part of it.

2:14:22

Why weren't you racing in this Rolex 24 years?

2:14:24

Couple years. I got more years. I was. So race season. We are first race for Ferrari. Challenge kicks off in, like, the middle of March at the Thermal Club outside of Palm Springs, where you guys have been very cool. I'm actually going to race a support race at Le Mans this year. I'm going to. I'm going to race the Monaco vintage races in an old F1 car. My goal is to run Le Mans in the next, like, five years, like, 24 hours at Le Mans month.

2:14:26

Wow. That'll be amazing. Well, we'll be.

2:14:54

You know, the guys you got to have come on here, though, because you like cars. I don't really know him, but George Kurtz is like, oh, he's got him on.

2:14:57

He's a sponsor.

2:15:03

He's very good. Yeah, yeah, yeah, yeah.

2:15:04

No, his. His race at. At Daytona was insane. I mean, from getting taken out in that opening corner to getting the win.

2:15:06

Yeah. Absolute legend.

2:15:15

Well, F1's gonna be F1. It was funny. Door dropped. Russell was talking that he thought F1.

2:15:17

I thought.

2:15:22

I saw, like, a interview with him where he thinks it might be a lot closer this year because there's a bunch more teams that could be in the mix. It's still obviously very early on, but, like, the car that Adrian Newey built it at for Aston Martin is totally wild.

2:15:22

Yeah, we're putting. We're putting together a market map of F1 teams based on the tech companies that are sponsoring them.

2:15:34

So that's amazing, by the way, because there's.

2:15:42

I mean, I think the easiest way to understand motorsports are definitely a beneficiary of the AI Boom.

2:15:45

So far, you guys got to come to. You got to come to an F1 race.

2:15:52

We're excited. Yeah, we. We were in Vegas. It was fun.

2:15:55

Well, yeah, I'm assuming we'll be at all three American races.

2:15:58

I hope so.

2:16:01

Anyway, have a great rest of your day, Mitchell.

2:16:02

We'll see you guys.

2:16:05

Cheers.

2:16:06

Goodbye. Let me tell you about Graphite code review for the age of AI. Graphite helps teams on GitHub ship higher quality software. And I'm also going to tell you about Gusto, the unified platform for payroll, benefits and HR built to evolve with modern small and medium sized businesses. And we have some breaking news. Sam Altman has responded to the Anthropic ads. He says first the good part about the Anthropic ads. They are funny and I laughed. I like it. But I wonder why Anthropic would go for something so clearly disappointed. Honest Our most important principle for ads is that we won't do exactly this. We would never obviously run ads in the way Anthropic depicts them. We are not stupid and we know our users would reject that. I guess it's on brand for Anthropic doublespeak to use a deceptive ad to critique theoretical deceptive ads that aren't real. But a Super bowl ad is not where I would expect it. More importantly, we believe everyone deserves to use AI in our community committed to free access because we believe access creates agency. More Texans use ChatGPT for free than total people use Claude in the United States. So we have a differently shaped problem than they do. If you want to pay for ChatGPT plus or Pro, we don't show you ads. Anthropic serves an expensive product to rich people. We are glad they do that and we are doing that too. But we also feel strongly that we need to bring AI to build billions of people who can't pay for subscriptions. Maybe even more importantly, Anthropic wants to control what people do with AI. They block companies they don't like from using their coding product, including us. They want to write rules for themselves for for what people can and can't use AI for. And now they also want to tell other companies what their business models can be. We are committed to broad democratic decision making. In addition to Access, we also are committed to building the most resilient ecosystem for advanced AI. We care a great deal about safe, broadly beneficial AGI and we know the only way to get there is to work with the world to prepare. One authoritarian company won't get us there on their own, to say nothing of the other obvious risks is a dark path. As for our super bowl ad, it's about builders and how anyone can now build anything. We are enjoying watching so many people switch to Codex. There have been 500,000 app downloads since launch on Monday. And we think builders are really going to love what's coming for them in the next few weeks. I believe Codex is going to win. We will continue to work hard to make even more intelligence available for lower and lower prices to our users. This time belongs to the builders, not the people who want to control them. So he fires. Fires back.

2:16:07

Yeah, it was. It was more responses. This is why I said in general, like again, I think Sam has to admit that the ads are like pretty entertaining and it's just such a wild move and, and unexpected from anthropic given that they've been trying. I mean, I'd say from a brand standpoint, going with the sort of edgy adult humor was risk unexpected. But yeah, ultimately it's deceptive. Like they're trying to Mislead people about OpenAI's ad product.

2:18:41

Yeah, it's not corn syrup.

2:19:16

Like the entire strategy of the campaign is clearly not to drive downloads.

2:19:17

It's fud.

2:19:21

It's fud.

2:19:22

Yeah.

2:19:22

And yeah, I think there will probably be a Harvard Business case study on this.

2:19:24

Maybe we should do TBPN business case studies. I think that'd be fun. Should generate those also. It's just anti ad, which I don't like.

2:19:31

I like ads.

2:19:39

Here's another ad label box. RL Environments, Voice Robotics, evals and expert human data. Label Box is the data factory behind the world's leading AI teams. And we have our next guest in the Restream.

2:19:40

Is it time to puff?

2:19:53

It's time to puff. We got Simon from Turbopuffer. Let's puff in the Restream waiting room. Let's bring him into the TVPN Ultra realm. Simon, how you doing? Good to see you.

2:19:53

Good.

2:20:03

Good to see you guys.

2:20:03

Hey, John.

2:20:04

Hey, Jordy. It's time to puff.

2:20:05

It's time to puff. What's new in your world? Were you thinking about doing an over the top attack ad on a competitor or are you locked in, focused for this Super Bowl? What's the strategy? How is the business growing?

2:20:07

I mean, you were so generous to give us a front and center logo to the super bowl as a new entrant. Yeah, as a new entrant to the new world. I can't. I don't really understand American football.

2:20:19

But.

2:20:36

It seems right up your alley. Because when I think of it from first principles, it seems like a sport that has been designed to show ads.

2:20:36

Yes.

2:20:43

Which I think that is a great American invention. So I have to respect that. It's very beautiful.

2:20:43

Yeah.

2:20:49

How has the agentic coding boom been processed to you. What's the effect? Are you seeing any slowdown? There's been hype waves and sell offs and all sorts of turmoil. Has the recent sort of re acceleration taking you to surprise? Did it hit you like a flashbang?

2:20:51

Whatever is going on right now is hitting us like something that's good.

2:21:13

What is going on? Break it down for us. You've been waiting to use this effect all day. Yes, all day. And the chat demanded it anyway.

2:21:18

Bobby says billions.

2:21:31

We need a billion puff.

2:21:32

Yes, we need a puff one. We need a puff one. Like a puffer F. It's sort of just like, you know, when you want me to get off camera, you get a puffer fist. That's counting down.

2:21:34

Well, we have smoke grenade. We have a smoke grenade that kind of issues puffs of smoke. But maybe they need logos all over that. That would be good.

2:21:42

I mean, look, the world, the world's puffing. I think you asked about coding. I think coding was the first vertical that really started puffing really hard. And I mean, in general, the way that we look at the world is that we can certainly take all of the knowledge in the world and we can distill it into a few terabytes of weights. And those terabytes of weights are extremely useful at reasoning over data. But in some way they have to reason with the data. And to reason with the data you need a search engine, right. That has all the canonical data and is that tool to search through the data. I mean, that's what cursor and cognition and other coding tools are doing with turn. Turbopuffer is to search through the code. And so I think we're seeing that in other verticals now. We see it in legal, right. And we see every vertical sort of starts by just pushing a bunch of things into context. But then really what they want is to allow the agent to search by itself. And that's where turbopuffer comes in as the search engine. It can index petabytes and petabytes of data to allow these agents to search.

2:21:48

Yeah, I mean we're talking about ads. Is do you have a burgeoning ads business? Not ads in TurboPuffer, but powering ad tech systems. Because I can imagine a lot of new platforms. They have a lot of content, they have a lot of ads. They need to match those. And so your system should be valuable there. Is that a stretch or is that a good application?

2:22:49

I think most of the ad businesses already have extremely sophisticated systems for this. I'm thinking of New ads generally it's. Yeah, yeah, no, we haven't seen that vertical really come alive for us yet.

2:23:11

Yeah, I imagine it becomes important. And what about other sort of applications or sectors that are growing in terms of this technology adopting this technology puffing broadly?

2:23:22

Something that we're seeing puffing pretty hard right now is that people want to basically build their own Google. Google. They want to take the entire web and make it searchable and allow the agents to reason over either literally the entire public Internet or data sizes that are the internal data that's of that kind of volume. Right. Hundreds of billions of documents. And recently we launched some product to support that where you can basically now build your own Internet index on top of turbopuffer for an extremely reasonable cost. So what we showed is that we could take 100 billion documents and make them available to you it at P50. So that's the median latency of less than 50 milliseconds. So this is pretty remarkable for an off the shelf SaaS product and something that we're seeing that some of the most sophisticated customers in the world need. And I don't think there's any other easy way to do it than to puff.

2:23:41

Yeah, we talked a little bit last time on just compute bottlenecks, AI bottlenecks. Anything that you're that's keeping you up at night night. In terms of data center build outs, semiconductor build outs, even if it's not directly related to your business. Is there anything that you see that could potentially put a damper on the growth of the booming AI industry?

2:24:34

I mean I think everyone is struggling to get compute when you get enough skill. And I mean it's good news in CPU like land. I'm not contending for the GPUs yet, but we are starting to see, I mean you have to do that at any scale. Right? You have to start requesting from the cloud providers, hey, I want hundreds and hundreds of this machine in this particular region for these customers. So I think that's something that everyone's facing. Obviously we're watching the DRAM prices and things like that like everyone else, but hopefully that comes down.

2:24:59

Are there any kind of like nar. You have an interesting view into the usage of pretty much every popular AI product. Obviously you can't talk about individual, the sort of usage of any individual product, but is there any sort of narrative violations that you're seeing broadly every single day depending on new products getting announced, people are saying, you know, it's over for this company, it's over for that company. Or I just turned from this product. But overall, what are you seeing?

2:25:34

I think, I mean, what we see is that the companies that are building the most exciting product are trying to operationally, very, very large amounts of data, and that's difficult, difficult to copy. And so, I mean, we have survivorship bias in the businesses that we see, but they come to us because they want to connect more data to AI than anyone else. I would say. Yeah, if you're a thin layer and you're not trying to vacuum up a lot of data, then it starts to get really difficult. So, I mean, I do see as with how powerful the models are now, you have to be investing in some pretty deep tech tech to make it.

2:26:05

Yeah, yeah. Are you doing anything on the front of like, making TurboPuffer intelligible to agents? So if someone goes to a vibe coding app or a CLI tool and they just say, solve this problem for me, the agent pulls turbopuffer off the shop.

2:26:41

I mean, we try as much as we can to make the docs very easy to read. I think it's not completely clear how the LLMs make the decision of what the best database is, but I think that we're trying to do what we can to make sure that the LLMs like puffin as well. We have made API decisions around what it would be easier for the LLMs to catch.

2:27:05

Yes.

2:27:33

Yeah, yeah. I heard an interesting anecdote about someone building a CLI tool. They wanted to make it easy for an LLM to grab or an agent to grab. And so basically they just rewrote the CLI to include every possible hallucination or mistake that could be made. And people often do this with URLs where they will say, well, some people might type the wrong URL sometimes, but. But LLMs hallucinate in a slightly different way. They don't necessarily fat finger a single key, but they might drop the wrong token or use the wrong phrase or use the wrong term. And so they rewrote their CLI to include every possible permutation that an LLM could think about in terms of update, edit, quit, exit, all the different keywords. They just created hooks and functionality for all of them. So even if the LLM shows up, up and is like kind of clumsy, it can still use it and so it still likes it. But I don't know. I don't know where all that goes on the commercial side. This is more for, like, open source.

2:27:33

Yeah, I think, I think maybe it's a little overblown how much you have to design for this. I think it's the same with humans. We fat finger stuff. And the LLMs are a lot more resilient now than they were even just six months ago. So to me, it's just always good API design. If you're designing a good API or good CLI or good interface, you should sort of be able to predict like, oh, if I did this, it's probably going to work this way. And I think the LLM treats it exactly the same way. So I would maybe take back that we're doing anything super intentional for the LLMs, because I think the same design that's good for humans is good for LLMs.

2:28:36

Yeah, yeah, that makes a lot of sense.

2:29:09

What, if anything, has changed since the funding round?

2:29:12

Yeah.

2:29:15

Not much has changed. We got logo rights to Ramp, so that's exciting.

2:29:17

Congratulations.

2:29:23

Second ram mention of that.

2:29:24

They were like, finally, you raise a bit more venture capital, we'll let you use our logo.

2:29:25

There you go.

2:29:29

I mean, you know, it's, it's in, it's. Yeah, I think we're really, really happy to work with that team. And I think there's lots more logos waiting beneath the surface to be on earth. So the murderous row will continue to evolve as we are, earn the trust of more and more of the lovely companies that we work with.

2:29:31

Early on, when companies were discovering TurboPuffer and realizing, hey, this is really good and this is going to be sort of key to enabling all these different features and functionality that we want, was it actually a challenge getting logo rights because they didn't want their competitors to be super aware of it? Right. You want, like, hey, if I can have an even an extra month before for another company that I might have some overlap with, I'll take it.

2:29:55

Yeah, I mean, we have some verticals in. There's a hedge fund that doesn't really want us to talk about because they see it as a competitive advantage. Right. And so, I mean, we like that tension for sure. I think honestly, in the beginning it was just. I'm sure that some of the companies that bet on us early were like, okay, this is like three or four people in Canada and we're betting our whole business and search on them being able to make it. And I've certainly heard from one of our early customers that that kept them awake at night. Now the company is ten times that size. Right. And of a different caliber, and we owe them everything to have bet on us that early on. But it was certainly a challenge in the beginning, not just because they saw it as a secret stake sauce, but also because they were afraid of what their customers would think.

2:30:21

Right.

2:31:15

Here's this, like little like dinky company in the woods of Canada. Like what do they know about writing search engines? I think they were concerned with their customers thought. And we've just tried to do everything that we can by keeping our uptime as good as possible and walking everything carefully and building good software to earn that trust. But I think every startup that gets logos is. I think the founders have to really fight for it and they have to build the relationships to make sure that people trust they're going to use the logo in good hands and they're not going to plaster it all over contexts that they wouldn't feel okay with.

2:31:16

How are you guys approaching vendor selection today? When are you running the calculus of should we build this ourselves or pull something off the rack?

2:31:50

We were just discussing this recently where we've started our data stack now is just using cloud code or cursor agent in a repo and just giving as much information as possible. What's going on? We're starting to more and more ditch just the tools because it's so easy to compose the charts and things like that with agents. So the shallow SaaS software, I think it's dire there because it's just so easy to replace it now. But I think there will also be a bit of a hangover from that. Right. There is real value here in like permissions and enterprise, like all these enterprise features. So it will probably go very much in the, in the direction of everyone's going to build their own little shallow thing and then some of them are going to realize that this is a, this is a bit of a nightmare to maintain and then pay for it Again, could be dire for a little bit, but I think that, that I don't think it's going to. To swing. I still think there's a lot of value in that.

2:32:02

Yeah. What's the worst logo crime that you've ever seen someone do? Is that just go in the customer database search? Has any employee from a hyperscaler signed up for a trial account and then slapped the logo on the landing page? What would you tell a young founder not to do in terms of throwing a logo on a landing.

2:33:00

Landing page?

2:33:23

I think we, we like. I think logo crimes is actually great diction. I think we see this a lot where you see like just you go on a page and it's like Microsoft, right? And it's all these, it's like, okay, like we were talking about this internally because. So for example, there's a, there's one small team inside of Vercel that uses Turbo coffer.

2:33:24

Yeah.

2:33:44

And I think a lot of companies might then just put the Vercel logo on the website and say vercel's Puffin. But I don't think that's fair. Right. It's an auxiliary use case. It's one person who's using it for go to market and doing a phenomenal job over there. And so what we're trying to do is like, okay, that's that that logo might not make it to the front page because that would not be an honest association. And then on the use case page we're going to be having attacked this as core and auxiliary. And then under accelerator ancillary, right. It will say, okay, it's this particular team at this enterprise that's using it just to be completely honest about it. And I think that's. And then I think just sending a note to the teams that you work with of like, hey, is this a cool place to use your logo? You have to be very respectful about it.

2:33:45

Yeah, just ask for permission.

2:34:29

But you can't.

2:34:29

Yeah, you can't speed run it.

2:34:31

It takes the chat saying, vercel's puffin.

2:34:32

Breaking news. Put up a TBPN trading card. You got the scoop here. Yeah, we love gear.

2:34:35

Case study will come out. But as I said, it's an accelerator use case and that's still cool. And we love the trust, but it's not like v0 is puffin or something like that.

2:34:43

Yeah, no, that makes a ton of sense. Well, thank you so much for hopping by. We'll talk to you soon and have.

2:34:52

A great rest of your day. While we were live, Google reported earnings. Revenue grew 18%. Search 17%. Gemini has 750 monthly users and the stock is absolutely nuking. So say, just don't, don't look at your portfolio. Stay in Slack.

2:34:58

It's just barely over a 4 trillion dollar company though. So rough.

2:35:15

When are you coming to California? Next hours or Southern California?

2:35:21

I should say I'm going to California next week. All right, when are you coming up to Canada? I keep like every time we talk about this and then you guys are like, oh, it's so cold.

2:35:25

The whole country. I just feel like the whole country is un American from my perspective. It's just, I mean, have you met American?

2:35:36

Simon? I'm right here.

2:35:43

Okay.

2:35:45

I'm ready for you.

2:35:45

Okay, you got me. You got me.

2:35:46

Well, we would love, we would love to go Fishing.

2:35:48

Yeah, we need to do a weekend. Weekend trip. It's too hard to bring the show, honestly.

2:35:52

Yeah, family trip.

2:35:56

Yeah, that'd be great. Let's do it.

2:35:57

Realistically, it's great bringing everybody.

2:35:58

Awesome.

2:36:02

You dump them in the lake and you have a great time.

2:36:03

Risky with some one year olds, but we'll give it a try. Anyway, thank you so much for stopping by. We'll talk to you soon.

2:36:08

Great update.

2:36:15

Goodbye. Let me tell you about phantom cash. Fund your wallet without exchanges or middlemen and spend with a phantom card.

2:36:16

Chat says we need red suits.

2:36:24

Red suits. Yeah, we were talking about that today because we've been wearing the white suits a lot and now it's time for the red suits. We're going to have to call the Taylor.

2:36:27

Says Alphabet sees 2026 capex 175 to 185 billion versus 115 billion expected for the past decade. Mag 7 were free cash flow monsters who poured money into buybacks to support their stock price prices. That era is over.

2:36:37

Well, let me tell you about Vibe Co, where DTC brands, B2B startups and AI companies advertise on streaming TV, pick channels, target audiences, measure sales. Just like on Meta. Pavel says SaaS is dead, is probably oversold and sleepy companies get wrecked in technology shifts, is probably undersold. I think that's a good take. There's a lot of people talking about the SaaS. Pocalypical Patrick O' Shaughnessy shared a clip from Gavin Baker on Invest like the Best, saying why it's a mistake for SaaS companies to resist AI because it has a lower margin structure. We can pull this video up. When there's a transformative new technology customers are demanding, it's always a mistake not to embrace it. Let's play this clip. I wanna see what he said.

2:36:56

Application SaaS companies are making the exact same mistake that brick and mortar retailers did with E Commerce. So brick and mortar retailers, they looked at Amazon and they said, oh, it's losing money. E Commerce is going to be a low margin business. And so they did not invest in E Commerce. They clearly saw customer demand for it, but they did not like the margin structure of E Commerce. That is the fundamental reason that essentially every brick and mortar retailer was really slow to invest in E Commerce. And now here we are and Amazon has higher margins. So margins can change. And if there's a fundamental transformative kind of new technology that customers are demanding, it's always a mistake not to embrace it. And that's exactly what the SaaS companies are doing. They have their 70, 80, 90% gross margins and they are reluctant to accept AI gross margins. The very nature of AI is software. You write it once and it's written very efficiently and then you can distribute it broadly at very low cost.

2:37:41

Such an underrated point that the CFO just knows. My company has 30% gross margins. So that's what I'm building my entire business around. My company has 90% gross margins and that's how I'm informing every business decision. When the whole structure that you've built the foundation of your business on changes that can be a little tricky quickly. Yesterday we went to the Cisco AI Summit. I'm sure you saw it. Let me tell you about Cisco. Critical infrastructure for the AI era. We have a lot of fun talking to the team at Cisco. They were very hospitable. I had a great time, absolutely enjoyed talking to everyone. And we had a really wild, wide ranging conversation with Dylan Patel at the end. That's in the RSS feed. If you didn't get a chance to listen to the full thing. We go all over the place. Space, data centers, what Google's doing, what Apple's doing, what Microsoft is doing. We went everywhere. It was really, really great. Anyway, we can go through some more of the timeline. What is this? Omg. Cap says he died doing what he loved. Aggressively buying the dip on SAS. Like a guy averaging down on Blockbuster Video in 2010.

2:38:39

I'm averaging in.

2:39:48

Averaging down to zero. That is not good. There are a whole bunch of different takes on the SaaS apocalypse. Pierre Richelson has one. Why would I pay for SAS if I can prompt the software and run it myself? My brother in Christ. Have you heard of open source businesses? The last thing people want to do is be in charge of development and maintenance of software. You got to get an intern. It's a bull market in interns. The interns are back. The interns. You heard from Mitchell Green over at Lead Edge? He's hiring interns to build stuff for him. It's a bull market. Interns.

2:39:50

Everybody wants to be the one person billion dollar company. No one wants to be the billionaire.

2:40:23

Billion intern company of America. Yeah, that's the business. I hadn't actually heard that about Cloudflare. Over 1,000 interns. That's crazy. I didn't even know that about Shopify. That's very cool. Cluly got mocked for having 50 interns. Maybe they were just early. We at one point it felt like we had a ton of interns. We don't actually have that many interns. But interns are underrated. It's a good time and it's a particularly good time with the new tools to let someone loose with just a fresh start. They don't have to, they don't have to maintain the old system. They come in and they're able to start completely fresh with fresh tools. They can pick anything off the shelf. Whatever's the most cutting edge, that's what they bring into the organization. As opposed to saying, hey, you have to maintain this particular workflow or this particular technology or you have to plug into those rest of this ecosystem. Change management takes a long time. Interns allow you to kind of pull that forward, which is a lot of fun. Let me tell you about Cognition. They're the makers of Devon, the AI software engineer. Crush your backlog with your personal AI engineering team.

2:40:28

Unemployed capital Allocator says, wait, I'm confused. Is it the software engineers that are obsolete or software companies, or is it the software users or the software users, employers? Tae Kim says all of the above.

2:41:28

Yeah, the buchos is. You're not confused. None of these things will exist in five years. Everything, sell everything and quit the job.

2:41:42

It is interesting. Bloomberg had an article yesterday that shared that Google is expanding their footprint in India. Planning to.

2:41:50

That was a weird headline, right?

2:42:00

They'll have the capacity to add something like 10 to 15,000 new employees. And you know, I don't, you know, I would say like the, probably the catalyst there sort of changes or unpredictability to US immigration law potentially, potentially being a factor in that. But it was kind of looking at it from a Tyler point of view, just being like, hey, how AGI pilled is Google broadly is DeepMind if they're planning to scale headcount.

2:42:01

Yeah, the headline was from Bloomberg. It's millions of square feet. So that can accommodate somewhere between 10, 10,000 and 30,000 employees per million square feet. According to Gemini here, the industry standard is 60 to 80 square feet per person. I wonder how we're doing here in the TBPnulterdom. I feel like I have way more than 60 square feet to myself. Standard tech office is 100 square feet per person. Techno changes.

2:42:36

India is also going zero tax on data centers until 2047.

2:43:03

Whoa. What would the tax implications be of that property tax? You don't get, yeah, maybe no property tax because you can't possibly not pay income tax on the revenue that's generated from a data center. That would be a crazy, crazy thing. But we will see. But our next guest Is in the restream waiting room. We have KJ from Lotus AI. Welcome to the stream. How are you doing? Good to see you. I'm doing great.

2:43:07

Doing great. Good to see you guys too.

2:43:32

Thanks so much for hopping on. It's first time on the show. Please introduce, introduce yourself and the company.

2:43:34

Yeah, my name is KJ Dhaliwal. I'm the CEO and co founder of Lotus Health AI. Yeah, we are building an AI doctor powered by real doctors.

2:43:39

What's the go to market? How much do you like? Who do you want to sell to? Like the hospital network, the individual doctors, even the individual consumer?

2:43:48

Yeah, no, we're actually direct to consumer. So one of the big things that we believe in is going straight to the patient because innovation in healthcare sort of has always lagged at the hospital systems or the insurance companies. My background's in consumer. I built the largest South Asian dating app prior to this.

2:43:58

Oh, no way.

2:44:14

And so we just, you know, we thought, what are the things that people actually need?

2:44:15

Yeah.

2:44:19

America's struggling with healthcare, so we decided to just go direct to consumer.

2:44:19

Yeah, walk me through some of the. I mean the most basic could be, you know, diet plan, exercise plan, all the way to a doctor can refer you to a lab to go get an MRI and surgery and all sorts of stuff like where, what do you see as the early go to market, the landing zone, what do you want to be excellent at and then grow from?

2:44:23

Yeah, we're really focused on primary care today. So when you think about America today, 100 million people don't have primary care doctors. So that's sort of this huge sort of opportunity to, to really give people, you know, basic care. So that means, you know, sometimes you need a prescription, sometimes you need a referral or even a lab order. If you're going to go see a specialist like a cardiologist and you don't have a lab ready, that oftentimes that visit is a waste of time for the cardiologist and for the patient.

2:44:44

Yeah. Right.

2:45:12

So we can do a lot of that preliminary primary care virtually.

2:45:13

Yeah.

2:45:16

And it turns out 80% of all care in general is actually no possible virtually within the primary care setting.

2:45:16

How important is image processing? You know, oh, I got a mole like you look at this, like that's like a cloud or I got a scab and it's not healing and is it infected? Is AI ready to handle that sort of use case?

2:45:23

Yeah, definitely. So, like, you know, we're seeing, we're already seeing AI being implemented in obviously the imaging Space and the MRI space. But what we're focusing on, focused more on, is really getting the information from the patient, distilling it and summarizing it for a clinician to look at and make the final call. Sure. So that's really where Lotus is able to give Americans free primary care, essentially because we've brought the cost of care down by a factor of 10.

2:45:37

Yeah.

2:46:04

So you can upload images, you can, you know, chat with Lotus for hours. It'll do the intake similar to how you would answer questions on that clipboard when you go see a doctor. Doctor, and then the doctor can review all that quickly. And the big thing sort of, that we did is we sort of created this, what we call the personal health record, which is the richest longitudinal record on a patient. So when you sign up, the first thing we do is we ingest all your health data. So that's from all the ehrs, every doctor you've ever seen, your wearables data, your insurance claims data. That allows us to build sort of the foundation of the patient and then we can really unlock a lot of the sort of agentic workflows and the clinical workflows. On top of that, I just make care delivery a lot more accurate at scale.

2:46:05

Years ago, I remember this company, zocdoc was sort of like help you find a doctor nearby, almost like a yelp for doctors. I don't remember what their business model was, but is there any overlap in or learnings from that type of business where maybe there's a referral fee to a specific doctor? Does that make any sense? Or do you just want the customer to eventually pay or ads like, how do you see monetization evolving?

2:46:51

Yeah, zocdoc did, I think something interesting which they, they would basically have doctors pay some sort of affiliate fee to be listed on their platform. But actually getting paid for referrals in healthcare is, is a big no, no and kind of illegal. That's why doctor, you know, shouldn't be paid actually for referrals because then you can be incentivized to refer to a particular, you know, part. And so what we really believe in is in terms of, if you think about the largest tech companies in the world, they're consumer companies and they make money through sponsored content. So we actually think if we focus on premium content within fitness and wellness, which is a, you know, multi trillion dollar sort of industry already, that's sort of how we'll be able to monetize. But really what we're focused on today is bringing the cost of the care down.

2:47:16

So anthropic's attacking you like you're the. They weren't talking about OpenAI, they were.

2:48:05

Talking about you respond.

2:48:10

No, no, no. But I mean, how you, you are going to have to have the talking point that responds to that ad. I'm sure that super bowl ad will blow over, everyone will talk about it in the ChatGPT context, but you're going to have to have a strategy for how you don't do the hey, we're recommending lifts or we're being weird or creepy with your ads. How do you think about messaging and setting the tone, the mission, the values of the company now so that you never have a PR crisis in the future?

2:48:13

Yeah, absolutely. And I think that's so important to do that. Right. You know, trust and safety and privacy is very foundational to what we're building here at Lotus and being, you know, we want to not only stand for the, the patient, but also for the doctor. Right. And so when you look at the industry today, doctors are burning out.

2:48:42

Right.

2:49:00

We don't have enough doctors, so we're really giving them the technology to supercharge their capabilities. And yeah, in terms of, you know, advertising, it'll be optional and you know, there'll be a premium subscription. If you don't want to see ads, you can sort of opt out. Or there's other avenues we're exploring. Like your employers. Right. A lot of large employers that are self insured struggle to get their employees really good health care or at least primary care. So the employee has to take time off for work, you know, drive an hour, go see a doctor. Half of the times they're not seeing the doctor or seeing a nurse or a PA because we have such a shortage and it ruins their day and it, you know, reduces productivity for the company. So we have companies reaching out to us saying, hey, like we'll pay you 50amonth. Can you just give this to our employees for free? You know, so there's a lot of different revenue models we'll explore, but really what we're focused on is, you know, how do we get to a million, 10 million patients and grow this company to scale? Because that's really how we think this sort of primary care crisis can be fixed. Essentially, fee for service doesn't work in primary care. Right. And even value based care doesn't work because we don't have enough doctors.

2:49:00

Yeah. How are you reacting to the major LLM providers potentially going into this category? A lot of them have already made announcements. They've rolled out features, they have massive user bases. What's going to be the key differentiator that allows you to go the long haul?

2:50:05

Yeah, no, we think that's actually net positive for the, for the industry because, you know, people are starting to trust AI with their health data. The one thing they can't do is they can't actually treat patients. Right. So we can actually, because we have real licensed clinicians. Sure. Giving treatment, we can actually close that care loop. Right. We can give you that prescription, order the lab, refer you to a special professionals, give you a diagnosis. And, you know, in fact, you know, one of our investors is an executive at OpenAI, and so we've been sort of following that sort of product that they've been building for a while. But I think it's net positive. I mean, I think there's, you know, there's, you know, healthcare is a $5 trillion market in terms of spend, and a lot of that is just waste. Right. That can be sort of cut out and taxed.

2:50:21

Are you. Do you think Lotus is, is really a threat to urgent care where, let's say somebody has like a skin issue and they're used to the flow of like, I want immediate care. I want to get, like, if somebody gets, I don't know, like poison. Bad case of poison oak and they want, they need to get some treatment for it. They'll go to an urgent care, be able to get like a prescription or a steroid or something like that. Whereas with Lotus, you could just get, you know, basically, as long as you have your phone, you could immediately, like, get treatment. Treatment. Is Lotus threatening to those kind of businesses that rely on people just needing, like, convenience?

2:51:06

Yeah, no, actually, it's the quite opposite. Because what those urgent care centers today, and even emergency rooms, hospitals, their biggest problem is they're overwhelmed with patients coming in that don't need to come in for things that can be treated through telemedicine or telehealth.

2:51:45

Right.

2:52:02

We actually refer you to urgent care because we obviously recognize and our clinicians recognize that there are obviously certain things that can't be treated virtually and need to be done in person. And that's where we refer you to inpatient or urgent care sort of care where the doctor needs to touch you or do a procedural thing. But it turns out, like I said earlier, 80 to 90% of care can be done virtually, and that's sort of the big bottleneck. So what ends up happening is all these patients end up going to urgent care or, you know, hospital ER rooms And then the patients that actually need that care end up not getting it because, you know, these systems are overwhelmed today.

2:52:02

Well, we are in the Lambda Lightning round. You raised some money. Tell us what happened, what's the deal?

2:52:42

Yeah, we just raised, you know, a total of 41 million in our seed and series.

2:52:48

Congratulations.

2:52:56

Leiner Perkins and CRV Code led that round. And then for great names, very proud to have these guys.

2:52:58

Well, congratulations on all the progress and I'm sure we'll see you back here soon. Great to meet you.

2:53:06

Cheers.

2:53:10

We'll talk to you soon. Goodbye. Let me tell you about CrowdStrike. Your business is AI. Their business is securing it. CrowdStrike secures AI and stops breaches. And we will continue our Lambda Lightning round with with Nick Sharp from adeo. He's the co founder and CEO.

2:53:10

Welcome back.

2:53:27

Hey guys, thanks for having me. Excited to be here again.

2:53:29

Yeah, good to have you back.

2:53:32

Give us the update on the last few months. I feel like the timeline has been loving Adeo. Feels like there's a ton of momentum.

2:53:33

Yeah, well, that's good. We've had a. It's the end of a busy day here in London and it's been time. A long, long one, but an exciting one. So I think it was about five months ago that I was on the show when we raised our Series B. And since then we've been busy deep in R and D land and today we're showing the world the fruits of our labor and launching a product called Ask Atio.

2:53:44

Amazing. Break it down.

2:54:08

Yeah. So it is a entirely new way to interact with Atio with your CRM. And it is a new conversational AI interface which essentially does. Does something which has been the holy grail for CRM for a really long time, which is making sense of all of the ton of data that you're generating all the time. That's calls, that's emails, that's all of the interactions you're having with customers, product data, etc. And essentially makes it intelligible and allows you to take action on it. Good example might be I now run a workflow at the end of every day which essentially goes through all customer calls, all customer emails and just flags things that are important for me.

2:54:09

So it's like customer feedback or a place that you need to step in and have a conversation with a customer yourself, et cetera.

2:54:53

Exactly. But imagine we have like 30, 40 people in our customer facing team now. So to do that previously when we have hundreds, thousands of customer interactions a day would be impossible. And so, yeah, it's a very, very exciting step for us.

2:55:00

What's been your reaction to. It feels like even this year a lot of people have been saying AGI is here and yet there's a bunch of net new CRM companies being formed. You guys have the benefit of having been around long enough to have a very strong foundation, yet not necessarily be fully submitted in your ways and still a lot of functionality. And so it feels like you were already an AI native CRM. And so it must be kind of funny to see other CRMs coming in that are maybe trying to claim they're more AI native.

2:55:17

Totally. And what's funny for us is that there seems to be a level of consensus now that something's going to happen in this market, which has not been the case. Right. So the consensus seems to have been building more and more and more. And in the last few months we've just seen a, we've, we've seen things fully flipped in the other way where everyone now believes that these kind of incumbents are going to be disrupted. And so we're, we're excited by the, by the, by the excitement of everyone else and the kind of belief in the market. And you know, we, we have the job now of making sure that we continue to, to be at the front of the pack, which is sort of lots of short term, just being very paranoid in the short term, very optimistic in the long term and just kind of keep building, keep going.

2:55:53

What are you paying attention to on the sort of sales agent side? Just in general, this is a big opportunity for adeo, but I know you guys will want to integrate with a bunch of players as well. So far we've been hearing people are using a bunch of different, different agents on that front. But what are you seeing?

2:56:42

Yeah, and the way that we think about this is I think when the market is, when there's as much innovation and turmoil, etc. As we're seeing right now, you kind of want to see, you want to see how things play out a bit. So we're taking some pretty big bets, but equally we're not building a closed platform and we're letting, we're, we really want people to integrate with our platform and we want our customers to go in whatever direction is right for them and make sure that we can support that. So we took a huge bet on ecosystem and building a really strong SDK. We've kind of quietly put out an MCP server in beta, so. Mcp.atio.com MCP but we're basically, CRM is always going to be the center of a GTM stack. And so as well as doing our own things in agentic workflows and that kind of stuff, we've also taken a huge bet on supporting all the other great companies that are doing stuff there as well.

2:57:02

Awesome. What can we expect from you guys this year? You're calling it now, this is the last thing you're going to ship for the year. No kidding.

2:58:05

Yeah, exactly.

2:58:13

Job done.

2:58:14

Well, for us, this is day four of FY27. So we're just at the beginning and honestly it's going to be a bit of an onslaught. We have a huge, huge amount coming. We're going to follow up with this launch pretty quickly in just over about a couple of months with another really, really big launch. And then of course we're in this interesting phase as founders or as company builders or whatever, where you now have to do two things. You've got to scale your company and you've got to hit all of your really ambitious targets. But at the same time, everything needs to be rebuilt every three months. So the way that you do go to market, the way that we think about product, et cetera, we're in kind of deep R and D mode across the company. And so you're trying to be simultaneously be a extremely scrappy early stage company, but then also trying to kind of compound growth and serve a growing number of customers. So we're up to about 7,000 customers now, which is triple this time last year. And so, yeah, lots of things going on at once.

2:58:17

How are you thinking about outbound? Actually sending emails, going into more of the AI agent for sales rep or place a sales rep. That's a space that feels like people are trying stuff but nothing's really caught. How far away are we from that?

2:59:35

The interesting thing is that.

2:59:56

You might.

3:00:00

Have seen this, the ramp post that was making rounds. Yeah. And these channels are getting saturated and it's a constant game of cat and mouse.

3:00:01

Right.

3:00:09

Because the more emails that get sent, the better the Gmail or whoever else is detecting them and shutting them down, etc. So we take an ecosystem bet on those things. Again, especially when it comes to outbound, we really want to be, we want to be the place where the customer context live, where, where your team live, et cetera. We integrate with a lot of these players now over time you want to be able to provide your customers with a pretty out of the box solution to a lot of these problems. So never say never. But right now we're mostly focused. And when it comes to outbound, we're focused on just partnering with the best people we can.

3:00:09

Amazing.

3:00:49

Awesome.

3:00:50

Well, thank you so much for taking the time to come.

3:00:51

Thanks for staying up.

3:00:52

Thanks for staying up late. We appreciate it.

3:00:53

Long day.

3:00:55

Long.

3:00:55

Yeah, exactly.

3:00:55

And we will.

3:00:57

Thank you guys.

3:00:57

Yeah, we'll see you back on soon.

3:00:58

Yeah. Have a good one.

3:01:00

Cheers.

3:01:01

Goodbye. We have one last story that we should get through before we get out of here. Citadel's Ken Griffin says Trump White House has, quote, enriched family members. The Wall street figured Republican donor offers rare public criticism of perceived administration self dealings. They're calling him Ken doesn't like Grifton, the billionaire investor. Sick of Grifton, he said. This is a direct quote from Ken Griffin, the founder of the hedge fund Citadel. He says this administration has definitely made missteps in choosing decisions or courses that have been very, very enriching to the families of those in the administration. He said at a conference in West Palm Beach, Florida on Tuesday that was hosted by the Wall Street Journal. He added, quote, that calls into question is the public interest being served. In response to Griffin's comments, a White House spokesman said, quote, the only special interest guiding the Trump administration's decision making is the best interest of the American people. The fact that major stock indexes have hit multiple all time highs, real wages have grown and inflation has cooled since President Trump took office is proof that this administration is delivering for every American Trump and his family members have profited since he took office last year. This is the Financial Times. An FT investigation in October found the President's rapidly growing cryptocurrency empire had already reaped more than 1 billion in pre tax profits over the prior year, partly attributed to a digital currency boom buoyed by the White House's own crypto friendly policies. Companies backed by Trump's sons have been awarded contracts with government agencies and benefited from administration policies and cryptocurrency and prediction markets. And a lot of this goes into the decision making to sell Nvidia chips to the uae. Right.

3:01:02

And yeah, I meant to ask Dylan about that yesterday. This idea of like do you think any lab CEOs or hyperscalers were looking at that? The reporting that came out over the weekend that the spy sheik in the UAE had bought half a billion dollars of world or sorry, half a billion, invested half a billion into World Liberty financial. Yeah, bought 49% of the company right before the inauguration and quickly we approved a pretty massive sale to China. And I wonder how many people were kind of reading that story and being deeply frustrated just given, hey, like, I would have.

3:02:47

Oh, I would have bought that I.

3:03:25

Was a happy buyer.

3:03:27

Yeah, yeah.

3:03:27

Trump of course, said. Trump, of course said he didn't know anything about. Anything about the investment.

3:03:28

Yeah, yeah, yeah. There's a big question. I mean, obviously the Wall Street Journal, I think, had a piece on the timing being suspicious. The Trump administration's denied that there's anything wrongdoing going on. Maybe there'll be an investigation at some point. But the question that's in my mind, aside from that one, is just, is the UAE suitable for selling GPUs too? Because if it's something that's really obviously a thumbs up and everyone's happy about, then it's a lot less, less controversial to do it, I would think. Because we're now at a point where many folks that I look to for guidance on like the geopolitical chip issue have come around to the idea that maybe we do want China to actually be buying Nvidia chips. And there's maybe some good arguments there. Initially it was like, cut everything off. Now the ball has moved to don't sell them SML equipment, don't sell them lithographic equipment.

3:03:36

Yeah. Come around to, we want people to stand in AI, to be on the American AI stack. Is it better for the UAE running on Nvidia versus Huawei?

3:04:29

Yeah. I mean, yeah. Dylan's point was like, not this. These things don't happen in isolation. So if you, if you don't sell Nvidia chips to China, it's not that they're going to, you know, retaliate specifically with, with Huawei or SMIC or SME. It's that you could like, they could like, you know, know do something in a completely different part of the world. Like Dylan's example was like, be more aggressive about Africa or something like that, like, or the rare earth elements. So there's like a million different debate points going on and all of those sort of come together. But with regard to the uae, I mean, I guess that there is, there's a risk that you send a bunch of Nvidia chips and they just stay on the, they stay on the container ship and then they just get forwarded straight along to, to someone who's maybe even less friendly. But I don't know that that's been a historical risk. There's been a lot of chip smuggling, a lot of it's gone through Malaysia. I don't exactly know how much of a risk the UAE poses. So I'd love to talk to Bill Bishop about how that fits in. We'll have to keep digging in, but obviously I think we're fans of Jimmy Carter approach to leading the American people.

3:04:38

Divest, divest.

3:05:54

Stay focused. Lock in. It's the great lock in. It's the great lock in. You got to stay focused on the. On the job at hand.

3:05:56

Anyway, anyways, crazy day. Mike Isaac over says by the number of OpenAI employees I see tweeting about anthropic super bowl ad, Anthropic should be paying OpenAI earned media fees.

3:06:03

That's funny.

3:06:17

Oh.

3:06:18

Also related to Ken Griffin. I like this post from High Yield Harry where it's MrBeast with Kim Kardashian and Ken Griffin's just way in the background. And High Yield Harry said a guy like me would have taken a photo with Ken Griffin instead. And I don't know who he's imagining him being. Is he imagining that he's Mr. Beast or Kim Kardashian? But it's funny to imagine being at this party and saying, I gotta get. I gotta get a photo with Ken. Gotta get a photo with Ken.

3:06:19

Whoa.

3:06:44

Okay, we gotta close with this. This G Wagon. Look at this G Wagon. You mentioned this.

3:06:45

Pull it up.

3:06:50

I did not pull it up. I did not see this. It's. It's right after. Yeah, look at this.

3:06:50

This. What is going on, Blake? I'm gonna pull up. I texted him to get some more context. So Blake runs an account called Found Objects on Instagram, and he is selling a very special G Wagon that's effectively like a. It's. It's one of 10 ever made.

3:06:55

Why is Vladimir Putin seen next to it?

3:07:14

Because he was. He was rolling it in the.

3:07:17

He was rolling in.

3:07:19

This was part of the presidential crazy vehicle. So it's basically like a Maybach converted with it with. Into a G Wagon.

3:07:20

Yeah.

3:07:28

In order to. You would have to do an insane conversion to get this into the US But. But for the right person, it could be a really good.

3:07:29

This has Mark Zuckerberg written all over it. Yeah, he's got the extendo. He's got the. Send it to west coast customs. This extended G Wagon is so.

3:07:37

Yeah, go to the picture at the very end.

3:07:47

Yeah, the picture at the very end.

3:07:49

Goes incredibly hard standing in front of it.

3:07:51

It's like, why is he not getting into it or getting out of it? He's just walking next to it. It's very, very cool.

3:07:54

So this can be yours for the lower price.

3:07:58

So, you know, the person that's seen this. This is not AI. This is considerable.

3:08:01

This is real.

3:08:05

This is real because there was an AI car caught on bring a trailer that caused a firestorm in the car auction community. The pictures were absolutely hilarious. You'd look in the car and the car was generated, I guess on cobblestones. And inside the car the AI had hallucinated more cobblestone so it looked like you had cobblestone. Cobblestone floor mats and you could, and you can imagine getting into.

3:08:05

So basically, before you pay, you know, 225k for a brand new G Wagon, first consider this 2010 G55XXL. It's a fantastic example and we'll end on a white pill. I guess we didn't need the red suits for Google today. Yeah, the stock has recovered. It's now up one one and a half percent after.

3:08:35

Investing in AI and the market loves it.

3:09:03

Logan also just tweeted, they just crossed 10 billion tokens per minute on Gemini and 750 million MAUs. I've seen enough. Give them $100 trillion valuation.

3:09:06

They're 4% of the way there. They just need a 10x and then 10x again and they'll do it. Well, I'm rooting for them. I'm rooting for the entire financial market. I'm rooting for all companies because I love business.

3:09:18

We're rooting for technology because we love you.

3:09:29

We love you.

3:09:31

And we'll be back tomorrow. We have a fantastic show planned.

3:09:31

Leave us five stars. Plant the bomb. I'll tell them about where to leave us a review. Leave us a review on Spotify podcast. Five stars. Please sign up for our newsletter tbpn.com, follow us on all our social media platforms and go to Yahoo. Why not?

3:09:35

Why not?

3:09:52

The Yahoo team, the marketing team, they were like, we need a jingle. We need someone to go sing it. This guy gets in the booth and starts singing and they're like, go harder, Yahoo. They're like, that's not hard enough. Goodbye.

3:09:54

Nice work, brothers.

3:10:10

I'll see you on the next one.

3:10:12