searchlore

0:01

[Music] [Music] Hey, hey, [Music] hey. Heat. [Music] [Music] Heat. Heat. Heat.

0:51

[Music] [Music] [Music] Double [Music] kill. Hey. Hey. Hey.

1:34

[Music] [Music] [Music] [Music] [Music] Are you guys doing other talents or this? Like, should I log out?

2:38

[Music] [Music] You came to this world behavior overnight success.

3:23

We came to this world to shape our [Music] future.

3:38

We dream to feel the new Feel the [Music] music.

4:09

The temple of technology. [Music] Let me take you.

4:30

[Music] [Music] You're watching TVN.

4:47

Today is Wednesday, May 14th, 2025.

4:50

We are live from the Temple of Technology, the fortress of finance, the capital of capital. Ben is back. Ben is back. Ben is back. Welcome back, Ben. Welcome back.

5:02

He found his happy place.

5:04

He found his happy place. It wasn't in a wander.

5:06

Although it could have been in a wander, but now it's in the studio in the temple technology. Ben is back, baby.

5:12

That was a huge milestone for the show. Yeah.

5:14

To actually survive for it was remarkable days without Ben.

5:21

We were on a shoestring, but we pulled through and he's back and we're excited.

5:25

Also, thank you to Sequoia Capital.

5:25

They sent us some beautiful candles. They smell fantastic.

5:29

highdimensional cardamom and violet embedding diffusing through hidden layers of spice back propagating to a musk and sandalwood base. So, thank you.

5:38

This is from their AI ascent conference.

5:41

They uh sent extremely well. Extremely well executed.

5:47

And they sent us a nice little note here.

5:48

It's from uh from Sonia and the team at Sequoa. Beautiful.

5:50

She says uh thanks for having me on the show to chat all things AI.

5:55

It's no Himalayan Birkin, but hope you enjoy the candle. So, thank you, Sonia.

6:00

We'd love to have you back on the show.

6:01

I really enjoyed the conversation. Yeah, it was great.

6:05

And we're gonna be talking to quite a lot of venture capitalists today, John. Yes.

6:09

It's a It's a gladatorial It's a gladiatorial battle.

6:12

It's a It's a one man versus 100 gorillas situation.

6:14

One show versus 100 Anderson partners.

6:17

Let's take a look at the lineup.

6:20

We do have uh four founders here, although one of them is raising a fund.

6:24

Uh Josh Browder uh you know from do not pay but he's announcing his fourth venture fund. That's pretty crazy.

6:30

We were just talking about how how few uh company how few firms get to fund four. He did it.

6:36

Uh we got the founder of Granola.

6:36

Uh Brian Johnson's coming in the founder of Don't Die.

6:40

We're going to get to the bottom of how not to die.

6:41

Don't die before he comes on. Uh yeah.

6:44

And Matteo from Eight Sleep is coming on announcing the Pod Five.

6:48

Let us talk about the Pod 4.

6:51

Now it's time for pod five and then we're going live to A16Z's limited partner day.

6:56

This is the day of the year that they check in with all of their investors.

7:00

So they're all together and we got them all calling into the stream one after another.

7:03

We got six GPS and the co-founder of the firm.

7:06

So that'll be a lot of fun.

7:08

Uh but first I wanted to set up the conversation because I think that if you're talking to Andrees, what's the most interesting tech trend that they seem to be they have a unique perspective on?

7:18

And from my perspective that is open- source artificial intelligence.

7:22

Um they Mark has been very outspoken about his love for art for the open source community.

7:28

They've backed a bunch of open-source uh AI companies and it's been you know received very well in the open source community and some investors love it, some don't.

7:37

And so my former colleague at Founders Fund, John Ludig, wrote a piece that we're going to be diving into all about uh the future.

7:46

He claims the future of foundation models is closed source.

7:48

And he makes the argument that open source is not ever going to fully win.

7:52

Not that it can't be helpful or useful, but we've also been having some other conversations with folks that say open source might be a tool of economic warfare that we need to be engaging with and maybe we haven't considered all the different angles.

8:05

And so, uh, his piece is from a year ago.

8:07

We'll see how it holds up today and then we'll talk to a bunch of the Andrew and partners about open source AI.

8:12

But I wanted to kick it off with uh Ben Thompson Instatey talking about OpenAI's restructuring.

8:16

We covered this when the news broke, but he kind of packaged it up in an update with a few different uh pieces of information talking about how Microsoft's involved talking about the hiring of Fiji and the acquisition of Windsurf.

8:32

And so he starts it off with a quote from the Wall Street Journal.

8:35

says, "OpenAI abandoned a controversial effort to place its juggernaut artificial intelligence business under control of a for-profit entity and will instead remain under its founding nonprofit board."

8:45

We covered this story, this exact story in the journal when it dropped.

8:49

And Sam said, "Hey, it's not that big of a deal.

8:51

It's not that big of a deal."

8:53

And then people were like, "Okay, it's not that big of a deal."

8:56

Then we won't talk about it on podcasts. Yeah. Yeah.

8:57

Does he want us not to talk about this?

8:59

Is that what he's saying? It's not a big deal.

9:00

Don't make a three-hour podcast out of it about this one thing. Well, that's one thing. It's no big deal. Well, too bad.

9:07

We're making a podcast about it and it's four hours today, by the way.

9:11

Uh uh and so the move could complicate the company's future fundraising efforts.

9:16

Obviously, it's very difficult to invest in a nonprofit even though there was a wait for some cash to flow through.

9:22

Unlike traditional boards, which must must act in the best interest of shareholders, OpenAI's nonprofit board has a fiduciary duty to humanity.

9:30

I was joking that if they have a fiduciary duty to humanity, if you're human, you can just ask them to do things, can't you? Yeah.

9:35

You can just submit them tickets.

9:37

I was telling Will, make me a to-do list app now. Yeah. Do it.

9:41

I'm human, so you got to You're responsible to me. You're in service. You're in service of me. Yeah.

9:44

Uh but OpenAI started working on this change uh after Sam Alman was surprised, fired, and then reinstated in 2023.

9:52

its big investors, including Microsoft, wanted uh watched his temporary ouster from the sidelines, unable to wield official power over the outcome because of the nonprofit structure.

10:01

They couldn't just sue, hey, this is not in in the interest of shareholders.

10:04

Uh and the nonprofit board was able to make a very convoluted argument for why they were so suddenly changing CEOs.

10:12

Uh the conversion would have changed OpenAI's business to a public benefit corporation uh while preserving some form of the current nonprofit uh that controls the company.

10:22

Altman's rival estranged co-founder Elon Musk had tried to block the conversion in court.

10:26

Open Eyes said it had scraped had scrapped the Boulder plan after discussions with civic leaders and attorneys general in California and Delaware, which would have been interesting. Yeah.

10:35

And so they probably just went to the courts and said, "Are you guys going to veto this if we try and do this?"

10:40

And they said, "Yes, absolutely.

10:42

We're going to veto this."

10:44

And okay, we're not even going to try.

10:46

That's what that that's what that reads like. I don't know.

10:48

Um, but anyway, um, the framing of this news and the tenor of the overall reaction has been interesting.

10:53

OpenAI is characterizing this as a clear step back from their initial for-profit plans, and most people seem to be taking them at their word.

11:00

A small contingent, including Elon Musk, are on the opposite extreme, claiming that this is a total betrayal of OpenAI's nonprofit roots.

11:08

Unsurprisingly, Strateery falls somewhere in the middle.

11:10

Ben, of course, with a twist.

11:12

with a twist. I do in fact think this is still a meaningful change but also a fair one that does say stay stay somewhat true to the entity's roots and frankly that's the part that worries me most and so interesting of course that the open-source company is now closed source they're maybe releasing an open

11:30

source model I don't know if you want to pull up the poly market on whether or not openai will release an open source model or what the timing is but it seems like they they do that at some point it it doesn't seem like a right now it's sitting at uh Will OpenAI release an open source model before July? Poly

11:44

Poly Market has a it at a 48% chance. Wow. Really 50/50 basically. Yeah.

11:50

Still feels significant on 356K of volume. So not not small. Yeah. Anyway, so we'll see.

11:59

See how this Microsoft's obviously big player in here.

12:01

They kind it seems like they need to sign off on whatever happens and they want a stake and then there's also some crazy numbers about exactly how they how much they've invested.

12:10

And so uh the core problems facing OpenAI as a nonprofit were straightforward.

12:15

Developing AI costs astronomical sums of money.

12:17

Money at that scale is only available from investors who expect a return.

12:21

And OpenAI has no meaning no means to deliver a return to investors. Yeah.

12:27

Basically, you can raise a few hundred million dollars as a nonprofit from people like Elon and Peter, things like that, lots of other but beyond that, you start to run out of um you know, they could have just perpetually raised. Yep.

12:40

They could probably have gotten to a point where they were perpetually raising hundred million a year, but not$10 billion a year.

12:46

And my and my take on this was like it's beautiful that that that the future of AI must be capitalist.

12:54

And it's actually very good.

12:56

And I think Ben Thompson kind of agrees with this.

12:57

Uh so he says now you can certainly make the case uh that the most appropriate thing for OpenAI to do at the point that they had this realization and this was and everyone started realizing that scaling laws applied in 2018 after they released the paper improving language understanding by generative pre-training.

13:12

And this was confirmed with GPT2.

13:14

They saw that they didn't really train they didn't really change the algorithm.

13:18

They just increased the scale of the data and the compute and it got way better.

13:22

And so they're like, "Okay, well, we thought if it was just genius application, genius algorithm design, all you'd need to do is just raise $100 million, get a hundred geniuses in a room, turn them loose, and they would come up with the most beautiful, elegant algorithm for AGI."

13:37

Y give them enough time, they'll do it.

13:39

But that wasn't the that's not the solution.

13:41

They need compute and they need energy and that costs money.

13:43

And so um they should have wound down the nonprofit at that point is what Ben Thompson argues.

13:48

Um, and he says, "I wonder if Alman regrets not just doing that."

13:54

After all, and this is critical, OpenAI had at that point created very little of commercial value.

13:57

To review the timeline of that critical 2019 period, February, OpenAI releases GPT2.

14:04

This was the model that showed simply scaling up model size, data set size, and compute produce more generalizable results.

14:10

March, uh, OpenAI creates the OpenAI LP, a cap profit entity that could be invested in in exchange for a share of profits.

14:19

and July, Microsoft invests $1 billion in open ILP with the right to 100x its return.

14:26

So they could make a hundred billion to 100x. Yeah, simple ask.

14:28

Well, the funny thing here is that I'm pretty sure a h 100red billion is less than what Microsoft earns in a year.

14:36

So it is like a substantial sum of money, but it's not that much money to them.

14:42

Uh to put in context, 100 billion is 17 billion short of Microsoft's total operating income last year.

14:50

And so it' be like a nice EBIT was 131 billion.

14:53

But it sounds it sounds crazy, but I mean they put in a billion dollars.

14:58

So the question is like what's more valuable?

14:59

They grew 20 they grew 28 they grew EBIT EBIT 28% from 2023. It's amazing. Satcha on fire. Come on the show soon.

15:10

Anyway, of course, OpenAI and the rest of its investors almost certainly don't want to give up such a valuable asset.

15:17

Um, I'm not sure that they will have the choice.

15:19

So, the question is, should should o should should Microsoft go for revenue share or perpetual rights to OpenAI's models that they can vend into Azure.

15:27

Um, and they'd probably want to stick with the uh with the perpetual rights.

15:33

Uh the fact of the matter is that Microsoft put up uh the money to make OpenAI an entity worth fighting for and they would do well to give up future profits from OpenAI to secure profits of their own.

15:42

Assuming this deal happens, that leaves the nonprofit board, which has final say over the company and some amount of shares in the public benefit corporation.

15:50

Uh this is ultimately in line with OpenAI's original structure.

15:54

So I think it's a reasonable compromise.

15:56

I don't like it though, but not for the same reasons as those who think that any for-profit entity is wrong.

16:00

Rather, I don't like the idea of an entity unconstrained by things like fiduciary duty or the profit motive having ultimate say over what is shaping up to be such an important company.

16:09

I get the theoretical argument for people motivated by the common good.

16:13

I myself subscribe to the perhaps more cynical view that base motivations like profits and doing right by shareholders are important constraints against utopian visions that all too often end badly. And I completely agree.

16:26

He's very very convincing on that point. Yeah.

16:28

Um and then uh we can go through a little bit of the of the Simo Fidget Simo and uh and Wind Surf news.

16:36

So in an Altman email, Simo will enable our traditional company functions to scale as we enter the next phase of growth.

16:43

Altman keeps the top job and will increase his focus on research, compute, and safety systems.

16:48

Uh Simo is a technologist at heart, but a very pragmatic technologist.

16:52

She loves consumer companies that have a big impact.

16:55

That's why she went to eBay. And what else? like ads. Ads. Oh, yeah. Ads. Yeah.

16:59

So, u uh now the question is, you know, you interview anyone from OpenAI, when are you going to get promoted to CEO?

17:10

Because now if there's two CEOs, why not 50? Three, four, five. Yeah. Yeah. Unlimited. Why stop at 50, John? Yeah. Just keep going.

17:17

I mean, this is an entirely uncharted territory.

17:19

We've seen the co-CEO model before. Yep.

17:21

We have not seen scaling all is all you need.

17:25

CEO scaling law applied to CEO law applied to CEO.

17:32

Imagine do you think that's the value if value occurs to the most CEOs?

17:35

CEO drives the most value.

17:37

So get multiple CEOs more CEOs.

17:39

They need 10 more of CEOs. Yeah. Six billion CEOs.

17:42

Then they'll be responsible to humanity.

17:44

This is this is the way it's very likely that the company that develops AGI will be entirely CEO driven. Yeah. CEO.

17:51

Everyone acting like Yeah. They run the company. It's great.

17:57

Um, and so yeah, I mean it's obviously like a big shift to they're going all in to be the next big consumer app and they're necessarily going to build an advertising product.

18:07

And I was trying to think about this like there is a world where you charge people $200 a month or $20 a month or zero a month and the $20 a month subsidizes the rest of the of the cohort, but you have to differentiate the product at that point.

18:24

And so you're essentially giving poorer people dumber AI, which feels a lot worse to me than giving everyone equivalently smart AI or everyone the same ads or everyone the same content, but you either watch ads or you don't or you have Exactly. Exactly. Exactly. So I I I I truly Yeah.

18:47

I maybe it'd be nice to be able to pay to remove the ads, but but I I I do I don't like the world where where certain people based on their economic conditions don't have access to the smartest models.

18:58

I think that's very very true.

19:00

Who are we talking I mean, who were who are we I honestly think it was MadFez that we were talking about offline about how that this dynamic that we've had where billionaires have the same iPhone as somebody working at at right.

19:14

I that is not going to be the case in AI and that's kind of the issue that you're getting at right now.

19:20

Well, no, I think it I think it ultimately will be the case in AI that it will be the same because I do think because I do think the ads and the economics will lead to GPT5 being available to everyone equivalently.

19:33

And maybe there's less ads for certain people, maybe there's certain usage limits or certain things that you pay for.

19:38

Yeah, the argument is so much better for that if I'm if I'm paying 20 $400 a year for OpenAI as a proumer, right?

19:49

I use, you know, whatever.

19:50

I'm on the $200 a month subscription.

19:54

They could get probably 10 grand a year in ads in value from me as a user in ads. Yeah. Right.

19:59

Hey, we got to raise the price on you. Yeah. To get rid of those ads. Yes.

20:04

That's what but but but but no the argument is that like Google I would imagine you know like OpenAI at some point would have an incentive to just get so good at ads that they can extract the maximum amount of value.

20:13

Um it's so cocky to be like yeah my Arpoo at Facebook they they make 20 grand off of me because they show me so many AMG G63s that they just print off of me.

20:27

They just know they found the infinite money glitch.

20:29

They've shown me so many G Wagons that that Mercedes has proved.

20:31

We got another G Wagon addict figures this year.

20:37

He's already got two and he's a third. That's ridiculous.

20:40

Anyway, um but obviously, you know, the company is shifting to it's the what what do they call it?

20:48

The unlikely consumer uh tech company.

20:51

The un the unwilling consumer tech company.

20:53

Uh they just bought Windsurf.

20:54

Uh it's model agnostic which is very interesting because you can use windsurf with uh anthropic claude.

20:59

And some other news that was interesting uh Harvey just announced that they're going model agnostic. Interesting.

21:06

What were they built on before? OpenI.

21:08

Open AAI and OpenAI's fund was invested in in Harvey. Yeah.

21:12

And so this is the point that Ben Thompson makes which is that there is a world where chat GPT is powered by claude and no one notices or cares because all you care about is the app that being a great front door to AI and you go there and as long as you get good responses.

21:30

Does it really matter what exact model you're using? Not really.

21:34

The actual underlying model matters a lot less than how that model interacts, the other tooling, the tool use, all these different things.

21:41

And so uh we got to dig into this more.

21:43

But yeah, the other the other uh interesting thing here and Ben calls this out.

21:46

He says the unusual part of this this appointment is that Simo is currently a public company CEO of Instacart.

21:52

And the best way to understand her decision is that it is better to be the number two at a generational company instead of the number one basically anywhere else.

22:00

And it's interesting to um you know in many ways Instacart was a generational company. Yeah.

22:08

But uh but it's 11 billion dollar right now.

22:11

Um and it's been around there since IPO.

22:14

It's actually fairly up since IPO.

22:16

I believe $25 when it IPOed. Now it's at 45. So it's done well.

22:23

Um do you remember they were I think they were over 30 in the private markets. Oh, really? Yeah.

22:27

But still, I mean 11 billion that's what 3% of Open AAI at the current valuation or something like that.

22:34

But the the takeaway is that if you're running a startup, why don't you just go recruit a CEO from public company to come work for you? Totally. Totally.

22:41

Set your set your bar high.

22:44

Massively appreciated approach for recruiting. Yes.

22:47

Just go look at public markets, see which companies have, you know, executed well over the past, call it eight quarters, and then just go pick out, you know, your favorite CEO.

22:58

So I want to shift over to John Ludig's piece on his Substack uh Lutig's learnings.

23:04

Uh he says the future of foundation models is closed source if the centralizing forces of data and compute hold open s open and closed source AI cannot both dominate long term.

23:16

So he was thinking about um you know this was this was during this wave and this big meme that uh open source would just completely dominate and commoditize the foundation model layer.

23:27

Uh, some of that's happened, but then we've seen missteps with Llama 4 and there's some interesting context in here, but it's a year old piece, so it's interesting to reflect on how it's held up.

23:36

And he got into a big debate with Martin Casado, who I believe is coming on the show in just a couple hours.

23:40

Um, and I'd love to hear how Andre's view of open source AI has evolved because they've invested in some open source AI companies and some closed source AI companies.

23:50

And so I want to know how they're thinking about it.

23:52

Um so uh John writes two uh two seemingly contradictory but equally popular narratives about the future of foundation models have taken hold.

23:59

In one future AI centralizes scaling laws will hold and value acrru primarily to scaled closed source players.

24:07

In the other future AI decentralizes foundation models have no moat.

24:11

Open source has caught up to closed source and will have many competing models.

24:15

Today both both narratives seem true.

24:17

We have powerful closed source models and a thriving ecosystem of synchro sanct open-source models.

24:23

Llama 3 recently put open source on the map of GPT4 class models.

24:27

Meanwhile, an unusual open- source alliance is formed among developers who want handouts, which is very controversial to say, academics who embrace publishing culture, libertarians who fear centralized speech control and regulatory capture, Elon who doesn't want his nemesis to win AI, and Zuck who doesn't want to be beholdened to yet another tech platform.

24:45

as an accelerant of it's like not letting it happen.

24:50

Simply not happening again.

24:50

I will spend a hundred billion dollars this year to make sure this doesn't happen again. Yeah. Yeah. 100%.

24:56

And he's and this is what keeps him up at night. Yeah.

24:59

Cuz it it might be happening for sure.

25:01

Um but I think he's doing okay.

25:04

Like the the llama models, they're not fantastic, but they're definitely good enough for enterprise use.

25:08

And that's that's really the win case for uh enterprise AI or open source AI is is it's seeing fantastic adoption in the in the uh in the enterprise where you can fine-tune it and tweak it and really understand the cost structure and you can stuff it in all different parts of your enterprise.

25:25

So as an accelerant of modern software, open source maintains a cherish a cherished place in tech.

25:28

Who can argue with free stuff, decentralized control and free speech, but open and closed source AI cannot both dominate in the limit if centralizing forces hold scale advantages will compound and leave open source alternatives behind.

25:41

Despite re recent progress and endless tiering, open source AI will become a financial drain for model builders.

25:47

Basically, this idea that uh Zuck will run out of of investment dollars to pour into Llama if there's not a strong economic engine there.

25:59

And that's the main that's the main argument here.

26:01

Uh and so open source software started as an act of charity.

26:05

The world owes the likes of Linus Tovolds and Fabris Bellard for endowing humanity with Linux, Git and FFmpeg.

26:11

Because free stuff is popular, open source became a great premium marketing strategy.

26:15

Think data bricks or Mistral.

26:17

And sometimes a market equilibrium in itself, eg Android, a cheap smartphone option reinforces Google's search monopoly.

26:24

Um, companies that earned free marketing from open source eventually succumb to business physics.

26:29

Red Hat hid sent OS behind a subscription service.

26:32

Elastic Search changed their licensing after accidentally seating competition and data bicks owns the IP that accelerates ex Apache Spark.

26:40

So Apache Spark is open source, but you can't just spin up a databicks instance.

26:44

You have to license their IP.

26:47

So you have to use data bricks.

26:48

Unlike the charity work of open source in the early software era, today it is subsidized by businesses with their own data bricks. Oh, yeah.

26:54

Bought a company for a billion dollars today. Uh, got announced.

26:58

Yeah, they announced this at 300 a. m. Pacific.

27:01

I guess that's 6 Eastern, so not that great.

27:04

They're just on the Ashton Hall grind.

27:06

Play the Ashton Hall sound.

27:10

We're announcing something at 3:00 a. m. They are banana.

27:12

Many people have called data bricks the Ashton Hall uh press data analytics. Data analytics. Yeah.

27:19

Uh, and so, uh, John writes, uh, that begs the question, if Meta is only pursuing open source in so far as it benefits themselves, which he believes, like they're not doing it completely charitably. It's not a charity. It's not a nonprofit.

27:32

What is the tipping point at which Meta stops open sourcing their AI?

27:36

Uh, so sooner than you think.

27:38

Exponential data frontier models were trained on the corpus of the internet, but that data sourc's commodity.

27:41

model differentiation over the next decade will come from proprietary data both via model usage and private data sources.

27:49

And so the feedback loop in chat GBT is what's valuable.

27:51

The thumbs up, the thumbs down, the the follow-up questions.

27:54

You you you issued a response, did they close the app because they were satisfied or did they ask you 25 more for follow-ups to get the real answer?

28:03

So open source models have no feedback loop between production usage and model training.

28:07

So if you're running Llama, that data doesn't go back to Facebook. Exactly. Exactly.

28:11

And then there's exponential cap capex, a lagging edge model that requires just a few% of Meta's 40 billion in capex is easy to open source.

28:20

Nobody will ask questions.

28:21

But once you reach 10 billion or more in capex spend for model training.

28:23

Um shareholders will want a clear ROI on that spend.

28:28

The metaverse raised some question marks at a certain scale too.

28:31

Diminishing returns on model quality within meta.

28:33

There's a large upfront benefit for meta building open source AI model even if it's worse than the frontier closed source counterpart.

28:41

There are lots of small AI workloads.

28:42

Think feed algorithms, recommendations, and image generation where Meta doesn't want to rely on a thirdparty provider like they had to rely on Apple.

28:47

But it's unclear whether Facebook products benefit much from models approaching AGI quality.

28:53

It's equally possible that Meta's model improvements will be very particular to their own internal use cases.

28:59

And that's like what we talked to that uh that that the company that trains small uh language models task specific.

29:08

Yeah, they had task specific models.

29:10

And so you could imagine that a lot of meta's a AI workloads are not going to need AGI level univers. That's right. Fastino Fino AI. Yeah, Fastino.

29:23

And so uh so you can imagine that there is a Fastino task specific specific language models.

29:28

And so you can imagine there is a Fastino like team inside of Meta actually figuring out, okay, we just need to translate every caption.

29:37

Let's let's build a model specifically for that or we just need to flag everything for profanity, right?

29:42

Like they could build a model for that and they're not going to need AGI level massive models for that.

29:48

And so uh Zuck is not running a charity. He's a savvy capitalist.

29:50

While Meta can justify scaling capex on incremental models for their own ends, their open source strategy will only make less sense over time.

29:57

And so then uh John goes into developers and consumers.

30:03

Um as a developer choosing an open source model, what do you get in terms of cost, model quality, and data security?

30:08

Open source models have the illusion of being free, but developers bear the inference cost, which are often more expensive than comparable LLM API calls.

30:17

You either pay a middleman to manage GPUs or and host models or pay the direct cost of GPU depreciation.

30:21

Um in capitalist America, free is never really free.

30:26

You should wonder how you'll ultimately be monetized.

30:27

This isn't Linux where a single developer built the product as a gift to humanity.

30:33

These are cash incinerating endeavors whose only way out is to eventually monetize you.

30:37

You're probably committing to a closed source compliment in time.

30:41

I'm not afraid to be monetized. Yep.

30:45

Are you afraid to be monetized?

30:45

I'm fearless about being monetized. Monetize me. Go right ahead.

30:48

Even Even Android eventually monetized via Google Play.

30:52

like a pretty good, you know, metal song, you know, monetize man.

31:00

Uh, on model quality like housing, healthcare, and education, a paid version is generally better than the free version.

31:05

Even within software, the open source winner is rarely the best product.

31:08

Android is worse than iOS.

31:08

Open Office is worse than Office or Google Docs. GDO is worse than Unity.

31:12

Freecat is worse than Solid Works.

31:14

A coralary is that engineers focused on the best platforms make more money.

31:18

They're more likely to build cuttingedge products.

31:22

Everyone's celebrating Llama 3. Yeah, here's the kicker.

31:24

Everyone is celebrating that Llama 3 is on par with GPT4 a year later.

31:27

A year later, the product quality gap between iOS and Android or Mac OS and consumer Linux has stayed wide for a long time because the best software creators are aligned with paying customers.

31:38

When you choose closed source models, you're not making a point in time decision on model quality.

31:44

You're paying for future model improvements where the road map is aligned with paying customers. Yeah.

31:47

Uh, and then he talks about data security a little bit.

31:51

Some enterprises need the utmost data security, financial services, healthcare, legal, but I'm not you I'm not sure using open source models on prem or via thirdparty cloud hosting is actually safer than using thirdparty LLMs in the cloud.

32:00

This is a legacy belief from the early internet era where an on-remise data center was the Fort Knox of data security.

32:07

As a customer, I'd trust Microsoft with healthcare data security more than my IT department's self-managed data center.

32:15

Oo, putting Founders Fund's IT department on blast there.

32:16

But very nice, very nice IT department over there. They're very nice. I know them.

32:24

And that bridge has already been crossed when 65% of the riskadverse Fortune 500 already uses Azure on OpenAI.

32:31

It makes you wonder who is dealing with data that is too sensitive for cloud-based LLMs, of course.

32:34

And so then he goes into national security, and this is why I wanted to pull up that Aaron Gin clip, which we'll play in a second.

32:42

But uh even if it makes eventual economic sense for model builders to build open- source, should they?

32:48

Advocates like Yan Lun claim that open- source AI is safer than closed.

32:51

But it makes me wonder if he really believes in Meta's AI capabilities.

32:54

Any reasonable extrapolation of capabilities with more compute, data, and autonomous tool use is self-evidently dangerous.

33:00

And we've talked to some other folks about this.

33:04

Um so let's play the Aaron Gin clip because I want to react to that.

33:09

But a new AI gunro initiative has already started.

33:12

They're already around the world. We know this.

33:13

We see them at, you know, different data centers that we work with selling their stuff and it comes as a fully complete solution with DeepS open like Manis models combined with infrastructure combined with support.

33:27

Manis has has that kind of distribution already. You're not talking.

33:31

No, like Huawei is designing for their their their open source ecosystem.

33:35

So, so Huawei is being the hands of meat out there building.

33:37

I think what J's asking is like, have you ever run into a country that's not China?

33:43

That's like, yeah, we're we have Huawei Ascend and then we have Deepseek and then we have Manis because Manis came out of nowhere just a couple months ago.

33:51

Feels like that would be a very quick ramp to be like it's deployed in another country as a part of the AI belt and road initiative.

33:57

Deepseek, I can believe Manis.

33:59

So, yeah, give give us the update on that. Yeah. Yeah. Yeah.

34:03

It's the point like Deep Seek and Manis are everywhere in Europe.

34:05

Like they're they're everywhere.

34:06

And again, it's open source.

34:09

It it open source works as a distribution strategy.

34:11

They're not open source because they give they they kind of cares.

34:14

They don't care about open source.

34:15

They care about defeating us.

34:17

And and and so the the approach that this is why uh OpenAI made that announcement of we're going to do like this combined deployments of software plus like infrastructure is because that's they're mimicking what China's doing.

34:29

doing. uh but but the proliferation of the open source models go look at open router like you can just see the growth in other models those other models are Chinese okay so there's this debate about now so so where I think John left off is that

34:46

if we're open sourcing all American AI and truly the best American AI you're in you're immediately giving it to nearpeer rivals geopolitically and that's risky what Aaron is saying is that if we don't do it, they will. And there could be

35:00

And there could be this AI belt and road initiative where a country that's a little bit more of a coin toss might say, "Hey, my options are pay OpenAI a ton of money or use the free Chinese model.

35:14

I'm going to go with the free Chinese model."

35:16

And that draws me closer to their economy and that pulls them away from us.

35:20

And so in terms of forging durable relationships with countries that are kind of on the brink or or could go either way, having a strong American secure open-source AI ecosystem might actually be an advantage to us in terms of like broad AI economic warfare.

35:39

And so yeah, and the question that's something that I've been gets at this a little bit.

35:43

I I think the question becomes how much does being able to deploy the truly bleeding edge leading models matter, right?

35:53

Does it matter if you're deploying, you know, GPT6 versus Y the open-source CCP equivalent of GPT5, right?

36:01

How big of a difference is is that? Right?

36:04

Everybody's going to have a different take.

36:05

Um but uh but yeah, we got Let's dig into dig into it today.

36:11

Uh, let's tell you about RAMP really quick. Time is money. Save both.

36:14

Easy to use corporate cards, bill payments, accounting, and a whole lot more. All in one place.

36:16

And let's invite our first guest into the studio. Welcome to the show. How are you doing? It's Granola time. What's happening? Congratulations. How you doing? We're doing great.

36:26

Uh, would you mind introducing yourself and the company a little bit? Totally. My name is Chris.

36:30

I'm the co-founder CEO of Granola.

36:30

Um, how much you wanted me to tell about it? Tell us everything. Yeah.

36:36

I mean specifically the evolution specifically the evolution of the product because I know people think notetaking but you're doing a lot more.

36:44

Uh so give us kind of the what was the initial pro uh the initial uh minimum viable product.

36:53

What's the product today? Where is it going? Totally.

36:54

So uh we launched Granola a little under a year ago and the idea was the best way to take notes during meetings and to do that in collaboration with AI.

37:04

So, it's an app that lives on your Mac.

37:07

It looks a lot like Apple Notes.

37:09

You can take notes however you want, but it listens to the conversation in the background and when the meeting's over, it'll like rewrite your notes to like flesh them out to make them great. Um, yeah.

37:20

And that that's the basic product.

37:22

We tried to keep it super simple, uh, really minimal.

37:26

And, um, I mean, I think I think that design resonated.

37:30

I think we can talk about that.

37:31

But there I think there are a few things about the way we built that that people really liked.

37:34

And then I think what we noticed happening was that once you start using granola, what happens is you have a meeting and like a notification pops up saying, "Hey, you this meeting starting.

37:43

Do you want to use granola for it?"

37:45

And once you start using granola for lots of meetings, it kind of becomes this like searchable second brain repository thing.

37:54

Like you can kind of be like, "Hey, what did that person say?"

37:55

And you kind of turn to it and we notice lots of users doing that.

37:59

And then kind of the evolution like what we launched today is now you can do that across your team.

38:04

So like now you can you can chat we added a bunch of support for these great reasoning models and you can chat ac across meetings across your whole team and ask questions and it's a stepping stone towards like the bigger vision of where we're going. Yeah. But yeah. Uh tell me about search.

38:20

Are you stuffing every granola uh note into a vector database or are you just doing a huge context window?

38:28

What are the different strategies to actually surface insights?

38:32

It I mean it's it's definitely evolving.

38:34

What what we found is like rag or like stuffing it into a vector database is really good for certain types of like information lookup like search.

38:42

What what it's what it is terrible at is basically if you ask questions like hey um like coach me like how how could I be better in my one-on- ones, right?

38:52

better in my one-on- ones, right? or um when let's say you're trying to sell something it's like hey when I get this type of question like how good are my answers and how does that compare to the answers that other folks on my team give like it completely rag completely breaks with that and what we find is people

39:08

turn to granola to ask a lot of these kind of analysis type questions like hey what are the biggest burning fires right that I should be focused on or something like that and um rag's not suitable so we actually make a lot of use of stuffing these massive context windows with the right with the right transcripts and and not what's the favorite context window to stuff? Um I

39:26

Um I know Gemini's got a big one, but they all have trade-offs, but llamas do the bigger scare you.

39:35

No, no, the big ones scare me.

39:37

Uh open AI's context windows. Perfect.

39:44

Um we just we just today starting today we let users choose.

39:47

Before that it was, you know, we do we have everyone's going model agnostic. This is a trend, right? Yeah.

39:53

I mean, wind surf is model agnostic now.

39:56

Who else were you saying?

39:57

Uh Harvey just went model agnostic.

39:57

Uh this seems to be the trend. Really?

40:02

You're drinking the commoditization Kool-Aid. Yeah.

40:04

The the product person in me feel I have mixed feelings about about surfacing it to the user because it's like one more thing a user has to think about and your average person shouldn't have to think about.

40:14

We should do the best thing. So it defaults to auto.

40:16

So we try to choose the right model which for a lot of it is is like Sonic Cloud 3.

40:21

7 like the reasoning mode of it because we find that that for the type of queries that we're getting is it works the best.

40:26

But you know if you're a power user now you can choose Gemini 03 like whatever you want. Cool.

40:30

Uh I got to ask uh what was your reaction to notion's launch yesterday?

40:34

Uh many people said uh it was a granola clone.

40:40

Do we have a shots fired?

40:40

Yeah, we need like a shots fired sound effect.

40:42

But, um, yeah, I'm curious to get I'm curious to get, you know, your reaction to it.

40:48

Um, I'm sure it's not even the first time that somebody's taken a little inspiration from you guys over the last year.

40:54

I mean, I it feels like it means we're doing something right is is my first reaction.

40:59

Like I it'd be easier if I didn't like Notion.

41:03

We really like Notion, right?

41:04

Like we think they're great.

41:04

Like Ivan actually DM' me on Twitter letting me know it was coming beforehand, you know, as in like, hey, wow, that's really sweet.

41:11

It's really cool, you know? like we respect you.

41:13

I hope this leads to us pushing each other to build better products for people. So like I I like them.

41:19

Um what it means for I mean I think the world changed overnight with AI and we're like 2% of the way there figuring out like what does that mean in terms of products and tools and there's just a ton to build, right?

41:32

I feel like transcription from the beginning has always been a a commodity, right?

41:38

Everything is going to have transcription.

41:39

So I think the question is how do you go from what we have today to like a true AI assistant that kind of understands you and helps you like be a thought partner and do most of the work you want to do. Yeah.

41:51

It feels like there's like three different like approaches to building a new AI enabled note-taking app or something like that in this world.

42:00

It's like uh green field project with a startup that has a fresh clean slate like you guys.

42:06

clean slate like you guys. There's notion which is like a well-funded scaled company but still in founder mode can still like we saw this with like Figma roll a bunch of new features in move still pretty agile compete and then there's the hyperscalers which are just like we have a monopoly and yeah it

42:23

might take us 10 years to add this feature but you're probably still going to be using an iPhone so uh yeah expect Apple notes to get an update in 2035 or something like that but I'm interested to know kind of like do you see the market that way and and how are you thinking about uh the broader competition? Because I imagine that if

42:41

Because I imagine that if you just look at the overall like people taking notes, Apple notes is probably still the number one product or something like that or notepad.

42:50

Um and so maybe that's the real opportunity and you and notion can both win.

42:57

Yeah, I think it's an interesting question. it. Okay.

43:00

So, one one angle is like I think we underestimate ju you remember when mobile came out and everyone tried to port websites to to mobile to the iPhone and like it took us a little while to figure out that that doesn't work and that mobile apps should look really different. Yeah.

43:17

The jump from mobile to from web to mobile is a fraction of the jump of like prei to post AAI.

43:25

So like I really think the future's going to look very different.

43:28

I think and in a in a very different future.

43:33

I think the the products that are kind of like invented in that new medium tend to do better.

43:37

You know, that doesn't doesn't mean like the the massive companies won't be able to add really cool AI features, but if you think about the AI features you use on a daily basis, how many of them were launched by startups, you know, that that came out in the last two years versus how many of them were launched by big companies?

43:54

Like it's like an interesting assessment to do.

43:55

So, like I see it like there's a ton of opportunity.

44:00

I really think like when Sam and I started Granola, it wasn't a our vision was never meeting notes.

44:07

Our vision was to be like a tool for thought that is contextually aware and AI powered like from the you know you can you can see the stuff that we wrote before launch and meeting notes just turned out to be a freaking awesome wedge.

44:21

Like we thought it would be a good wedge.

44:22

Turns out it's an amazing wedge.

44:24

Way better than we thought because there's like so much information in in our conversations that we don't is there also like a viral loop there where people are sharing their notes with each other on boarding I imagined. Yeah.

44:32

I mean like basically the granola product is nothing like there's no growth stuff built in like until today there's it was like single player product. Oh really?

44:41

It's just something really Yeah. Yeah.

44:43

There's something super inherent about the information in meetings and wanting to share that right.

44:46

So it's it's all been like word of mouth until now. Yeah. I think.

44:50

But I imagine I I if I take notes in a meeting, I could share a a static link and we're not multiplayer in that, but I could share my notes and then I'm seeing granola.

44:58

com and I'm just That's true. Yeah. Yeah.

45:00

You can you can share a link and and now you can even you can like put a bunch of meetings together in a folder.

45:05

You can share folder and someone who doesn't have granola can now see those and you know if they log in and then um granola. ai.

45:11

Sorry, I got the domain not wrong. Do you have I'm curious.

45:13

Do you have a do you have a strong thesis around agents?

45:17

You know, I can imagine at some point companies have agents that are calling people and companies want to kind of understand the the context of those conversations and what what happened what what kind of ne next steps need to take place.

45:30

Is that something that would live in granola or is that more at the CRM level?

45:34

What what what's your theory there? Oh man.

45:36

I I mean I I think the you mentioned CRM.

45:39

I think the product categories that we've had for the last 20 years, whether it's like okay, you have like a wiki company, you have a CRM, you have like your docs company or I think those are all going to get redrawn because I think in in an AI world, it's it's all about context, right?

45:54

And then how do you make use of that context to to do work? So yeah.

45:59

Yeah, if you have agents calling people, I don't think that's very different from other folks on your team calling people.

46:05

And then how do you make that context useful?

46:06

I think that could live in granola.

46:08

I also think agents are just one of those terms or I think we throw it around now, but we don't really there's probably like 15 different things that today we're calling agents that are going to be their own separate things in the future.

46:19

Um, and it'll be really interesting to see how how that evolves.

46:23

So, I got a real time uh follow-up question from our mutual friend Michael at Lightseed who uh was in is in the new round, but he asked uh uh is there an opportunity around agents to take uh action uh based on meeting notes, right?

46:39

Because in a beautiful world, humans get together, they have a conversation, you figure out next steps, and then things immediately start to happen uh autonomously. 100% like 100 like that.

46:50

like that. That's the whole future of granola is basically like how do we use this context for meeting notes but also in the future email slack you know Google docs whatever to help you do work and like our vision for granola is like AI should do all the busy work for you right and you should only be left with

47:07

the judgment so you can imagine a world you come out of a few meetings it's already kind of done all the the postmeating work or suggestions and all you have to do is kind of be like h a little bit more this way or yeah that looks good yep yep yep yep do that and then move on to your Next thing, what's the use of the funds? 43 million big 43 million big series B.

47:24

Uh is LLM inference a material cost to your business right now?

47:29

Is it just hiring more engineers, product people?

47:34

Doesn't seem like you're going to be training foundation models if you're moving to uh model agnostic, but what's the use of funds look like over the next 12 to 18 months?

47:41

Yeah, I mean I like I said, I think there's a we're like in this pivotal moment in history to build the future tooling of work and it's a big vision and we need to hire the best people in the world to go out and build that and there's a just a ton to build.

47:55

So that's the primary use.

47:55

Um LLM inference is it costs money.

48:00

Transcription actually costs more.

48:00

So we we transcribe millions of minutes millions of minutes a day right now.

48:04

So that actually does cost some money but people also pay for granola.

48:09

So like that the funds are really like can we go after this massive vision?

48:12

That makes a ton of sense.

48:14

Uh and what's uh what's what's the uh dispersion of the team going to look like over the next couple years?

48:22

I know you're in London now.

48:22

You are an American though. Yeah.

48:23

Uh what what what does it look like? Yeah.

48:26

I mean I think we So when we launched Can we were four people, we're 18 people now. We're still tiny, right?

48:33

Uh but we got a great group of people here.

48:35

So, we're definitely going to grow um the team in London because I think there's a lot of talent, but our ambitions are global, right?

48:42

So, we're definitely going to be opening up more offices probably in the US, probably multiple in the US. Um but all that's TBD. Awesome. Very cool.

48:51

Thank you so much for the milestone.

48:53

Congratulations and love to see Natt Friedman making some money.

48:58

Size Lord, we love having Chris. Yeah, he's great. Thank you so much, guys. We'll talk to you soon.

49:04

Have a good Uh, quickly let me tell you about public.

49:06

com investing for those who take it seriously.

49:08

They got multiasset investing, industryleading yields, trusted by millions, and they have a massive news dropping today.

49:14

Uh, now you can you can vibe code your own portfolio.

49:20

Basically coding indexes. Yeah. Yeah.

49:22

You you can go and type in I I I want to create a a synthetic index fund uh founder mode CEOs or I want to find CEOs that have been deered for more than 20 years.

49:34

people that have been quoted in their uh in their transcripts referencing Warren Buffett and it'll just pull it together, make it for you, and you can Very cool. Yeah.

49:43

So, you can do uh my personal favorite CEOs who don't believe in meetings. Fun one. That's cool.

49:49

But even better, CEOs that deadlift. Oh, is that building?

49:55

You can just go build an index based on I love it.

49:57

Uh guys that are very into Olympic weightlifting. Fantastic.

50:04

Well, we have Josh Browder in the studio. He He's a CEO. We'll ask him. Does he deadlift?

50:08

We got to know because he's going the portfolio out.

50:11

Welcome to the stream, Josh.

50:11

Great to have you on too long. How you having me?

50:16

Um, love the show and I I love Wonder as well. I see that Wonder. Oh, yes.

50:18

You're you're you're one of the first investors, right? We did an audit.

50:21

I'm proud to be the first investor. The first investor.

50:25

You found your happy place. Found your happy place.

50:27

First spot on the cap table. That's fantastic.

50:31

Uh so yeah, give us a little bit of an introduction and uh do you call yourself a CEO or a venture capitalist at this point?

50:38

You you you're on fund four so moving more into the VC world. What what's the story?

50:42

So definitely a founder at heart.

50:44

I'm the founder of do not pay which helps consumers fight companies and governments.

50:48

The very first use case was parking tickets.

50:49

Uh but today I'm announcing um my fund that I do on the side called Brower Capital Fund 4. Um, I'm very lucky.

50:56

We've raised $30 million from a lot of the best investors and I specialize in backing um undiscovered uncredentialed founders.

51:08

So, we're living in a world now where there's like these crazy seed valuations are like 50.

51:12

Um, everyone going to like one company gravitating towards like chasing hype. I'm the exact opposite.

51:18

I uh back more teal fellows than any other investor or fund and I really try and create my own hype and almost incubate them.

51:28

Obviously, it's um up to the entrepreneur to build their business, but I put them on the right circuit.

51:31

And um the very first company I did this with was I was the angel investor in a company called owner.

51:37

com in the preede and it's now yesterday they announced their billion dollar. Congratulations.

51:46

Uh what how are you thinking about this?

51:49

How are you thinking about the fund sizing?

51:50

I feel like most funds when they get to fund four, they're thinking three billion.

51:54

Um, but you've kind of kept it small.

51:56

What What are the advantages of that? So, um, I was lucky.

51:57

So, it took about 3 weeks to raise and I've done some raises in my career and, um, this this was, um, a very smooth process.

52:06

Um, and I deliberately kept it small and disciplined. I was lucky. I had a lot more demand.

52:12

Um, I want to do a good job for people.

52:15

I want to do a good job for the founders I back and my LPs.

52:16

And I think smaller fund sizes far outperform larger ones.

52:21

Um, in terms of what I'm backing now, I actually have a founder like living with me that I don't know too well.

52:27

I invited him into my home.

52:30

Um, another founder I put um his visa on my credit card um personally.

52:36

Uh third founder I co-signed for the so it's it's really just about putting people on the right track and being their first believer.

52:42

That's so that's amazing.

52:44

Um what uh what overall trends are you excited about or staying away from?

52:49

What's your take on defense tech?

52:52

Uh the application layer in in AI kind of built a the perfect company to benefit from foundation models.

52:57

Didn't have to train anything, but I'm sure do not pay is benefiting from AI and agents and uh reasoning models.

53:03

Um but what what are the trends that you're most excited about right now? So I'm founder first.

53:10

So wherever the top founders are going, I go.

53:12

Um, for example, a recent company that I was the first believer in with Corey, um, is called Pilgrim.

53:16

I think he was on your show a few weeks ago.

53:20

They they raised an amazing round, uh, I think from Kantos and others, which is a top defense investor. Also excited about AI.

53:28

Um, I'm investor in a company called Assured, which helps insurance companies process their claims, and they've received huge traction in the market.

53:36

Um I I think every AI every company will be an AI company though.

53:41

So I try and stay away from like founders who are like too smooth with their pitching AI first.

53:46

I I think about this with do not pay.

53:49

We've added AI to our business and really every company should do that.

53:53

So kind of these AI first pitches where the founders are saying all this jargon like we're building LLM operability dev tools.

54:00

What does that even mean?

54:03

So, I'd like to focus on like real problems and real solutions. Yeah.

54:08

I I think we actually met Jake Adler of Pilgrim together when we were camping a year or two ago.

54:13

I don't know if that's the first time you met him, but that was the first time I met I think both of you in person.

54:17

Um it was the first time I met him in person, but I this was more in the co era, so I was like doing my program over Zoom.

54:22

Um but I really enjoyed playing mafia with you guys.

54:25

I think you can um Oh, yeah.

54:26

If someone's a good founder or investor by how good they are at mafia. Yeah. Yeah. a lot of bluffing. It's fun.

54:32

Um I mean on the topic of like I I mean I love Jake, I love Pilgrim.

54:37

Um I don't feel comfortable uh due diligencing a company like that because it's so out of my wheelhouse in the biotech space.

54:43

Is that one of the benefits of being early stage and founder led is that you can kind of just bet on the founder and you're not you you don't need to do all the technical due diligence or do you like to dig in on that stuff and just get up to speed?

55:00

Um I I um I don't want to get into trouble by revealing any like specific numbers, but of my biggest kind of wins, they've been uh sub five or even around two average.

55:11

Um and so that kind of risk is baked into the uh valuation entry price. Sure. Sure.

55:18

And I I've seen I've been operator for 10 years now.

55:20

I I've seen things come and go.

55:22

And if you have a founder who truly it's their life's work and they're technical, they have some sort of good success and so they just have to keep at it and stay in the game and not run out of money and then they become really big.

55:35

Uh how long do founders stay under the radar today, right?

55:42

And and if you can invest in a founder at five posts, right?

55:44

They're typically you're either friends with them, you can add a tremendous amount of value, or they're flying under the radar in some form of another.

55:52

And then at c at a certain point they become kind of discovered broadly and they start getting you know these rounds start getting bit up.

55:59

It feels like with the internet today, uh, even X, you know, a kid can drop out of college and immediately get, you know, uh, three term sheets in a week, especially if they can go viral, which is like a common thing that we're seeing now with X is that some founder does some sort of stunt, gets a ton of attention, um, and then there's a question about whether they can build a product, but it certainly drives venture capital interest.

56:23

But yeah, I'm interested to hear your take on that.

56:25

So I think we're seeing a rise of like super credentialism.

56:29

Now VC top VC firms are even tracking math competitions in high school and you have these founders they tell the perfect story.

56:35

They're like we I won Intel. I won math. I went to MIT. I dropped out of MIT.

56:39

I worked at OpenAI and now I'm starting a company.

56:42

Those ones are super chase. Everyone's chasing them.

56:47

There's lots of competition.

56:47

Um but for me I really think it's about creating one's own winners.

56:51

And so the best entrepreneurs I've backed, they didn't have those credentials coming in and maybe they needed to be put more on the right track in terms of what they were working on.

57:02

And um a lot of sometimes investors even passed on them.

57:06

Um and so um I really I'm not saying that math competition winners make bad founders.

57:12

Obviously there's some huge wins with that.

57:13

But just because someone's good at math doesn't mean that they make a good founder.

57:16

and there's like more innate qualities that other people aren't signaling on.

57:21

How do you think about um kind of the geopolitics of entrepreneurship right now?

57:27

Um we're in kind of this this odd realignment towards America.

57:32

There's there's some VC funds that are pushing really hard to invest in Europe.

57:36

There's other folks that are great at I mean you said you you you paid for someone's visa.

57:40

Clearly that sounds like you're bringing a founder from somewhere else to America.

57:47

uh what what does the trade imbalance look like in the founder ecosystem?

57:51

Is America still bringing enough uh talented entrepreneurs here?

57:56

Is it still fertile ground for uh the global entrepreneurship community to come to America or Silicon Valley and build generational companies?

58:03

So, I'm originally from England as you guys can probably tell from my accent and I think that America is by far the best place to build a company.

58:11

People are more ambitious here.

58:12

the capital availability is like 100x.

58:15

Um I I think a concrete example is like the most valuable private company in the UK is like Revolute which is like a bank clone.

58:21

The most valuable private company in the US is like Open AI or SpaceX.

58:26

Um there like order of magnitude bigger.

58:29

Um so I encourage any founder I back to move to the US and also to move to the San Francisco Bay area.

58:36

Um, unfortunately, um, still being in the Bay Area really matters a lot, but companies don't have to start in the Bay Area.

58:42

They could start in Arizona or wherever they drop out from, but they should really move here.

58:48

Um, Pilgrim is a great example.

58:50

Um, Jake started in Canada and then moved to the Bay Area.

58:53

Um, I also think there's a huge arbitrage on the 01 visa.

58:58

Um, it's people, it's known as the Genius Visa, but it has like over 95% approval rating uh, depending on the year.

59:06

And so you get someone the 01 visa and they're a true genius.

59:09

They move to Silicon Valley and then they hit the ground running.

59:14

Uh what about advice for young people?

59:16

Do you think uh learning to code or being able to code is still a positive signal or in the age of cursor and windserve and devon and AI tools it's becoming less relevant.

59:26

I think the biggest thing is like distribution and being an online first and having these skills of like distributing your product coding and being technical is important but it's just one skill.

59:36

Um like I backed the WAP founders um WH as an early check and they were like really deep in like sneakerbot communities and things like that because everyone is starting a company now and you need to break out.

59:49

I feel like a lot of Stanford students are just starting companies just for the resume.

59:53

Um there are so you there's so much noise and you have to kind of wade through that.

59:58

Yeah, the [ __ ] founders are definitely doing it for the love of the game.

1:00:02

I mean they're they're killers. It's amazing. What a great company.

1:00:06

Can you give us uh I know it's not why you're on the show today, but can you give us the update on on do not pay it seems like in so many ways uh the company was incredibly well positioned to benefit from uh improvements at the model layer.

1:00:20

Uh give us the update there. Yeah.

1:00:22

So, Do Not Pay is an incredibly efficient business. It's a very small team.

1:00:27

It's only 14 people and and seven of those are like expensive employees and seven contractors.

1:00:31

Um, and when I started Do Not Pay, it was really about templates to get people out parking tickets.

1:00:37

And admittedly now, ChatGpt can write you a template letter.

1:00:42

But AI has improved our product by 5 to 10x.

1:00:45

So, instead of helping people with parking tickets, now we have um AI agents.

1:00:49

They log into someone's Comcast account and start negotiating with Comcast to get people's bills down.

1:00:53

And the big companies have AI and we have AI and it's like AI versus AI warfare.

1:00:58

But we're we're more motivated than the Comcast of the world.

1:01:03

And so we're winning like thousands of these negotiations a week.

1:01:06

Um, and I think it's a good lesson that I tell all the people I back is like you can't rest on your laurels.

1:01:12

constantly have to be innovating, change the product and adapt with the time. Yep. Makes a lot of sense. Fantastic.

1:01:21

Well, thank you so much for coming on the show.

1:01:24

We appreciate you and good luck with the new fund.

1:01:26

I feel like uh I I my only concern for Fund 4 is that some of these bigger multi-stage funds are going to sort of pay a private investigator to follow you around and just sort of like see who you're meeting with, you know, ahead of time and be like, "Oh, Josh is trying to do another he's trying to do another, you know, 500k on 5 mil.

1:01:44

Why don't we offer them five on 25?"

1:01:50

Has that actually been a problem for you at all?

1:01:51

Have have you are are the like the bigger funds the the crazier dynamics is is that is that a trend that uh early stage VCs should actually be worried about? No.

1:02:00

I think Josh's playbook is he meets them he buys 10% before anybody knows about it and then says hey you know you guys you know take them on a road trip.

1:02:10

They're not ready for the big funds.

1:02:12

They need about a month to kind of get and stuff like that. Yeah, that makes sense.

1:02:18

So, it's not it's not strictly like they wouldn't be sniped even if they could be found because they're just not they're just not in a place to do those types of rounds. That's interesting. Makes a lot of sense.

1:02:27

Uh well, thank you so much for joining. Uh this was fantastic.

1:02:29

Great to hear the update, Josh.

1:02:30

And uh yeah, congratulations on the new fund. Thank you. See you guys soon. Byebye. We'll talk to you soon.

1:02:41

Well, that's a great uh cue to talk about Linear.

1:02:42

Linear is a purpose-built tool for planning and building products.

1:02:46

Meet the system for modern software development.

1:02:48

Streamline issues, projects, and product road maps.

1:02:50

Jord's been using it for the past 35 years. He loves it.

1:02:56

I wish fantastic uh planned his whole years before I, you know, was in the womb.

1:03:02

I was I was running on Linear.

1:03:04

No, you know, a funny little anecdote.

1:03:05

Linear sending us some merch. Oh, yeah.

1:03:07

And and the the automated email to notify us that they were sending us merch that's not even available online was like perfectly designed. Designed. No way. That's crazy.

1:03:20

Like it was truly incredible. I'm like, "Okay." Yeah. Culture of design.

1:03:24

Every single touch point.

1:03:24

We should also talk about numeral numeralhq. com. Sales tax on autopilot.

1:03:31

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

1:03:32

It's sales tax AGI, John. It is.

1:03:34

We got to get uh we got to get blueprint on numeral.

1:03:37

Yeah, we should just spend this entire pitching him bit.

1:03:41

What are you doing for sales tax?

1:03:43

What if he's just like I'm already on Numel? It's possible.

1:03:44

It's a highly efficient business.

1:03:46

I'm excited to talk to him about it. Uh benchmark series A.

1:03:50

Um anything else you tracking on Poly Market these days? I saw.

1:03:53

Um I'm still obsessed with the world's largest tech company.

1:03:58

Nvidia is surging right at the same market cap as Apple but uh um Nvidia has the uh the momentum. Yeah.

1:04:11

Largest company end of May it's the gap has been closed John.

1:04:13

Oh yeah it's still so Nvidia is at 42%. Microsoft is at 50%.

1:04:22

And uh Nvidia has a lot of momentum. Oh yeah. Nvidia is 42% now. Wow.

1:04:26

who really are ripping today.

1:04:28

They were at 20 something just yesterday. Good to be Jensen. Good to be Jensen.

1:04:35

I I was I was posting this that that if Nvidia does it, if they pull it off and they're the biggest company, they're going to have to teach Jevans parad $60 billion.

1:04:43

It's only a $60 billion gap. Jens one more.

1:04:48

Jens one more. He needs to he needs one more one more deal with conference in the global you know he's been on the conference circuit but it's working like he seems to be at a new conference every single week but uh clearly he's getting

1:05:02

business done at these events and he's striking deals and I'm sure it's like you know he meets someone super important like the head of a country and then says my people will talk to your people and we'll get and we'll ship you a 100,000 GPUs. Yeah, he definitely has

1:05:13

Yeah, he definitely has envelopes or not not he has napkins with with full fully fleshed out contracts on them.

1:05:22

You know, they just he just How many zeros do you want behind the one of the number of GPUs you're ordering? Yeah. Of hoppers. Um, fantastic.

1:05:27

I don't know if Brian's here yet. Let's check on him.

1:05:33

Um, and yeah, I mean, maybe we can do some timeline while we wait for him.

1:05:38

We got some posts we can talk about.

1:05:38

Uh, Mischief collabed with Mercedes.

1:05:40

This is a fantastic looking collab.

1:05:43

They created a trash can with pedals.

1:05:47

Seeing this almost made me angry because I want I want every piece. Yeah.

1:05:54

You were like you were like, "Oh, like what did they do? This is so frustrating." Yeah.

1:05:59

Like did they did they do something that we were thinking about doing or something? No. No.

1:06:02

Just incredibly well executed.

1:06:05

You can kind of scroll through these and I guess they're going to be selling.

1:06:08

There's an AMG coat hanger that looks like a fantastic coat hanger to hang your jacket on.

1:06:15

The couch is just awesome.

1:06:19

Uh, we got Brian Brian studio.

1:06:19

Welcome to the stream, Brian. How you doing?

1:06:22

Hey guys, how's it going? It's great. It's great.

1:06:26

It's great to have have you on the show finally.

1:06:27

We How'd you sleep last night? Really well.

1:06:29

Actually, no, that's not true. Um Oh, no. No.

1:06:31

I just started a new sauna protocol. Okay.

1:06:33

And 200 degrees Fahrenheit. 200 degrees?

1:06:39

Wait, what what's a sauna that we do?

1:06:41

Isn't it like 130 or something? That's crazy. Uh, yeah.

1:06:44

I mean, I can get my sauna.

1:06:46

So, I've been nerding out about saunas for a decade now.

1:06:48

And and currently, my sauna can get up to like 165 and then it'll trip like a a heat limit.

1:06:58

So, I'm not I'm not even able to get to that range yet.

1:07:00

You're in the minor leagues.

1:07:02

Yeah, it's in the minor leagues. Okay. So, 200 degrees.

1:07:04

What else is special about the new sauna routine? Yes.

1:07:06

I I found I read this yesterday. Let's see here.

1:07:08

So, a Yeah, at 200° it takes about 20 to 30 minutes to bake an egg until the white is set and the yolk is still soft.

1:07:17

So, yeah, it's hot cuz I' I've done seven days now and my sleep has been disrupted.

1:07:27

So, I'm still getting I'm still getting good sleep scores.

1:07:29

I'm my total time in bed, but my nervous system is is pretty high. So, Sure. Yeah. So, I'm going to wait.

1:07:36

So, so what do you have to Are you wearing gloves?

1:07:38

Are you wearing like any type of protection on your head at 200 degrees?

1:07:43

What What do you need to actually make it not, you know, cause your hair to fall out or something like that? Yeah, exactly. So, today's day eight.

1:07:49

I think I have the system sorted.

1:07:51

So, I I sit on this um this ice pack and then I have another one that I put inside my shorts to cool the testes because it does lower sperm quality. Sure.

1:08:02

So, make sure the the balls need to be cooled.

1:08:04

And then I have another ice pack laid over my head uh because it can make hair brittle and yeah cause scalp irritation.

1:08:10

So those seem to be adequate.

1:08:13

People used to never people used to never believe me when I said I had I got an ice pack on Amazon specifically for use in the sauna that is like a more like a cup.

1:08:21

They stopped they discontinued them.

1:08:23

I think it was just me and Brian buying them on there.

1:08:26

The special cup ice packs.

1:08:26

So where did this idea come from?

1:08:29

Is there new research or did you just discover old research or I mean I imagine that you want to know that there's some sort of data out there about the effects of this or are you just coming up with this idea on your own?

1:08:41

idea on your own? Yeah, I mean our process has been we've rolled through all the scientific literature and said what things have the very best evidence to improve health span lifespan and then we've just gone systematically down the list and so sauna's always been a target

1:08:54

of ours but most the data in sauna is observational which is a lower grade of quality data than interventional where you say we're going to do this thing and it's going to solve that you know whether you're going to see whether it has an effect or not and so we just finished up with hyperbaric oxygen

1:09:09

therapy and now we're pointed at sauna so we completed about 40 measurements for sauna baseline and now we're testing like even 7 days in like I'll tell you I just did this analysis this morning my um I had a five 6% drop in my central blood pressure a tw uh which is it's it's an improvement a 21% improvement in

1:09:28

my central pulse pressure uh 50% improvement in my augmentation pressure and 25% improvement in my uh augmentation index so we're looking at all these vascular age markers so even 7 days into it looks like it's having you some some meaningful effects in my uh vascular health. Uh how are you thinking about the

1:09:46

Uh how are you thinking about the potential of hyperbaric oxygen chambers?

1:09:50

It feels like the next frontier, the people that have a cold plunge in a sauna at home are probably, you know, thinking about adding one.

1:09:56

You've you've used them quite frequently, but but uh is that is that the next uh trend that that you're expecting to see broadly on the hardware side?

1:10:06

Yeah, hyperbaric was the best therapy I've ever done. Wow.

1:10:12

And so we've we've tried so many things and the what was unique about it, it's whole body uh health improvement.

1:10:16

So it improved my brain and my gut and skin and inflammation markers and uh like uh protein markers that that cause Alzheimer's were down.

1:10:29

Um trying to think others.

1:10:29

Is there an equivalent to uh like altitude that you're at or are you just like in the stratosphere or is it like you're you're in you're in Los Angeles but you're effectively in Denver or is it more like K2 or Everest or is it like space? Yeah, exactly.

1:10:47

So what what happens is you it's a hard shell chamber and then you pressurize to two atmospheres which is the equivalent of about 33 feet underwater. Under water. Got it.

1:10:56

So it's yeah it's pressurized and then you breathe in 100% oxygen.

1:11:00

So when you breathe in oxygen the atmosphere is about 21%.

1:11:03

So you're getting oxygen through your body but at this high concentration plus the the um the the pressurization you basically flush your entire body with this 100% oxygen.

1:11:13

So it pushes it via the plasma.

1:11:16

So it gets it out to all the extremities and that has an effect of growing new blood vessels of getting oxygen to parts of your body which you just don't get at.

1:11:24

So it's that 90minute exposure to oxygenation to the entire body that which has this whole body effect.

1:11:30

It seems like a lot of the interventions that you do kind of have there's some that have power law outcomes like sleep, diet, exercise, and then there's some that are more on the long tail.

1:11:40

long tail. uh how have you thought about as just being like a science communicator relaying the relative importance of these because some of these seem like they're extremely expensive for a minor gain but it's worth it for you because you're running the experiment and yeah this is your

1:11:55

life uh but for other people it might be a distraction to say oh I just need a hyperic chamber and then I don't and then I can eat McDonald's right yeah exactly I mean I think the the big the big change for me was in the past few months I've been able to figure out how to compress everything I've learned into one biioarker. Mhm. And that I just Mhm.

1:12:13

And that I just didn't think that would be possible because before if you talk about health it depends on what era of guru you're in.

1:12:22

You know what sometimes it's like NR or resveratrol or cold plunge or sauna or like take your given you know celery juice or whatever.

1:12:28

Um and the marker I think that is actually durable over time is resting heart rate before bed.

1:12:34

M and so the reason why this is so good is it basically is if you show me your resting heart rate I'll see your soul.

1:12:41

It's an accounting of your life decisions.

1:12:44

And so the thinking is this that like so tonight when you go to bed if you lay your head on the pillow and if you have a wearable just take a few deep breaths and then look at your resting heart rate.

1:12:55

If you don't have a wearable just take your pulse you know do it for six seconds times it by 10.

1:12:59

Let's say your resting heart rate is 60 beats per minute.

1:13:03

Your singular goal now is to drive that down.

1:13:06

And so over the next week or two weeks, you want to have that be 55 beats per minute instead of 60.

1:13:12

And when you see that happen, your sleep is going to be improved.

1:13:14

And when your sleep is improved, you're more likely to exercise.

1:13:19

When you exercise, you're more likely to eat well.

1:13:20

When you eat well, you're more likely to.

1:13:22

So it creates this really positive domino effect.

1:13:23

And the other thing it does is it it's a check against behaviors that drive your resting heart rate up at night.

1:13:30

So, a lot of our worst behaviors are between like 5:00 p. m. and midnight. We're tired. We're stressed.

1:13:37

We want to just be left alone because it just all feels too much.

1:13:41

And often times, food can be a soothing mechanism or other things, you know, a drink.

1:13:45

So, activities in that late night, even though they seem soothing to us, they drive up our resting heart rate and that necessarily reduces your sleep quality, which then reduces your your likelihood to exercise.

1:13:57

So I basically I now uh work with all kinds of people helping them understand which things increase heart rate and which things decrease heart rate.

1:14:06

But once you build your life around that it solves all of the fundamental problems with health that you want at a basic layer.

1:14:12

And then once you get that solved you can then of course ladder up and do other stuff but really that's it.

1:14:16

Like that is the marker that you can peg all your electricians on.

1:14:20

It's a really great starting point. Great.

1:14:24

What ingredient are you worried that we'll dis that we're consuming a lot of uh that the sort of maybe broader biohacker community is consuming a lot of today that in 15 or 20 years from now we'll discover is is actually uh terrible for us because we've you know we've done that we've done that a bunch of a bunch of times in the you know fat is probably the the biggest one.

1:14:41

is probably the the biggest one. Um but but is there anything that that comes to mind that you're kind of wary of that eos you don't you're maybe not 100% convicted on yet but you're you're starting to watching it spend more time thinking about I'll tell you we we got

1:14:55

pretty surprised I was taking rapamy this is considered by many in the longevity field as you know a very promising drug and if you I mean anti-aging scientists disagree on almost everything but there was probably more consensus around rapamy around than anything else out there and I was taking it for several years. We meticulously

1:15:14

We meticulously tracked my blood levels to make sure we were on a a good dose.

1:15:17

But we all we stopped last year uh because the side effects were too significant like lipids were being messed up.

1:15:23

My blood glucose had some alterations.

1:15:25

I had soft tissue infections and then a month after we stopped a paper came out that showed that 16 epigenetic clocks showed that it actually accelerated age.

1:15:34

So this is not the end all beall to say rapamy is toast.

1:15:39

This is just to say that it's not like a slam dunk easy one. It's complicated.

1:15:44

And so we stopped I stopped taking that that drug, but I guess like it's a good example to your point of we I think we're in the very early stages of truly understanding anti-aging and biology and certain interventions.

1:15:56

I think we're good on sleep and exercise and you know eating nutritious food, but even then that branches out into some complexity.

1:16:04

But I'd say rapamy is probably to your point a really surprising outcome for where cons like emergent consensus has been and now we're not doing it anymore.

1:16:14

And so there's a few other drugs we've been looking at.

1:16:16

We have a similar path we're looking at right now, but that's probably one of the bigger ones that I would identify.

1:16:20

Do you think uh anti-aging will be solved in some form via pharma or a collection of lifestyle decisions, right?

1:16:32

Because in many ways lifestyle decisions today, you know, allow you to have a longer health span, lifespan.

1:16:38

Uh but in in some ways, you know, we had the the founder of of New Limit on on the show last week, which is working on longevity.

1:16:46

In many ways, it feels like if we can all just extend our lives long enough to get the, you know, uh the the sort of pending miracle drug that, you know, adds whatever another hundred years, 50 years, what whatever that that number ends up being.

1:16:59

But I'm curious how you look at it. Yeah. I Yeah.

1:17:03

I think we need much more robust interventions.

1:17:07

The the the type that Brian's working on with New Limit.

1:17:09

Did you guys have Brian on? Yeah. Yeah. Yeah.

1:17:13

I I love what he's building. I'm an investor there.

1:17:14

I think that that's a very promising path.

1:17:16

And what a lot of people I think don't understand with what I'm trying to do is I'm not suggesting that one can jump past 120 or find a path to not die by going to sleep and exercising and eating well.

1:17:32

I'm suggesting that this is the most disruptive moment in the history of the human race.

1:17:37

We're giving birth to super intelligence.

1:17:39

It's going to transform society in ways we can't understand.

1:17:43

And in this uh approaching an event horizon, I'm suggesting literally the only thing we know is that none of us want to die in this very moment.

1:17:54

And so I'm really trying to give birth to a new ideology, a new way to understand ourselves as a species.

1:17:58

And but you know, ideology and philosophy is very hard for people to engage with.

1:18:01

So I speak with about things that people understand like going to bed on time and not eating junk food and you know not basically doing things to slowly kill yourself that in a few years time I think we'll look back at this moment and we'll say that was insanity in the ear you know before the early before the late 2020s humans used to do things that was some form of slow suicide. Can you believe that?

1:18:26

That is just incomprehensible.

1:18:28

So really, I guess I'm I'm trying to get at this from a a philosophical, ideological, spiritual, cultural perspective uh by the practices I'm doing now, but I'm I'm I'm grateful that people like Brian are working on the more robust interventions that actually will meaningfully change uh how we age.

1:18:46

Are you worried there's not enough companies like New Limit, you know, taking on this monumental task?

1:18:51

It it feels like, you know, at least within the tech community, there's a bunch of ways, you know, there's there's a million easier ways to to make, you know, hundred million dollars than longevity. Yeah.

1:19:04

And that automatically, you know, people self- select out of that path. Yeah. Exactly. Yeah.

1:19:09

I I I'm not surprised at all.

1:19:12

And I my di my diagnosis of the situation would be that currently the strongest incentives are pointed at the accumulation of wealth, power and status.

1:19:23

You know, if you look at our day and say how much effort do we put behind those things and we do that because when death is inevitable, we've chosen a yolo game of we're going for epicness, right?

1:19:35

of a of achievement or of being remembered for a given accomplishment or of status and power in society.

1:19:42

And we just haven't crossed over that threshold yet where people really believe that we can begin to dramatically extend our lifespans.

1:19:51

And so if if AI progresses to a certain point where it even opens the window or or an intervention makes a breakthrough where it's like huh is this moment really like you know people look at the health and wellness stuff they're kind of like I don't really care like if I lose an extra 10 or 15 years off my life whatever I don't want to live those live those years anyways I'd rather have preferred life now.

1:20:16

So, I think it's just we're just really caught in a cultural moment between the death being inevitable and don't die.

1:20:23

And it's it's happening right now.

1:20:26

We're starting to cross over, but I think there will be an event or a few events in the coming years, we'll change it and the humanity is going to snap and say, I think this might be it.

1:20:34

Like, and I think we'll have a big cultural shift.

1:20:36

How do you think about like this internal motivation as something that can overpower an unhealthy lifestyle?

1:20:42

I'm thinking about Warren Buffett drinking six Coca-Cas a day kind of rejection of everything that we know about health and wellness.

1:20:53

And yet he's been phenomenally successful and performant in a very stressful important job for what seven decades or something like that.

1:21:02

Uh Donald Trump's in a similar category.

1:21:07

the it seems like there's something to the motivation and like the human will potentially, but is that just all pseudocience in your mind? Yeah.

1:21:13

I mean, people they they do reference Buffett a lot, right?

1:21:19

And and it's like also it's like someone's grandma who lived to be one or two who drank and smoked every day. Yeah.

1:21:25

And you know those examples are just not practically relatable, right?

1:21:30

like those those are genetic differences uh that they got they won the genetic lottery and they lived despite their choices.

1:21:36

But if most people do what they do, they would die prematurely.

1:21:40

But I think it really comes back to the ideology that that Warren Buffett played a game in life about asset accumulation, right? And he won the prize.

1:21:51

I mean, he's arguably the best investor of all time.

1:21:52

So, he really played that game very well.

1:21:54

And so the question is if if you were to basically take a Warren Buffett like persona today, what would they do?

1:22:00

Would they try to amass assets or they try to amass existence?

1:22:05

And so what I'm suggesting right now is the the highest ROI investment anyone can do is to stay alive.

1:22:13

And that's what I'm trying to do is I'm I'm trying to basically I've attempted to become the healthiest person on the earth, like the, you know, with the best biomarkers to demonstrate a new game of status and accomplishment is here.

1:22:26

And it's about health and wellness and it's about your ability to exist.

1:22:29

And so this will take time because it's a massive shift in our culture culture, but really it's just like we've all whether we've agreed to this explicitly or not, we've all agreed to this wealth, power, and status game.

1:22:41

and we play it independent of whether whether we're aware of it and I'm just trying to say that that is a game it will shift and a c a new cultural emergence will will arise and don't die will be the prominent thing we'll focus on uh last question I have do you think

1:22:54

people pay too much attention to scaled data sources like you know the concept of blue zones right is it should individuals be focused on you know highly personalized health and the approach that you're taking which is you know sort of real- time experimentation

1:23:12

rapid testing because even you know I'm curious if you have any type ofsis here but uh you know I think many people have said that you know Japan having this you know immense longevity might actually just be pension fraud uh and I don't you

1:23:26

know because people have an incentive to like that's funny people have an incentive to say like yeah grandma's still kicking you know keep giving us the you know 10k a year whatever I'm not making any accusations myself but um you know retire next year. People love to

1:23:39

People love to read into, you know, oh, this region, they're drinking two glasses of wine, so I should drink three and I'll be okay. Yeah. Yeah.

1:23:50

I mean, what you're touching on is a real problem where if if somebody encounters the idea of wanting to be in good health for whatever reason, like let's just say uh a person wants to be hot, right?

1:24:02

They want to look good and be hot.

1:24:04

Then the next question is what do you do?

1:24:06

And once you step out into the world and you say, "What do I do?"

1:24:10

You're faced with hundreds of different opinions and they're all in conflict with each other.

1:24:15

And then if you take another scenario where somebody's got something that they're dealing with, they have constant headaches or they have sleep apnnea or they have an autoimmune disorder like or take your your thing, they've got a complicated situation. What do they do?

1:24:26

And so up until this point, you've had to go about and navigate the world of experts.

1:24:30

you know, now these AI tools are better than uh almost all the experts in helping people solve their given problem.

1:24:37

And so the next thing to close this loop is now that we've got a source of advice that is reliable and that everybody has access to is now really a function of measurement is can you actually get data that then improves the AI's ability to make very precise assessments.

1:24:51

make very precise assessments. And so that's what I'm working on personally is I've become the most measured person in human history and I've seen firsthand when you have data it's incredibly helpful to figure out what's happening and when you don't have data you're in the dark poking around and so we're trying to figure out how to scale data

1:25:11

and access because once you pair that with the AI then the the question of what to do becomes so simple which again is why I believe we're at this moment of shifting dramatically where the complex lexity and the friction between someone saying I want to be healthy and knowing what to do is extremely short whereas now it's like so big and so complicated it's like honestly not doing it. So I

1:25:31

So I think we really are like this moment is so rich and has such high potential.

1:25:36

It hinges on data with closed loop with AI.

1:25:41

I think we're off to a new race. It's exciting. All right. Very exciting.

1:25:45

Well, come back on when you have more news. Yeah. Always welcome. This is amazing. Thank you so much. Yeah. Good hang. Be well. Cheers. We'll talk to you soon. Bye. Um, we got Matteo.

1:25:54

We got Matteo coming in in the waiting room.

1:25:56

Let me tell you about Vanta really quickly.

1:25:57

Automate compliance, manage risk, improve trust continuously.

1:26:01

Vant trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation whether you're pursuing your first framework or managing a complex pro program.

1:26:13

One of the fastest growing vendors on ramp. Oh, really? At scale. No idea. And believe that.

1:26:16

We're excited to have Matteo from Eight Sleep today. Massive day. Congratulations. Huge news.

1:26:23

I gotta update my ad reads.

1:26:26

I'm always talking about the Pod Four Ultra. It's the Pod Five today. Pod five. Congratulations.

1:26:32

And this is this is particularly huge for the podcasting industry because I've said this before, but you know, in the highstakes world of professional sports, sleep is extremely important.

1:26:44

But arguably a higher stakes world exists. Technology podcasting.

1:26:46

You need to get your sleep if you're going to be podcasting 3 4 hours a day.

1:26:50

And that's why we need sleep. You'd be proud.

1:26:55

I I legitimately get angry at John when his when his sleep scores are down. It's it's it's terrible. It's tearing us apart.

1:27:01

I take I take this game this podcasting game very seriously. Very seriously.

1:27:05

And I take John's sleep very seriously.

1:27:07

And uh you should Jord's been smoking me.

1:27:09

Uh but yeah, give us the update.

1:27:12

Explain to us what how the launch is going. What's the overview? Break it down for us. Yeah, of course.

1:27:19

I mean, first of all, at 8 sleep, we develop technology to improve your sleep.

1:27:23

That that's literally what we do.

1:27:24

And the way we do it is by focusing on controlling and personalizing all the environmental factors.

1:27:33

We started with temperature in the cover, which was a cover is a cover that you can install on top of your mattress.

1:27:39

It tracks everything about your sleep and health, and then it changes temperature to give you better sleep.

1:27:44

And then we took temperature to the next level now with pod five.

1:27:47

So you also get the blanket and you can sleep in this sort of microclimate.

1:27:51

And guys I I know I'm biased because I'm the CEO of eight sleeper. It's freaking cool.

1:27:58

I have been using this product for nine months and I have given it to some friends or people inside the company and they all love it.

1:28:06

Doesn't matter if you like it cold or warm. anecdote.

1:28:07

John's John's wife messaged John this morning when she saw the launch and she was like like just sending exclamation points like we have to get one. Yeah, she will love it.

1:28:21

Do you know if she sleeps hot or cold? Extremely. She's cold.

1:28:23

She wants it to be a blazing fire. An inferno. An inferno cocoon. Yeah.

1:28:28

So, so she will love it because now she will be surrounded by temperature.

1:28:33

Cold temperature everywhere.

1:28:34

It's like going in a freezer and closing the freezer and you just leave that there.

1:28:38

And I like my side freezing cold.

1:28:40

I'm I'm minus 10 and she's plus 10 every night.

1:28:43

But she says you can tell him I love him for inventing this.

1:28:49

Uh it's really like the greatest upgrade of the entire year to our house. It's fantastic. That's awesome.

1:28:54

And so temperature is one. Yeah. Right.

1:28:56

And we tackled that biggest factor impacting your sleep outside sleep medical disorder.

1:29:01

So there was plenty of evidence.

1:29:03

We didn't reinvent the wheel.

1:29:04

We were just the first ones to really commercialize a consumer product that would do something that it was already proven to work.

1:29:10

And then we got into elevation and so we introduced the base.

1:29:15

The base goes underneath your mattress and it raises your head if you're snoring and so you will see a drop in your snoring by up to 45%. Wow.

1:29:26

If you have problems of circulation with your legs, we raise your legs. Okay.

1:29:28

And the most important thing that very few people know, I don't sleep flat anymore.

1:29:35

Meaning sleeping flat is from the 80s.

1:29:38

You need to you you need to sleep in a position where you raise a little bit the head, a little bit your uh your your feet and it takes away any sort of pressure on your back. Yep.

1:29:48

Is snoring actually bad for you or is it just annoying to whoever's sleeping next to you? Sorry, say that again.

1:29:55

Is is snoring actually something you should be worried about or something you should be trying to mitigate from like a health perspective or like a personal sleep quality perspective or is it just annoying to whoever's sleeping in the bed next to you?

1:30:07

No, it's it's negative for you.

1:30:10

It negatively impacts your sleep architecture and it can limit the amount of oxygen that gets to the brain.

1:30:16

Because if you think the real true problem of sleep apnea, which is a form of extreme snoring, if if you will, is that when you have these episodes of snoring or sleep apnnea, less oxygen gets to the brain.

1:30:28

And so if your sleep apnea is very strong, the reason why you feel very tired in the morning is because you kept lacking oxygen in your brain during the whole night and your brain cannot rest. Mhm.

1:30:39

And so mitigating snoring is very important for your own health and is important for your partner so you don't bother them with with your noise. Yeah.

1:30:47

Now the latest base also comes with a speaker now. Mhm.

1:30:52

And the speaker plays soundscapes that you can play to fall asleep to stay asleep.

1:30:57

And then we have this partnership with Yubberman where we are introducing an SDR.

1:31:02

So non sleep deep rest is a sort of meditation and breathing.

1:31:07

You can play that with the with the voice of Andrew and then you just fall asleep and you sleep like a baby.

1:31:13

SDR I think sales development repres representative for Andrew Huber is just there pitching you on sleep.

1:31:19

Like you know I've talked to a lot of people that are considering sleep and I think it'd be really great for you tonight. Yeah.

1:31:23

I I think there's an opportunity to do an ad supported eight speaker.

1:31:28

We'll record ads and you can fall asleep to the sound of us just ads constantly. And even at at 1:00 a. m. Hey. Hey.

1:31:35

Have you heard about ramp?

1:31:38

Have you heard about ramp? Exactly.

1:31:41

I I I I can start playing ads for ramp. No, I'm joking.

1:31:43

Uh well, if they already have an eight sleep, we don't need to do an eight sleep ad again. Yeah. Maybe an Easter egg. Yeah. Yeah. Yeah.

1:31:51

When when you wake up certain days, there is a ramp ad.

1:31:54

By the way, my Tesla sometimes I start No, I go in the Tesla.

1:31:59

I use the radio very few times, but then there are ads on my Tesla. Oh, yeah.

1:32:04

When I want to use the radio. So, wow.

1:32:06

Everything's ad supported now.

1:32:09

You got to upgrade to the ad free version of your car.

1:32:10

It's a dark we've entered in that's a very black mirror, but the future it's going to be cheaper because it's going to be ad supported.

1:32:20

Uh what what uh what else have you looked at in the sleep world?

1:32:22

I know that you've launched uh some adjacent products.

1:32:26

I'm thinking like, you know, you're in blankets.

1:32:29

Are we going to see pillows?

1:32:30

Are we seeing, you know, supplements and other uh clothing and pajamas and all sorts of different things?

1:32:37

Like how do you view the sleep market?

1:32:40

And what have you gone after first and what do you see on the horizon?

1:32:43

Or even I uh like blackout curtains, right?

1:32:48

You try to black out try to black out your room is is uh you know there's yeah solutions. Yeah. What do you like?

1:32:53

What do you uh what have you already done and what's on the road map? Yeah.

1:32:56

So if we go back, you know, to the beginning, right?

1:33:00

We use AI and technology to improve your sleep.

1:33:02

So, anything that can improve your sleep is something we are going to do sooner or later.

1:33:05

Where there is technology is the top priority.

1:33:10

There could be other products that maybe they don't have embedded technology but that we develop based on our data because we are becoming one of the largest sleep labs in the world.

1:33:18

So, supplements we launch a first line of supplements that we develop with Peteratia and the reason is we have all these data we know what works, what doesn't work.

1:33:27

Sometimes you you read people saying, "Oh, should I take melatonin?

1:33:31

Should I take magnesium?"

1:33:31

We have run clinical studies.

1:33:33

But from a technological standpoint, I would say the two coolest buckets to me are one is environmental control and the second is body scanning.

1:33:42

And I'll tell you about both.

1:33:44

So environmental control is okay temperature with the blanket and the cover, elevation and sound.

1:33:48

Then you should imagine that yes there will be more uh thermorreulated products because temperature is uh is is the killer feature but we I want to control air, oxygen, light, noise, sound, sense, everything.

1:34:05

So everything that we can optimize because the key difference between us and the wearables is wearables do a decent job at tracking stuff but they don't do anything for you.

1:34:15

Our job at eight sleep is don't wear anything and we are going to do the work for you. Yeah.

1:34:19

And so this is the environment.

1:34:21

Then doesn't matter how long you sleep, it's still a lot of time.

1:34:25

And so the goal here is to work on body scanning technologies where we can scan your body every single night to detect if you are developing certain health conditions before you even know. Crazy.

1:34:37

And that is part of our R&D.

1:34:37

Uh have you been able to look across the anonymized sleep data and noticed any interesting trends?

1:34:45

I'm thinking right before uh tax day or how do people sleep before Christmas or is everyone excited?

1:34:54

Uh anything that you can tell us about uh how the global sleep community trends one night or another?

1:35:01

Do people sleep better during winter when it's cold or summer? I don't don't even know.

1:35:06

Have you learned any aggregated data? Anonymous of course. Yeah.

1:35:10

Yeah, we have a look at some of stuff again as you said aggregated anonymize.

1:35:14

So obviously we we don't know anything about an individual but like for example I think the city where the the town where they sleep the best is in northern Australia.

1:35:22

Uh we have seen people sleeping more after COVID meaning after COVID the trend generally changed.

1:35:30

Women tend to sleep longer than men.

1:35:33

Uh during the election day people went to sleep like a couple of hours later. Oh wow.

1:35:38

And then we can start tie it or connect it to certain specific events, right?

1:35:43

Even sport like Super Bowl, all that kind of thing, which is pretty cool.

1:35:48

Or you see it if there are bad days in New York and is there no day in New York, all that kind of thing.

1:35:55

Um, and it's pretty interesting to see these patterns this in terms of I would say consumer behavior.

1:36:01

Then with AI, our models are reaching a point where are smarter than us and they start connecting the dots of things that we don't even medically know.

1:36:12

Meaning let's say that you join it sleep and and let's say you you you share that you have certain health conditions.

1:36:19

Our model start understanding that uh there could be based on these parameters you might have a certain likelihood to develop a certain health condition.

1:36:28

It can go could be completely disconnected from sleep. It could be diabetes. It could be Parkinson.

1:36:34

Uh it could be a a neurological disease.

1:36:36

But by correlating all these aggregated amount of data, hundreds of millions of nights of sleep, we can predict the future of health.

1:36:49

Uh I got to ask, how did the Charleslair partnership come together?

1:36:51

Was that a was that a two plus year kind of you know let me start with a cool story and then I'll I'll tell you but so the day we sign so I go to Monaco for the day we sign and I'm at the hotel and we need to go to this place where we sign and then we do the first photo shoot and so Charles comes to pick me up and he comes in the Ferrari because he drives the Ferrari and so I said man you're the you're the best Uber driver I ever had by far. Yeah.

1:37:26

Being driven around Monaco by Charles is is a good good fantastic everybody should do it once in their life, you know. Oh yeah.

1:37:37

Um the interesting thing is Charles bought the pod uh two years before the partnership. We didn't even know. That's always the best.

1:37:44

And yeah, it's always the best.

1:37:46

And so when we got connected two years later, he already knew about us. He he loves the product.

1:37:51

And when I pitch him our health vision, health is very important for him.

1:37:56

And he's like an AI geek.

1:37:59

You would never expect that, but he reads everything about AI.

1:38:01

So when we meet, we never talk about Formula One because I don't want to bother him talking about work and we just talk about AI and health. Very cool.

1:38:13

And so when we met, we had dinner.

1:38:15

It was actually before the the F1 GP in Miami last year and it was a Thursday night and before the end of the of the of the dinner we said look we we want to work together. We'll figure it out.

1:38:25

Uh I'll I'll do some steps to I mean I'm I'm sure you you have a decent cost and you will make some steps uh towards me because I'm a startup and I don't have money.

1:38:36

Um and within a month the deal was done. Wow. Quick. That's awesome. That's amazing. Well, yeah.

1:38:42

I uh I I get uh you know, I'm not excited if I have a bad night's sleep and I got to go podcast for, you know, four hours, but uh I can't imagine getting a bad night's sleep and then needing to go uh race track at, you know, hundreds of kilometers an hour. So, brutal. Makes sense.

1:39:01

Well, congratulations to to you and the team.

1:39:03

It's uh an amazing uh milestone and uh we are, you know, Yeah. stoked to be partners. Proud to be partners. Yeah, of course.

1:39:11

You guys are doing a great job, so I'm a big supporter and uh thank you for everything. Thank you so much. You're the man. We'll talk to you soon. Thanks for coming on. Talk soon. Bye. Cheers. Bye. Fantastic.

1:39:26

We love a big product launch on TVPN. Fantastic. So good. Uh we also love adquake. com.

1:39:33

Out ofome advertising made easy and measurable.

1:39:35

Say goodbye to the headaches of out ofome advertising. Say see you later.

1:39:39

See you later to those headaches.

1:39:41

Only AdQuick combines technology, out of home expertise, and data to enable efficient, seamless ad buying across the globe. Simply too good.

1:39:50

Um, and we have we're going into the Andre Horowitz LP meeting.

1:39:57

We sent one of our producers out there set a uh a Zoom room.

1:40:02

Hopefully, it looks really nice.

1:40:02

And we have an absolutely stacked lineup of Andre and Horowit's general partners and some operating partners.

1:40:08

And we should be able to take you on a whirlwind tour of what's going on in Andre World, what they're investing in, uh what trends they're looking at, and what deals they've done because uh we really have a full full list here finishing out with Mark Andre.

1:40:23

Um but this should be a lot of fun.

1:40:25

Um Eric will be kicking it off.

1:40:28

He's been on the show twice.

1:40:28

Uh we we we broke the news here. Not really.

1:40:31

He came on the show shortly after announcing that he was joining Andre Horowitz.

1:40:34

Um and uh he's been a close friend of us of ours for years and it's always fun to talk to him about media and venture and tech and all of those things.

1:40:44

So he'll be kicking us kicking it off with us, breaking down how the transition's been, what it's like working at Andre and how he's settling in.

1:40:53

Has he ripped any huge checks?

1:40:54

We're going to get to the bottom of it.

1:40:56

Um, so we will bring um him in in just a minute.

1:41:00

And in the meantime, why don't I tell you about bezel? Go to getbzzel. com.

1:41:04

Your bezel concierge is available to source any watch on the planet. Seriously, any watch.

1:41:09

Mark Andre's known for wearing what?

1:41:11

What does What did he wear on Rogan?

1:41:14

He wore a uh which one was it? Omega. An Omega.

1:41:20

The watch of James Bond, the Speed Master.

1:41:22

the recently voted by Nico Leonard as the most iconic watch in history.

1:41:26

Um he had a whole debate with all of his team, the great YouTuber about watches if you're interested in following him.

1:41:32

And uh uh so maybe we'll see what what Mark's wearing.

1:41:35

Maybe we'll see what Eric's wearing.

1:41:37

I don't know if we'll get into watches.

1:41:38

Uh we always uh we always try and spice it up, but some people some people want to stay on message surprisingly.

1:41:44

They want to they want to focus on business, but you know, we try and throw them off.

1:41:49

Try to throw TJ off with some car talk.

1:41:50

He wasn't really having it, but we'll get to him one of these days.

1:41:55

We'll bring him back for a car segment eventually.

1:41:58

Um, anyway, let's see if Eric is ready. We do.

1:42:01

We have a red, green, blue, yellow card. Let's see. Yellow.

1:42:03

We're We're almost ready.

1:42:06

Um, so let me tell you about Wander. Find your happy place. Find your happy place.

1:42:14

Book a Wander with inspiring views, hotel grade amenities, dreamy beds, top tier cleaning, and 247 concier service.

1:42:20

It's a vacation home, but better, folks.

1:42:22

The other thing that we could talk about is uh four years ago, The Verge wrote of Elon Musk's Tesla bot announcement. Don't overthink it. It's a joke.

1:42:33

It's a distraction and an empty promise.

1:42:35

This was the event where Elon hired a dancer to come out dressed as a robot and is dancing around and it looked very silly.

1:42:42

And Elon is saying LOL because now um of course the Tesla robot is dancing and it's clearly very real.

1:42:51

But somebody was saying it dances like Brian Johnson. Really?

1:42:53

Brian Johnson's got that he's got that very specific style.

1:42:57

The the interesting thing is that um you have to know know that like Elon likes humor.

1:43:02

He likes telling jokes and and he also knew what was capable.

1:43:08

He could have just white label.

1:43:10

He could have bought all of Boston Dynamics and had a dancing robot on stage, but he chose to put a human in a suit probably to troll people, I imagine.

1:43:19

Like that had to have been he had to have known what would happen that people would see that and be like, that's obviously a human in a suit that looks goofy.

1:43:27

And so, you know, we know that you didn't do a real robot, Elon. This is a joke.

1:43:32

So he almost I think I I feel like he was baiting the media a couple years ago because what he was proposing like a humanoid robot was not impossible by the laws of physics.

1:43:40

Elon of course lives by the laws of physics and he gets to take his victory lap today of course.

1:43:45

Um anyway uh what else is interesting going on?

1:43:49

Oh HBO has rebranded again. Max is now HBO Max.

1:43:51

Um and Derek Thomas this is so that was absolutely brutal. I mean it was crazy.

1:44:01

It was crazy that they went from HBO Max to Max.

1:44:03

They rebranded three times as Max.

1:44:05

They have three different logos and now they're bringing back HBO Max. They're maxing. They're HBO Maxing.

1:44:12

Somebody said back then, and I I think this was on point, was that uh D HBO is such a good brand. Yeah. To drop it is crazy.

1:44:23

It was always It was always a crazy crazy move.

1:44:24

But why why even call it HBO Max?

1:44:26

Why not just call it HBO?

1:44:29

Yeah, they should just call it HBO. the max.

1:44:31

Well, that's what ESPN did recently.

1:44:32

Uh it was in the it was in the business section today.

1:44:34

Uh ESPN prices its new service at $29.

1:44:37

99 a month and it's called ESPN. What a concept.

1:44:39

You know, the internet is as important as the linear TV over the over the air TV.

1:44:47

And so make your OTT uh option just the same name. It's it's genius. It's so basic.

1:44:51

Uh Derek Thompson was saying, "If Max is paying a marketing company a million dollar annual retainer for naming strategy, this is pathetic.

1:44:56

But if Max figured out that it can add and remove HBO from its streaming title every two years to create a bunch of free outrage publicity, I sort of respect the play. So, who knows? Is it 3D chess?

1:45:07

Are they playing 4D chess? Who knows?

1:45:09

Anyway, uh we got Eric Torrenberg from Andre and Horowitz in the studio.

1:45:14

Welcome to the stream, Eric. How you doing? Third time. He's back. Three Pete.

1:45:21

Three Pete champion and kicking off a record.

1:45:23

I have has eight Andre and partners ever been on one podcast? I don't think so.

1:45:27

I said Nixon went to China, Torrenberg goes on TBPN.

1:45:34

Yeah, look what happened. Look what happened.

1:45:35

Uh you open the floodgates. Uh it's been fantastic.

1:45:37

Uh how has the transition to Andre been?

1:45:40

We asked you when you joined uh what your plans were.

1:45:44

Were you hitting the ground running?

1:45:44

Are you ripping checks on day one?

1:45:45

Uh you're a couple weeks in the gig.

1:45:47

I expect some sizable checks to be written. What you got for me?

1:45:51

I feel like TVPN is my uh you know, what did you get done this week? You know, makes time.

1:45:55

It's my third week, fourth week.

1:45:57

And um no, it's it's been amazing.

1:46:00

I mean, just the the scope of the firm. Uh almost 600 employees.

1:46:04

I mean, I don't I think people fully appreciate all the uh the the amazing things that that are happening here.

1:46:10

I've been spending a lot of time in the different deal meetings at the different funds, just getting up to speed and how we operate, meeting the meeting the broader team, starting to build out this this media org.

1:46:18

And um I feel like you know I talk to employees at OpenAI all all the time.

1:46:23

They're like we've got so many things cooking and I I feel like there's a little bit of that here too.

1:46:28

Just Ben and Mark are so entrepreneurial and and that leads onto the different vertical fund leads that you're going to hear you everyone's going to you're going to chat with today and people are going to hear from and um yeah it's just been incredible. Yeah.

1:46:38

Is is Andre the biggest venture fund ever in history now?

1:46:43

I can't imagine there's a bigger one.

1:46:46

I think if you count the original ships that went looking for whales harpooning, they probably had like something like a thousand spread across maybe a few ships.

1:46:54

Yeah, because I mean there's like this narrative in venture generally that like Andre's really big and no one's ever done it at this scale before and like maybe that's weird or different or good or bad or whatever.

1:47:03

whatever. But the other way to think about it is like think about the other major financial institutions that are huge much bigger sometimes an order of magnitude or two orders magnitude bigger and are they in founder mode or are they like 200year-old behemoths and all of a sudden when you frame it like that I become very bullish on the founder mode

1:47:24

organization even though all there's obviously all these competitive dynamics and different strategies but it is very interesting to see a new financial institution we got this from when we were talking to David on the Um uh and uh it it was just very interesting to see that uh this is David Haber uh that that he he came from Goldman uh worked on firmwide strategy there. Told us some

1:47:43

Told us some of the history of how Goldman had developed all these pods.

1:47:47

I was doing a deep dive on their uh their special situations group and how that spun up and spun out during the OA crisis and like it's such a storied firm but I don't think you can get there by maintaining some small size.

1:47:59

You have to become an institution.

1:48:01

That's exactly what what what what you've done and I think Mark's core insight, you know, self-resing the world and the sort of the implication behind that that the winners are going to get bigger and bigger and there's going to be more of them is just perfectly coincided with with the firm going super big and you need a lot of capital to to support that strategy and it's it's worked out tremendously. Yeah.

1:48:19

I mean yesterday we were talking about this weird dynamic where uh you would expect the power law to apply to startup outcomes and we've seen this from time and time again.

1:48:28

is very obvious and it also just makes sense based on market dynamics, how you know power acrru etc.

1:48:34

But um venture capital funds also follow a power law and and and and you would imagine that also the size and scale of the firms might also follow a power law as there are increasing returns to scale.

1:48:47

So it's been interesting to see kind of that play out and in many ways the power law just keeps applying no matter what lens or frame you take there's like there's going to be a big outcome.

1:48:56

Uh anyway, Jordy, anything else you want to run through?

1:49:00

Can you break down LP day?

1:49:00

Like what are what's actually going on today?

1:49:04

What are we zooming into? Uh break it down.

1:49:06

The thing I'm trying to triangulate is what's what's the next coinage?

1:49:10

You know, what's the next American? We got to break it here.

1:49:14

That's the trillion That's the trillion dollar question.

1:49:15

Or or it'll drop as a big piece either in the Wall Street Journal or on a blog and then we'll we'll get you the same day to break it down. Yeah. Yeah. Um 100%.

1:49:22

down. Yeah. Yeah. Um 100%. And um you know today and tomorrow is going to be us presenting sort of you know one our our biggest portfolio companies or some of our best ones that you you'll you'll you have Alex tomorrow coming on the

1:49:36

show from Worldcoin but then also the different leads of you know Jen who who runs uh LP day is is coming right after this and and also just our fundraising uh I believe over $40 billion um you know to to date um but then also the the heads of the different vertical funds 40 billion. That's huge. That's That's huge.

1:49:53

That's the most deserving.

1:49:56

That's the biggest size gong ever, I think.

1:49:57

Uh maybe beaten out by Stargate.

1:50:00

But uh but other than Stargate, you know, you're up there all in Jen, you know, sort of go deeper on it.

1:50:08

But also just um David George who leads our growth fund which has done phenomenally well.

1:50:12

Um you know, Martin on the on the infra side, an who works uh AI AI under underneath him.

1:50:17

Uh Du Dave Yulovich on the American Dynamism side.

1:50:22

Um, and and the LP day is really just us talking to our LPs about uh, you know, what what is the latest in terms of how we're doing, what what we believe, what's happening in the market, where we think things are going, and it's really giving them kind of an inside view, and we're going to release some of the the best talks.

1:50:36

Um, but and of course, you're going to hear from Mark Andre himself about uh, about where we are, where we're going.

1:50:42

In terms of uh kind of like advice for founders, what is the role of an LP day in the life cycle of a startup CEO that becomes a fund driver?

1:50:52

I've seen some times where founders go to these LP days and then couple of months later there's a direct investment from one of the LPs.

1:51:00

Is that part of the dynamic or is it really just an opportunity to kind of uh come and and share like the story of the business with the wider business and investment community? No, 100%.

1:51:09

So we have a very influential uh LP base.

1:51:12

Jen will get more into specifics and it and it's we're more diversified than your average vent firm.

1:51:18

So we have um LPs that um you know mo a lot of firms really focus on on endowments.

1:51:23

We've got some incredible endowments as well.

1:51:24

Um but we also have a lot of other partners including big international partners and a lot of companies are looking for you know especially some of these uh AI companies uh you know will tell you about sort of the open source AI companies that he's worked with looking for these global global partners.

1:51:38

So, so yeah, you know, we have worldcoin, etc.

1:51:42

So, there are a bunch of companies that have international presence that are growing their international presence and want partnership from from from big players uh globally and some of them are here and yeah, we we're definitely looking to Yeah, I mean we saw that yesterday with Nvidia.

1:51:54

Jensen went to Saudi Arabia, announced some big deal, the stock popped.

1:51:57

It might be the biggest company in the world.

1:51:59

biggest company in the world. is so back and you know Jensen gets invited to these things but if you're you know two years into your entrepreneurial journey you might not get a seat at the table but an LP day could be an opportunity to meet someone from that community correct

1:52:14

yeah yeah no 100% Ben Haritz was was also there as reported which is why he's uh he's not joining TVPN today but more uh we'll give you the the the scoop there perhaps Jen but um but yeah no there's there's there's a lot going on in terms of uh international partnerships So expect to hear more from us on that front. Yeah. Now I'm sure the Yeah.

1:52:31

Now I'm sure the fund is aligned on most of the strategies and kind of topics and hot takes, but there's probably a lot of internal debate as well around how the these future tech trends will play out, especially the things that haven't been codified or really decided.

1:52:48

codified or really decided. Uh I are there any topics that you're hearing the partnership kind of go back and forth on or debate internally whether it's future of AI or how geopolitics plays out or anything re-industrialization different

1:53:03

strategies what is keeping the partnership uh you know excited in uh in big partner meetings I'll give you a couple um so so one is more of a meta uh topic which is this idea of you know a lot of people say a lot of things about us and how much should we engage you

1:53:22

know should we just focus on the work just just working with founders just just you know uh driving great returns for LPs or when do we correct the record when when do we step in versus hey it's not uh you know it's not worth our time you know there's this famous um this rap

1:53:36

beef um yellow once called out Eminem or maybe it was MGK MGK called called Eminem and then Eminem was like oh man now I got to answer this fool I'm going to like build up his profile in responding to him and so sometimes is a little bit of that. Um, you know,

1:53:49

Um, you know, there's one example.

1:53:51

Um, Eric Newcomer, who like Ben Smith, unclear if that's a real name or not.

1:53:55

Um, if that's I don't want to dox him. Yeah, don't dox him. Don't dox him here. Don't dox. Yeah, we don't dox here.

1:54:03

uh you know, Eric Newcomer who who is a friend of mine. He's a good guy.

1:54:06

But on on his podcast, he did say um uh or or certainly imply that he was one rumor that our crypto fund has made a lot of its money betting on speculative things and then selling off and that couldn't be further from the truth.

1:54:20

Uh you know, uh we we are not a memecoin fund.

1:54:23

Uh you know, we hold mo most positions and so that's there was at least like one good company in that fund, right?

1:54:30

Like am I thinking of Coinbase maybe?

1:54:32

Like the biggest potential one. Yeah.

1:54:35

That Chris Dixon got into it.

1:54:37

Like the seed and the series A and the series B and the series Z like basically every single round. Yeah. That turned out okay. Industry.

1:54:45

That's that's an example of disinfo.

1:54:47

There's a lot of disinfo out there.

1:54:47

And so we we debate how much we should weigh in.

1:54:50

Of course, me coming in, I'm like, no, let's let's correct the record a little bit.

1:54:53

Let's uh so that that's one on the on the highest level how the firm operates.

1:54:57

operates. Um, in terms of where things are going, you know, I'm excited that you'll have both Martin and An uh today because in AI we have a big and David George, we have a big debate internally around uh defensibility and is it going to look like um you know previous

1:55:12

platforms where if it's not state that's the moat because state can be exported it's it's context context is the moat and that you you chat with chat uh with chat GPT and it gets to know you better and and that's what's going to keep you um or is it uh is it brand um is Are these things just moving so fast? You

1:55:28

You know, we hear all the time about people who are building cursor for X that uh that brand is what keeps people going even if if if if it's not context or or are they just not not as defensible as previous players?

1:55:39

But um you know the numbers don't lie and the the user growth and and the revenue growth is just astronomical.

1:55:44

And as aggressive as we have been, we we talk internally like we should be being more aggressive.

1:55:49

We should have been more aggressive.

1:55:51

There's winners in every category and it doesn't seem like they have dominant market share.

1:55:55

market share. that that's a mistake that we've made historically is like thinking oh you know OpenAI won this so you know there's not going to be another uh you know big big provider and on either the model layer or or or the apps and so it seems like the market is expanding where where OpenAI continues to to do

1:56:10

phenomenally well and also there's there's winners in across the stack and and room for more so um I'm sorry how has your mindset shifted being at a you know platform multi-stage fund now where you Historically, if you were investing, maybe you could do personal checks across different, you know, stages, but ultimately you were very early stage focused. Is it is it has it been

1:56:35

Is it is it has it been enjoyable being at at a, you know, such a scaled fund where you can realize like, oh, I didn't get into the series A of that company, but I want to meet the founder now, so and and really get to know them.

1:56:48

So, we have an opportunity to be in the B or the C or the D, right?

1:56:51

like there's there's really no no sort of um cap on on when you guys can get involved and partner with the company.

1:56:59

And that's a totally different mindset of of allow of being able to think like truly long term. Yeah.

1:57:03

Well, and it's a bit of both, right?

1:57:06

Because on the angel side or seed side, you can it's easier to invest in in people who are doing the the same thing or or sort of, you know, YC is in, you know, god knows how many competitors um you know, companies that absolutely hate each other.

1:57:18

I'm not going to mention names, but um again to get sense.

1:57:21

Um and uh you know bigger multi-stage firms just don't have that same luxury.

1:57:25

So you you got to pick and you got to be right.

1:57:27

Um uh and also to to your your point, yeah, if you don't hit something in the A, you know, you know, we weren't the earliest backer of Andre, but we're the second biggest backer now because we just caught it at the at the right time and just backed it all the way um afterwards.

1:57:41

So even if you don't get in at the seed or the A um and I spend a lot of time with the growth team, there's big opportunity to to get involved.

1:57:48

So that's uh that that's an encouraging sign.

1:57:50

And just once like I didn't know what I didn't know.

1:57:54

It's kind of like I I used to live in Detroit and then I moved to San Francisco and I was like why do I live in Detroit?

1:57:59

Why did I live in Detroit?

1:58:00

And similarly like um just being at a at a place like this I didn't I didn't quite realize what I was missing. That's great.

1:58:06

Well, let's kick it over to Jen.

1:58:07

I know we're running over time already.

1:58:08

I'm sure we're going to be bumping up against stuff all all day long, but thank you so much for hopping on and kicking it off.

1:58:13

Kick TVPN correspondent signing off. Thank you.

1:58:15

Always great to see you, Eric. Great to see you.

1:58:19

Uh yeah, keep these soundboard going.

1:58:22

Jen told us that she wants some soundboard. We'll hit it hard.

1:58:24

She's the operating partner at uh Andre Horowitz and is responsible for fundraising.

1:58:29

LP day is obviously a huge day for her.

1:58:31

So, we will bring her in in a second.

1:58:33

We'll ask her what's on the minds of the LPs today uh and whether or not America is back.

1:58:40

Welcome to the stream, Jen. How are you doing, Jen? G. Good to see you all.

1:58:44

There was a world where we didn't talk and I just hit the size column button over and over and over and over and over and it's just a 10-minute segment of just us. Yeah.

1:58:51

Why don't you kick us off with how much money you've raised for Andre so we can just get Jordy an easy layup here. All right, let's do it.

1:58:58

By the way, for the audience, I told Jordy ahead of time, my sole intention is to have all the sound effects pop off the entire time I'm talking here.

1:59:06

Let's get the National Ball sound effect. Yeah, there we go.

1:59:09

Welcome to the stream, Chad. Welcome to the stream. Let's go.

1:59:12

By the way, by the way, as Katie said a few uh weeks ago, first time caller, longtime listener, but before we get into the content, I have to say I have to do just a quick plug for your true early fans.

1:59:24

You know, I know you had to do the gift guide for the Super Bowl of Pronatalism, right?

1:59:30

You had to cover tariffs.

1:59:30

You had to cover drug pricing. There's a lot happening.

1:59:32

But before we get to the real topics, it's been a while since you've talked about them on property rankings. Yes, it has. It has.

1:59:38

We need to bring that back.

1:59:40

Why don't you come on with us?

1:59:41

Do why don't we do why don't you join us for the for the updated power ranking of the bond properties.

1:59:49

This is what really this is the news that moves Silicon Valley for sure. Yeah. Yeah.

1:59:53

We can get kind of like a goat debate going. Yeah.

1:59:55

Anyway, by the way, I've got so I've raised over 43 billion in capital over the course of my career.

2:00:03

Most of that was at congratulations.

2:00:07

Uh but I I spend all my time across all the funds.

2:00:09

across crypto summers, tech winters, everything in between.

2:00:11

Uh I'm the professional yapper to the capital behind the capital, so the LPs.

2:00:15

Uh so that's where I spend all my time.

2:00:17

Uh what's it like having everybody or not everybody, but but many people under the same roof.

2:00:23

I'm surprised you have even 10 minutes to get away today.

2:00:26

You probably one of the more popular people in the room.

2:00:30

Well, it's definitely the quiet before the storm right now.

2:00:32

So things are kicking off here shortly, but it's uh it's awesome.

2:00:35

Uh it is um there's a lot obviously happening in the world today and also uh our LPs are are busy people.

2:00:42

Last week was Milkin, this week was the Saudi conference.

2:00:44

Uh there's a lot on everyone's minds.

2:00:47

We could talk about what's happening with the endowments.

2:00:50

Um and you know people are are looking for some optimism.

2:00:53

We're probably the bright spot in their portfolio right now.

2:00:56

Uh and so as a part of that, it's exciting to gather everyone here.

2:01:00

Yeah, there's a lot of like hot take and micro themes around, oh, if there's a if if Harvard goes forprofit or something, that's going to flood a huge amount of endowment money out of venture capital or Yale's going to sell their stakes or something like that.

2:01:13

But can you talk about some of the more macro themes if we zoom out the last 10 years or more?

2:01:17

How have trends changed in the LP base of venture capital more broadly? Yeah.

2:01:25

Well, the asset class has definitely gotten obviously a lot more sophisticated, right?

2:01:27

It used to be a small cottage industry and the asset class has just grown alongside of course the outcomes and so as a result of that you know the sophistication level of not only the venture capitalist but also the expansion of LPs have commensurately changed as a part of that.

2:01:43

So you just don't have endowments or foundations or family offices investing that's expanded to sovereigns now who have a different type of opportunity cost of capital and then you have big pensions now actively uh playing the game as well.

2:01:57

So Kalpers, for example, very famously missed out on the last 1015 years of returns and now they're making up for for lost time as a part of that.

2:02:06

And so traditional pools that didn't used to invest into the asset class are now wanting a seat at the table largely again because increasingly the percentage of GD GDP that's being taken over by tech has just exploded.

2:02:17

I heard you mention to to Eric, you know, Nvidia is about to be uh this monster of a company just jumped up.

2:02:23

up. that about to be it's only a 60 it's only $60 billion gap between I checked this morning so get this so so Nvidia Apple and Microsoft independently are bigger than the entire stock market of most of the G7 countries

2:02:39

so Germany UK uh France like it is wild it's it's actually bigger companies are bigger than all the G7 countries except obviously for the US and Japan yeah I mean speaking of the international stuff a month ago people were buying French bombs It was a mistake. It's good to It's good to see. Um, yeah.

2:02:54

What is kind of the trade imbalance, the trade flow in terms of venture capital?

2:02:59

I imagine there's a lot of money flowing internationally into American uh venture capital firms, but VC funds also invest in uh invest abroad and and uh internationally.

2:03:07

Uh is there is there a trade imbalance there?

2:03:11

Is there more money flowing into America than flowing out or is it kind of 50/50?

2:03:17

Like how has that evolved?

2:03:17

Yeah, so our I will only speak for A16Z.

2:03:19

So, um, A16Z is still 90% plus in North America.

2:03:23

Um, that being said, you know, our portfolio companies though spend a lot of time abroad and focus on a lot of distribution as a part of that.

2:03:32

So, I heard your your Saudi correspondent yesterday mention that folks in the uh the defense in Saudi, you know, all of a sudden are like, we want Ander, we want Seronic, we want all these early stage technologies and and guess what?

2:03:43

Every country wants to modernize their government.

2:03:47

They want Silicon Valley tech on speed dial.

2:03:48

And so there's a massive interest in all areas, especially in our American Dynamism practice for new technologies and they want to replplatform to the new AI age.

2:03:58

And so we actually um at A16Z have brought over almost close to 300 plus companies uh around the world um spending a lot of time in in Saudi and also the broader GCC and particularly um yeah, it's funny.

2:04:12

I was watching the uh the Qatari uh motorcade welcome Trump earlier today.

2:04:17

like they went so hard on on it, but it's it's that kind of classic thing.

2:04:20

We're there to go to the best technology. Yeah.

2:04:22

First and for foremost and so as a part of that that is the leading driver of wherever uh trade actually ultimately goes to.

2:04:29

What is the takeaway from Saudi Arabia and what's going on with KSA and NBS?

2:04:34

Um how are you thinking about Saudi Arabia more broadly these days? Yeah.

2:04:40

So, we we spent a lot of time there um largely again to help our portfolio companies with with the expansion and and you know, NBS is pretty phenomenal.

2:04:48

Everyone's talking yesterday a lot about Trump and and how the vibe has shifted, right?

2:04:52

The America's so back, right?

2:04:53

And uh I feel like that deserved a gong like it is I was caught sleeping.

2:05:00

I was sleeping at the wheel.

2:05:02

I'm going to be on top of it. I got I got more here.

2:05:04

I'll I'll stay I'll I'll keep the soundboard up and just keep going. good.

2:05:08

Uh but but NBS is like an amazing leader, right?

2:05:13

Everyone's talking about Trump, but but I think the the the other part of this is like in the eight years since Trump's first visit, NBS was basically able to get every tech leader in the world at that conference, right? He's absolutely crushed.

2:05:26

And and the whole country, by the way, his people love him.

2:05:28

They all have, if you go to to the country and you see people's cell phones, they have NBS's photo as their their background on their phone.

2:05:37

And then if you WhatsApp with people, their profile picture is MBS, which makes it obviously very confusing.

2:05:41

But you know, when was the last time a world leader was as beloved as he has been?

2:05:46

And and that's largely again because he's very conscious of the fact that the future needs to be tech first.

2:05:53

He's got 70% of the population under 35.

2:05:55

And that population is watching Netflix and YouTube.

2:05:58

And so he's very conscious of the fact that he needs to modernize his country as a part of that.

2:06:04

He's retooling all the capital away from oil into tech as a part of that. Very cool.

2:06:08

Well, last question then we got to move on. Yeah.

2:06:10

What what what are people getting uh you are one of the co-creators of a mega fund.

2:06:15

What do people get wrong about the the in in the sort of megapun debate?

2:06:21

Because oftent times the people that are debating have sub nine figures of aum and uh you know I just think it's like unless you're playing unless you're in the arena scare me the big ones scare them but um but yeah what what what are people getting wrong about the the megapund debate today?

2:06:39

Yeah, it's it's a totally different game and obviously again I'll only speak for ourselves.

2:06:43

uh yeah what we're playing is is very different than you know obviously seed funds and small emerging managers are playing.

2:06:51

So it's very very uh difficult to paint the broad paintbrush across um every type of venture capital firm but you know I think the the premise of it still holds true.

2:07:01

So so I heard Eric mention this earlier you know Mark pioneered the term software is eating the world and the reality is software didn't just eat it door dashed it. It devoured it. ask for seconds, right?

2:07:10

And so everyone's obviously uh now playing catch-up as a part of it relative to this massive opportunity, but also more importantly, firms haven't scaled.

2:07:20

Most VC firms haven't scaled to reflect that.

2:07:24

And so from a capital standpoint, not only have we scaled the capital, but we've also scaled the firm as a part of it.

2:07:30

So if you look at every single team at A16Z, it's actually no bigger than the original incarnation of A16Z.

2:07:35

So the deal teams are actually pretty small.

2:07:40

They're no bigger than four to six people.

2:07:41

So, every vertical that you'll talk to today has a deal team focus that is much more oriented around not only that vertical, but also making sure that the fidelity of conversation that you have is super high because otherwise as you've scaled the firm, if you have too many people around a table, it ends up being like a presentation, right?

2:07:57

And so, totally when we think about the firm, that's the big part that we focus on.

2:08:03

like have we actually not only made sure we've scaled the capital to reflect the market opportunity, but also have we also scaled the firm to reflect how we actually want to run the organization and and making sure we're actually covering each vertical with that domain level expertise. That's amazing. Thank you so much.

2:08:20

I I mean I my my contribution to the mega fund debate is if you ask a series C or D founder, hey, are you happy that there's, you know, potentially hundreds of millions or billions of dollars available to you to raise in the private markets, are they going to be like, no, I'd rather do a small cap IPO and and, you know, be, you know, spend years dragging it out in the in the public markets before my in every statement.

2:08:44

I think that that should end the debate by itself.

2:08:46

So, um, Jen, it's fantastic having you on.

2:08:48

Uh, the crowd loved it, too. The crowd got us. Come back. Come back on again soon. This was great.

2:08:56

We'll do the We'll do the Aman power ranking together. Can't wait. Catch you later. Bye. Have a great event.

2:09:02

Yeah, it's uh it's funny.

2:09:02

Uh underrated narrative is that they are this mega fund, but I was looking at who invested in Josh Browder's new fund, $30 million.

2:09:11

It's like four Andre partners, right?

2:09:13

It's like Mar Chris like all the guys.

2:09:16

And he raised money from Andre for do not pay as well.

2:09:18

So there's been this like really such Josh's A16Z round I think it was a series A wasn't it like 4% dilution round.

2:09:26

Yeah it was really low fascinating but they clearly like working with him because they backed him on the fund. Yeah.

2:09:30

And he's taken such a contrarian approach to building the company.

2:09:34

Seven full-time employees. Yep. Dividends.

2:09:37

He was like the first company to be like I'm profitable. You remember this? He's a master of media. He's always on TV. Like how'd you do that?

2:09:45

Anyway, uh our next guest is runs in the family.

2:09:47

Uh welcome to the stream. How are you doing? Great. Looking fantastic. What's happening? What's happening?

2:09:51

Um uh my name is pronounced Anjene. Friends all call me An.

2:09:55

You should feel free to as well. An fantastic. Uh great to meet you.

2:10:01

Um I'd love to kick it off with a little bit of review of what you're investing in today.

2:10:04

Uh what you're excited to presenting at LP day and then a bunch of questions on open source AI and the specific trends that are going on in the in the in the market right now. Sure.

2:10:13

I have the fun job of spending the $40 billion that Jen raises for us every couple years. Fantastic.

2:10:20

So I I'm a general partner on our AI infrastructure practice, which is basically all the computing systems that people need and rely on to build great products and services on top.

2:10:29

And I basically spend most of my time um as a glorified customer support um and compute intern for scientists building foundation model companies, training models, all the stuff that happens, you know, at the intersection of of of research and then actually trying to figure out how to get that research out into the real world. Okay.

2:10:49

So I I started at the firm two years ago. Oh, cool. Almost to the day.

2:10:54

Um, and I came to the firm after a couple years as a founder.

2:11:00

Um, and then I I I founded a company called Ubiquity 6, which we then sold to Discord.

2:11:04

And um, I ran the platform there for a few years during the era where we went from basically being a chat app for gamers to which was about a million monthly active users to about 200 million monthly active basically in 6 months because COVID happened. Yeah.

2:11:21

And um the the homework assignment was an go build a developer platform business for us.

2:11:29

And it was around that time that I got a call from some friends who were running research at OpenAI who said um hey an we're we're we've just trained this model called GPD3 and we think we're on to this thing called scaling laws where if we can combine compute data and algorithms in the right combination, we just might have a shot at creating God. Classic. Um, okay.

2:11:51

How much do you need to get started?

2:11:53

They're like, sorry, our our audience, the live audience here got got a little excited.

2:11:59

A little excited about God in the box.

2:12:02

Um, well, you know, as a having just gone through the journey as a founder, uh, I was like, okay, how much you think that's going to cost us?

2:12:09

I mean, go God should be pretty pretty cheap in these times where where religion seems to have left the room.

2:12:13

Um, and you're like, yeah, it we think we can get by with five.

2:12:16

And I said, okay, five million shouldn't be a problem.

2:12:18

we can probably scrap that together next week.

2:12:19

And uh this was Dario who is a longtime friend and now founder uh and CEO of Anthropic said, "No, Andre, I don't think you realize I meant 500 million."

2:12:28

So that's going to take a little bit longer.

2:12:31

So So I I was one of the early angels into Enthropic at the time.

2:12:34

That gave me sort of my crash course to scaling laws for neural language models, right?

2:12:39

The idea being that you could get these these language models to to to predict the next word in a sequence and if you kept scaling up um both the compute the amount of compute you train these things on as well as the data set then you'd have no you'd be able to predictably improve the performance of these models.

2:12:58

I'd call that the pre-training era.

2:12:58

I'd call that the pre-training era. what I spend a lot of my time and so you know for the better part of the last three four years I've basically helped teams like anthropic I usually get a call from them before there's a company when they're scientists who who've who who get who get have this sort of moment

2:13:11

where they realize they're they're they're about they've unlocked something um repeatable where where you can very predictably improve the performance of these models on some axes um and they usually give me a call and they say we've we've had the the enlightenment moment we have no idea how to commercialize that and turn that into a business. Um, and so I've worked

2:13:32

Um, and so I've worked with teams like Mistral which is working um on open source models.

2:13:37

Uh, last year I led a our the first round into a company called Black Force Labs which is the who are the creators of stable diffusion which is the open source image model family that I think sort of introduced the world to the idea that open source language and image models can be pretty powerful.

2:13:53

And so that's that's what I spend most of my days doing helping folks you know get their clusters set up to train models to then build products on top.

2:14:00

And I think at this LP day the there's sort of two big things I'm getting asked about.

2:14:06

The first is okay an you know two years ago it was all about pre-trained models like the pre-training era AI scaling laws. What's going on now?

2:14:16

Why is everybody talking about post training right?

2:14:17

Um so that's that's one big question.

2:14:20

that's that's one big question. Um and then the second is hey what what is going on with with sovereigns like why are countries announcing a hundred billion dollars into data centers and why are they all talking about wanting their own AI you know champion company whether that's mistrol in Europe or um

2:14:38

whether it's Alam in in Saudi and so I would say sovereign AI which is really the idea that you want to control your own AI you want to control what it can and can't do um and the idea that that progress in AI is not coming necess necessarily just from pre-training, it's coming from from post training are the two big themes for for this year for me. Yeah. Um, on the on the open-source uh Yeah.

2:14:57

Um, on the on the open-source uh question, are you are you thinking about Mistral and Black Forest Labs as kind of like the Red Hat Linux model where they will be almost like implementers in the enterprise or is there a real world where you're underwriting against total and complete victory of the open- source uh of the open source paradigm at the foundation model layer?

2:15:20

And then related to the the international topic you mentioned like is Mistral uniquely um are you uniquely bullish on Mistral because it's an international company as opposed to an American company because it has this this specific advantage of being in Europe and and that driving extra value for the company.

2:15:39

Oh, I no I I think so, of course, you know, Europe is 400 million consumers.

2:15:45

So, just massive economy that's that's decided to gear up um to build the basically the single largest infrastructure build out I think we've seen in the continent in like 60 years.

2:15:56

They just passed an $800 billion defense bill they're calling Rearm Europe and like a huge portion of that is flowing to uh AI and computing teams for sure.

2:16:04

So, we can talk about that.

2:16:05

But that wasn't when I led the the series A into Mistral two years ago, $200 million round.

2:16:11

Um like Europe was we I did not expect that we'd be in the middle of like this massive sovereign AI buildout.

2:16:20

I mean the bet was very simple is that if you look at the history of computing infrastructure, there's basically two frontiers.

2:16:25

You've got the capabilities frontier and then you've got the efficiency frontier, right?

2:16:29

And the capabilities frontier is usually dominated by closed source.

2:16:32

If you look at, you know, databases and storage and networking and so on, you usually have a company like Cisco and so on that that pioneers some new capability, right?

2:16:39

Um and then or or actually in the case of Linux, like you said, right?

2:16:44

Um you had Microsoft show up and build Windows closed source, right?

2:16:47

And that usually opens up the aperture for consumers because consumers are often the first to flock to new use cases.

2:16:55

The enterprise cares about something slightly different.

2:16:58

They care about cheaper, faster, more control. Mhm.

2:17:00

And that's usually dominated by open source.

2:17:02

So, you know, two years ago when I was running the the platform or at Discord, we got early access to GPT4 and the OpenAI guys said, "Hey, we've got this new model.

2:17:13

It's going to come out in 6 months.

2:17:15

Can you guys just figure out what it's good for?"

2:17:16

And we ran an experiment.

2:17:18

You know, we we we ran a bunch of uh tasks through it and it was extraordinary on a couple of them and we said, "Okay, great. Let's go to production."

2:17:25

Well, it turns out when you want when you're working with sensitive data, right?

2:17:29

In our case, this was a social platform.

2:17:31

So GDPR, CCP, all of the compliance stuff, really critical that our data doesn't leave our servers, right?

2:17:37

You want control over what where the weights are running and what the weights can and can't do.

2:17:41

And so we looked around for a for an open source alternative to to GPT4 and we just couldn't find one. Mhm.

2:17:48

And that's when I realized, okay, for every dollar that that you're seeing in like enterprise or large company prototyping or proof of concept revenue, there's like 10 more waiting for the open- source alternative, right?

2:18:00

And really, there was no alternative until later that summer when Llama showed up. Mhm. Right.

2:18:06

Llama was the first time there was a comparable open-source alternative for to to a closed source models and and well, the creators of Llama left and started Mistral, so that that made the investment easy. That's great.

2:18:16

Uh can you take me through the deal for uh Sesame AI?

2:18:21

Obviously they went viral but the founder had had a previous Andre backed company I think in Oculus.

2:18:27

Um how did that come together and uh how is the company? Yeah.

2:18:32

And I and even your first interaction with the product that that blew you away because it in in many ways that that that their initial launch and and the website and and I think you guys had already you know completed the investment well before that but so many people that moment was truly eye opening for them where it was like okay I could imagine this feels like talking to you know a friend you know it felt it felt really real. Yeah. Right.

2:19:00

Uh yeah, Sesame is a fun story.

2:19:02

Um because we started that company two years ago at a time when like everybody in this space was like look the the this this idea that that we're going to have a new computing interface um was completely seen as like a crazy a crazy idea.

2:19:23

There's there's two parts of of this ask me.

2:19:26

one is the the hardware and we're building AI glasses over there.

2:19:32

over there. M um and the second part is the actual companion which is the voice interface you guys have probably tried out the the voice AI right and when you put those together the idea is that it's a it's a system that has so much context about everything you're doing about your life that it becomes the primary

2:19:49

interface to to to computing and I mean you know two years ago it it it I would say a lot of people had seen and watched like her the movie right yeah But the the the when I'd sit down and I'd describe to people that hey what you really need is is this beautiful marriage of hardware and software. Um

2:20:12

Um and that's what's going to come that's what's going to be the primary interface after smartphones.

2:20:19

People would just look at you like you're crazy.

2:20:21

Um and so I was like okay the the there's only two people I know who are crazy enough to believe this.

2:20:26

One was Brendan who had been an angel investor in my last company Ubiquity S.

2:20:29

Brendan was the CEO of Oculus. Yeah.

2:20:32

You know, one of the few people on earth who who' both built a hardware startup and sold it for billions of dollars.

2:20:37

I think Oculus sold in total for north of two billion.

2:20:38

And Ankit who was co-founder and CTO of Ubiquiti 6, which was my last company.

2:20:44

And Ankit was running the voice uh part part of the voice AI uh SDK and infra at Discord.

2:20:50

You know, about 60% of all discord voice minutes uh daily minutes are spent in voice channels. Wow. Wow.

2:20:59

That's when we realized people don't realize it, but for most for many many many consumers, voice is actually a much more frequent interaction modality than like looking at a screen. Yeah.

2:21:08

And so the idea was if we can combine those two, we might have a shot at building whatever comes next after smartphones.

2:21:16

So that that that was the idea, but the the key like insight I guess or bet was that it has to feel realistic.

2:21:23

You're not like Siri just doesn't work.

2:21:25

like you just nope out completely when you try to talk to Siri and you're like, "Okay, I'm talking to a robot." Yeah.

2:21:29

But instead, if we could get the companion the the to to cross the uncanny valley, if we could get you to think about talking to Sesame, and you know, we have two companions right now.

2:21:41

One is Maya, the other is Miles.

2:21:41

If you could get get you to think about using and talking to Maya and Miles as companions, not as robots, then that's when you'd really start using it in your daily life as an interface constantly to all the apps and services use. Does that make sense? Yeah, totally.

2:21:58

make sense? Yeah, totally. How do you in the context of Sesame what's been the thinking because obviously sounds like you've been on the board since day one or or you know I don't know the exact mechanics but how how has the company thought and and how do you think broadly across the portfolio around this a

2:22:15

balance balance between you know being heads down building and needing to capture the attention of of uh you know the tech community the broader you know potential user base because I feel like it's this interesting dynamic right now where there's so much to build like in some ways companies should just be heads down and almost be silent, right? And

2:22:32

And and this is like the SSI approach is to basically say like we're not you know maybe going to release anything that that we we consider less than um you know a really significant evolution.

2:22:44

Um and but Sesame kind of like popped its head up and said hey look at us look what we're doing.

2:22:52

But then now it seems like, you know, it's taking more of the approach of, you know, being willing to fly under the radar for, you know, call it the next who who knows how many months.

2:23:02

Um I'm I'm happy to report that it it doesn't get any easier the more money you raise.

2:23:10

I found like that on day one um we we didn't we hadn't raised any money, right?

2:23:14

So we we're just three guys sitting in a room and talking about ideas.

2:23:17

And um this I I I would say the heristic for most people who come from the software era is like hey ship fast, ship early, ship something you're embarrassed about, right?

2:23:28

And then iterate and AI is a little bit different.

2:23:30

Frontier is a little bit different because you often need a critical you you you need to have a a capability threshold that is sufficiently transformative enough that people would will put up with all the tons and tons of like friction that there still is to use AI today, right?

2:23:51

So you guys may have noticed but when you go to the Sesame site and interact with the demo, right? It's quite fast today. it.

2:23:57

One of the things we really the team like spent a ton of time honing was the latency of voice responses.

2:24:02

We had to make it feel like it was sub 200 millconds.

2:24:06

So it felt like you're having a conversation, right?

2:24:08

Two years ago it wasn't there.

2:24:11

It was is excruciatingly slow and we knew that nobody it people would miss the underlying personality of the model if we hadn't solved the systems problem of latency.

2:24:22

systems problem of latency. And I I find it comes down to this really delicate bal the the the overarching problem of like hey when do we ship when do we actually put this out in the world versus um being heads down comes down to this constant tussle between taste which

2:24:36

is who at the company has such a strong opinion that the product experience is good enough and the and the and the ruthless sort of machine learning approach of running evals right the idea that you you build a model you test it in in an environment you benchmark it and you see how it does on that eval. And if it hasn't improved that

2:24:55

And if it hasn't improved that evaluation score, you know, you keep going. Mhm.

2:24:59

And it it it it basically you you got to have like I've lost count of the number of hours we've spent debating that that tension, right?

2:25:06

And it's extremely uncomfortable for traditional software teams to do the the eval driven approach because you the the list of features you can kind of deterministically write out and say check check these are the things we need to ship build.

2:25:19

Once you've got the vo, the MVP, we ship. Mhm.

2:25:21

That's not how that's not how AI works, right?

2:25:25

AI research and post- training is is eval is eval driven, which is you have an intuition for what you what you need to do to improve the model, but you don't actually know until you run it through the an empirical test.

2:25:37

And so the answer comes down to how good are your evals. Mhm.

2:25:43

And so if I was to boil down what's what works for the best teams, it's that they have great product taste, but they also have great taste in eval what what is the right evaluation metric to build um for your team and then you you basically stay heads down until until you've you've unlocked both of those. Does that make sense? Yeah. Yeah. Fantastic.

2:26:00

Well, thank you so much for joining.

2:26:01

Uh good luck with the rest of LP. Come back on again soon. Yeah. Yeah.

2:26:06

We'd love to have you back and talk. Anytime.

2:26:07

You know where to find me. Fantastic. We'll talk soon. Bye. Cheers. Cheers.

2:26:11

Uh, next up we have Martin Casado, general partner.

2:26:13

He's on the board of Cursor, World Labs, DBT Labs, Kong, Fiverr, Idoggram, Ambient AI.

2:26:17

This guy has a lot of board seats.

2:26:20

Brain Trust, Coactive, Netlifi.

2:26:22

Um, uh, he leads the firm's $ 1.

2:26:26

25 billion infrastructure practice and has built quite the AI portfolio.

2:26:30

So, we're excited to have Martin on the show.

2:26:33

Uh, discuss open source AI, his views there, what's cursor doing.

2:26:39

Uh, welcome to the stream, Martin.

2:26:41

So good to have you here. How you doing? Good. Thank you.

2:26:47

You got to welcome some live the live audience.

2:26:50

The live audience is really just they couldn't control it. They're out of control. Yeah. Yeah. Anyway, thanks so much. Yeah. Yeah.

2:26:54

Thank you so much for uh for joining us.

2:26:56

Um I'd love to know your take.

2:26:59

We Let's just jump right into it since we're running a little bit behind.

2:27:02

Um cursor uh obviously it's becoming this front-end almost aggregation play.

2:27:06

It's the front end to AI coding.

2:27:09

It's uh potentially model agnostic. Uh what does that mean?

2:27:16

What have you learned from the cursor journey around how value occurs across application layer versus foundation model layer? Yeah.

2:27:23

So um I mean I think that there's two things you know that that led up to this cursor moment.

2:27:28

And one of them is one of the first use cases for AI was code, right?

2:27:33

we had like Microsoft Copil and a lot of people used it.

2:27:37

Um but like and so like you know there there was you know a bunch of users and their behavior was trained by Microsoft but the models weren't quite good yet and um in the last you know year we had this RL movement. Mhm.

2:27:51

And as a result, the models got way way better.

2:27:57

And it actually just happened to have like this kind of magic timing where it, you know, it did something very similar to what Cold Pilot looked like, but you know, use these greater models.

2:28:05

And now that they've kind of caught the wave, they're able to, you know, not only serve these amazing models, but like for every magic experience you see the the the users have, they've starting to like have their own smaller models.

2:28:17

And so it's just you know very unique point in time very unique uh position and then of course a very unique technology way behind it.

2:28:26

Going back to the foundation model layer uh can you talk about how your thinking around opensource in the foundation model wars has evolved.

2:28:33

Uh I remember you and John Ludig were talking about it last year.

2:28:39

Since then, we've had Llama 4, which kind of fell backwards, but then DeepSeek was a huge moment.

2:28:45

And uh we just had a a guest on the show yesterday who was talking about the importance of offering the world uh open- source foundation models that they can build on because if we don't do it, our near-per adversaries might.

2:28:58

How is your thinking evolved if at all?

2:29:01

And where are you seeing the most opportunity in the foundation model wars? Yeah.

2:29:08

So it's not clear to me that the model space is significantly different than you know the software space when it comes to open source.

2:29:14

So in the software space we've got you know 20 years of history of this.

2:29:17

Um you know the closed solutions tend to aggregate the value first and the quickest and the reason is is you kind of need to make money to change people's behavior right and you know this is like you know you need to have Unix before you have Linux and you need to have Oracle before you have my SQL etc.

2:29:35

SQL etc. And the same thing seems to me to be playing out in in in this AI space which is you know of course open AI was first and they've aggregated a ton of value and you know anthropic for code and um and then you know even if you look at the non- language models like you know image you know midjourney was

2:29:52

first and that's closed and 11 labs and so it seems to me like the first movers will close source and they'll continue to close source and they'll aggregate a lot of value but then for every first mover there's always the opportunity to enable the masses I mean like everybody else. Uh and there's always a lot of

2:30:06

Uh and there's always a lot of value there, right?

2:30:08

And this is kind of where these open source models come in.

2:30:11

And what's amazing and everybody should realize is just because Deepseek is out there hasn't changed the calculus or the growth of these other companies, right?

2:30:19

Is and fully are creative.

2:30:19

And so I think the only sin anybody can make in AI from an investor standpoint is zero someum thinking because the reality is is every time something new gets introduced, it's not at the expense of something else.

2:30:31

it kinds of finds new territory to gain value in and and open source is no different.

2:30:35

So it's a very important part of an ecosystem.

2:30:37

It's been part of the software ecosystem for a very long time.

2:30:40

It'll continue to be I do think that you know us um should have open source models and and I think it's very important for like continue to go presence in technology in that conversation.

2:30:49

Um but it feels very healthy to me um both the dynamic with existing closed companies and then you know new companies that can be built on on the open source. Yeah.

2:30:59

Are you a fan of aggregation theory or this idea of like the front door to artificial intelligence?

2:31:05

Cursor is kind of that for coding and we're seeing that in other applications that are popping up where uh the underlying model might not even matter.

2:31:14

And my where I'm going with this is that there's one world where uh another country wants their own foundation model.

2:31:19

So they know that the weights are trained in a way that they want aligned with their views on speech for example.

2:31:25

Uh but there's another world where the individual countries should be thinking about how do we create the dominant consumer app that people actually use because even if we have the model if nobody's using it because they're just opening up a certain URL and they go to another country's model.

2:31:39

What do you think of there?

2:31:41

I mean I think I think the story of AI has been the markets are much larger and they're growing much faster than anybody had expected and that's just resulted in actual fragmentation.

2:31:53

So, everybody had these theories early on, if you'll remember.

2:31:55

Oh, there'll only be one model.

2:31:56

Oh, the models are not defensible.

2:31:59

Oh, the apps are just GPT rappers.

2:32:00

Oh, like the apps are going to take all the value. And you know what? They're all wrong.

2:32:04

Like, everybody's been wrong.

2:32:05

Like, the only have been right are like, you know what?

2:32:08

These things are massive.

2:32:08

I mean, let's just take OpenAI, which is one of the most remarkable companies and maybe forever in tech, right?

2:32:13

But, you know, they were the first to image.

2:32:15

Remember that with Dolly? They lost. Yeah. Yeah.

2:32:19

They were the first to code with Copilot. They were the weights.

2:32:21

Copilot. They were the weights. they lost that they were the first of video and like you know Sora was amazing and remarkable but like they're not the leaders in video um and yet they're still the dominant AI company on language right and so what are the takeaways you can take from that like in my view it's you know every one of these is a massive massive market that's too big for one company and so like you know

2:32:42

open a great job focusing on chat GPT and you could have the same discussion between the model and the app layer right and so I know I mean you're talking about aggregation theory so there's two views of the world one view of the world is like everything consolidates into one company and there's another view which is there's like a model per thing that's a behavior I'm definitely in the second one and I just you know even though we're

2:33:01

relatively early we do have four years of experience with this and the story has just been disagregation it really has yeah that makes a lot of sense how are you uh how do you think about revenue quality in the enterprise in an

2:33:13

environment when you know the fortune 500 sort of has a mandate to just you know buy AI some some things they're going to that's a great yeah so okay so so there's there's a lot to this question and I'll try to be as succinct

2:33:28

as possible so the first most important thing is um AI today is mostly a proumer and consumer movement so even though the people are working in the enterprise it's like it's an individual use and individual behavior and it's not like a

2:33:41

big budget buying thing and so the I would say the vast majority of the companies we work with it's this more kind of bottoms up proumer thing and so that's clearly real individual demand This happens in all super cycles. Like

2:33:51

Like it was the same thing for the internet, right?

2:33:54

It was the same thing for PCs.

2:33:54

It was the same thing for like the iPhone, like the true smartphone like post Blackberry. It was like individuals.

2:33:59

And so, you know, from that standpoint, like I think we're very very comfortable.

2:34:05

This a real secular movement because it is individuals that are driving the behavior.

2:34:09

Um, you know, there there there's there's a second point to all of this.

2:34:13

this. um which is um in the early internet days you had all sorts of funny accounting cuz you know and it was quite frankly cuz people just didn't know how to do it successful but at the time people weren't even paying right so they would just kind of say everything and and of

2:34:29

course in these early super cycles you have all of that like they don't know the difference between like GMV and take rate and you know run rate is not a gap thing and so they kind of had deal there's clearly all these funny numbers out there um and you know that's just an artifact of massive massive growth and

2:34:44

success and like you know we'll rationalize it as an industry but for sure that's going on but it's less about like enterprise budget moving and it's much more about I've got this consumer phenomenon I have no idea how to really do the accounting you know and so you know it just requires from our standpoint you know just to kind of go

2:35:01

through things and kind of really normalize it that's great uh I mean we could talk for another hour about this we'd love to have you back on the show this is fantastic uh but thank you so much for even taking the time that you did more time this was great Love it. Thanks so much. It was a real pleasure. Thanks so much. It was a real pleasure. Cheers. Have fun. Cheers. Bye.

2:35:15

Um, yeah, we didn't even get to Feay Lee and World Labs, the deal that he led there.

2:35:19

Um, he spatial intelligence Martin, you know, I've seen his post, never heard him talk.

2:35:27

Potentially a Hall of Fame Hall of Fame yapper.

2:35:29

Can we get a soundboard for that?

2:35:31

Hall of Fame yapper for Okay, we got David George leaves the Growth Fund coming in next.

2:35:37

I'm going to want the Ashton Hall sound effect.

2:35:39

I'm going to want lots of sound effects for this one.

2:35:43

We're getting too serious right now.

2:35:43

We got to take it down a notch.

2:35:44

Everyone thinks, "Oh, they managed 40 billion." You got to be serious.

2:35:48

You got to be on Talking Points.

2:35:50

We have a fully executed lease on our new studio.

2:35:52

Oh, let's hear it for us.

2:35:54

Let's hear it for TVPN in the chat. Oh, thank you.

2:35:57

And you know why that's exciting, John?

2:35:59

Because we're going to be walk.

2:36:01

We're going to be able to walk around sitting for 4 hours.

2:36:03

Sitting for 4 hours is brutal.

2:36:06

Brian Johnson is shaking his fist. Yes. Thank goodness.

2:36:11

sitting is is, you know, don't they compare it to smoking cigarettes in terms of, you know, the terrible body. It's terrible for you.

2:36:17

And uh we're going to be walking around hitting the biggest gong you've ever seen that we've yet to reveal on the show.

2:36:22

Uh we'll be whacking that.

2:36:24

We're going to be you're going to be height mogging me, but I had an idea.

2:36:27

We can put a uh like a measuring tape up on the wall so that I can periodically walk over and prove that I'm Yeah. Yeah.

2:36:33

Anyway, uh we got David George in the studio.

2:36:37

Welcome to the stream, David. How you doing?

2:36:40

Welcome to the How you doing?

2:36:40

We're trying to keep the energy up. It's a big day. I love it. Great to see you.

2:36:45

Thanks for coming on the show.

2:36:48

Yeah, thanks for coming on. I love it.

2:36:49

Uh I would love for you to break down how the growth fund actually works.

2:36:53

Andre's known for these specialized funds, but then I know that the growth fund cuts across all of them and you see deals all over the place.

2:37:01

Uh do I have that even roughly correct?

2:37:03

How do you think about um supporting the companies in the portfolio that maybe get picked up at the earlier stage? Yeah, absolutely.

2:37:11

Well, look, fun to be on here with you guys. Uh big big fan.

2:37:13

Um yeah, I mean you have it right.

2:37:17

So we have seven different funds.

2:37:18

Uh I think if you talk to some of our early stage folks, they're very domain focused and that means you know they know the products, they know the markets, they know the technologies at a very deep level.

2:37:28

Um and then the growth fund where we sit is kind of across all of those.

2:37:32

So um you know ideally the way it works is um you know we can work handinhand with them and take kind of the best of the best ideas from each of those sectors uh and then invest you know stage specific uh at the growth fund and so what we do probably half our effort and dollars historically have been uh followons supporting existing portfolio companies and um that's worked great.

2:37:54

uh you know there's something we talk about all the time which is game film and like there's no better way to have game film than to uh you know have my partner having done the series A and Data Bricks and then you know we get to see it all along.

2:38:07

Um you know with that comes access um you know relationships with the founders all that stuff helps.

2:38:14

Um but then you know the other half of what we do uh is invest in net new companies that were not you know previously investments in A6Z for whatever reason.

2:38:22

And so, uh, you know, we started started this in 2019.

2:38:28

Um, I had a big list when I joined.

2:38:28

I'm like, "All right, I know the six existing portfolio companies that I want to go try and create a, you know, a new deal in."

2:38:36

Um, and then I knew a bunch of the ones that I wanted to invest in outside the portfolio.

2:38:40

And and and we've been fortunate to have some success.

2:38:42

You know, we've invested in uh outside companies like Roblox and Figma and SpaceX and Whimo.

2:38:48

um you know and then you know obviously a lot of success uh doubling and tripling and quadrupling down uh with existing companies.

2:38:55

Well, it's a great time to tell everyone that TBPN is sponsored by Figma.

2:38:58

So, thank you for the shout out.

2:39:00

Uh this is one the other moment.

2:39:02

Put some dollars in your pocket.

2:39:05

Put some dollars in our pocket.

2:39:06

Yeah, that was what didn't you do that deal in like April of 2020 or got announced back then?

2:39:10

That was a crazy moment in history. Congratulations.

2:39:14

I got a lot of stories about that one, man.

2:39:15

But we um so we had been courting.

2:39:18

I remember I met Peter Lavine uh when I was like right at the tail end of figuring out whether I was going to join and and uh he was like we need a growth fund.

2:39:27

I'm like okay like I was going to tell you that.

2:39:29

Um and he's like you know like we got to invest in this company Figma.

2:39:34

Like it's crazy that we got investors in this thing.

2:39:36

Like it really pisses me off.

2:39:37

Uh all right well we can go fix that.

2:39:40

And so um you know I joined and we courted Dylan for like you know a year year and a half um and then you know co hits we all think the world is going to melt.

2:39:49

We've been trying to convince Dylan to to do an investment with us.

2:39:53

Um and like the world's falling apart and he calls us and he's like I'll do a deal now and I'm like oh god.

2:39:57

So part of the dealend absolute dog this is this is brutal. Put you on the spot.

2:40:06

I know it was uh but it was great.

2:40:08

Obviously, we're thrilled to be as partners.

2:40:10

Uh that one was funny cuz we, you know, like if you take a traditional growth investing lens and look at a company like Figma at that stage, um you know, you could look at a spreadsheet and look at market reports on market size and you know what what you'd see is like, well, there's not that many designers and sure the number of designers is growing, but how compelling is that going to be?

2:40:30

Um and you know if we didn't have like depth of product and technology and market insights we would miss like the obvious you know thing which is like it's not just a design tool.

2:40:43

It's a it's a tool for um you know designers plus front end engineers and that functionality is emerging together.

2:40:48

Um and the portion of designers is massively growing relative to uh relative to engineers and you know there's even going to be business uses and so um you know Figma is an awesome one.

2:40:59

It's a it's a great example of a company where um you know traditional lens would probably miss it for what it could be.

2:41:05

Um but you know it's sort of a it's one of these things we call you know we call model busters.

2:41:09

It's like you do a little spreadsheet model and it like show these numbers and it like it turns out like it's all wrong like the market size is totally wrong and it can just totally exceed it. Oh, that's good.

2:41:23

Uh, can you talk about when it breaks the model busters, the aperture of growth equity deals?

2:41:27

Um, obviously the traditional growth round is postproduct market fit.

2:41:33

Uh, maybe, you know, you're up in the 50 million plus raised range.

2:41:39

Uh, but we're seeing these crazy deals get done with pre-revenue, pre-product AI companies raising hundreds of millions of dollars.

2:41:46

Are you putting that in the growth bucket?

2:41:48

And are you changing the way your team thinks about deals to look at to set yourself up for looking at those bigger but earlier stage deals or are you trying to think about how can we fit those into a venture portfolio or venture fund?

2:42:01

Yeah, let me just talk to you about like the buckets of stuff we do.

2:42:06

Um, and so you know there's a big bucket of stuff that we do which is you know traditional high-flying growth equity companies.

2:42:14

high-flying growth equity companies. And so a bunch of the companies that I listed uh would fit that um you know data bricks uh you know uh Roblox and those are like a mature businesses like they have you can do a DCF on them you can do market multiples like you can do

2:42:29

real real work but yeah yeah so that's one bucket and I I'll talk more about those um but I would say my observation um from you know doing growth investing for the last 15 years or so is um you know there's a bunch of instances of those that like end up being way bigger and better businesses than anybody would have thought. Yeah. Um and that Yeah.

2:42:48

Um and that shouldn't be surprising cuz if you just look at like the composition of the biggest and most important companies in the world, they're all tech companies and you know they you know their their their market caps have sort of extended beyond uh anyone's you know idea of what they could be both because they exceeded market size estimates and also they came up with new products.

2:43:05

Um and so there's both flavors of those when when we talk about model busters.

2:43:08

Um but that's one uh class of things that we're doing.

2:43:12

Um, in the AI in the AI world, there's another class of things we're doing which is is like it slaps you the face that it's working.

2:43:20

Like these are the fastest growing companies that we've ever seen.

2:43:23

Like we're investors in a bunch of the fastest growing companies of the previous generation like Whiz and Slack and Deal.

2:43:28

Um, and if you put them on a chart with companies like uh, you know, Curser or 11 Labs or XAI like they get to that 100 million stage way way way faster than that previous generation company.

2:43:42

Um, so you know that's a that's a bucket and you know there's there's not like hundreds of those yet but I think there will be hundreds of those as you know the the AI applications come um because the grounds ground works being laid for that.

2:43:54

Um then you know you asked about the sort of uh kind of you know earlier stage kind of companies that raise a lot of dough.

2:44:00

Um we do do those um but you know we keep an extraordinarily high bar.

2:44:05

So, um, you know, we're investors in, um, you know, companies like SSI, um, that are working on, you know, really exciting new things with the best talent in the world.

2:44:15

Um, you know, that was part of our thesis in investing in Character AI is backing Nome Shazir and and you kind of squint and say, okay, you know, this is a person that we can back and we know that, um, you know, he's going to create something special.

2:44:28

So, those are transformative.

2:44:29

So, it's pretty he's on the founder bet, but you can really go risk on.

2:44:33

Uh I have a question for you around kind of modeling market sizing in this AI era.

2:44:39

Everybody likes to say there's an opportunity for software to not just to basically get paid for end work.

2:44:47

So instead of creating let's say software for accountants, you can create software that does accounting services, right?

2:44:54

right? the TAM expansion there of you know these end markets for you know that that look more maybe more like payroll spend or services spend uh it starts to get really exciting because they go from you know hey this is a hundred billion dollar market to this is you know could

2:45:11

something could be you know multi-t trillion dollar market around uh labor my the question that I've been um personally kind of wrestling with is uh if you if software solutions are delivering services maybe there's a short period where you can capture, you know, a huge amount of that, you know, spend. But over

2:45:29

But over time that those opportunities, there will be other software providers that are, you know, competing to do that services work.

2:45:37

And won't that have like some type of deflationary effect and overall maybe not maybe markets aren't getting smaller, but maybe prices can't stay as high as they are in a in a sort of humanled uh, you know, service market.

2:45:51

uh, you know, service market. So what are your high level you know how are you thinking about market sizing given and I and I don't expect you to have like a perfect answer here because again um a lot of these companies are model breakers as you described the model

2:46:07

model busters maners the busters um uh so uh so yeah I love this topic because it's the it's the like it's the fun one that everyone is debating right now is like well we can go capture labor um And you know without like a super deeply developed and astute strategy like that's just not going to happen. Um

2:46:29

Um and so you know if you look at like some of the biggest technology breakthroughs of all time um the way that those get priced isn't the replacement of what they are delivering. Yeah.

2:46:40

Competitive forces come in and actually like there's return on capital there's like you know different competitive positioning and often that gets competed away.

2:46:49

Um so my rule of thumb is when you have these major technology waves just assume that 90% plus of the surplus is going to be on the consumer customer side and that's great because you can still like if they work when they work you can still create incredible businesses out of it.

2:47:07

So, you know, if you take the previous wave of technology change, um, you know, what would what would you or I uh or probably any uh average American person pay for an iPhone?

2:47:21

Like a lot more than $1,000. 10K probably. Yeah.

2:47:26

You know, like we've we've we've captured like 90% of the surplus.

2:47:28

And it still turns out that Apple has an amazing business on the back of it.

2:47:33

So you know technology can be kind of magical in that way.

2:47:37

Um in the previous cycle if you just take iPhone and cloud computing um those things together created like 10 trillion of new market cap across internet software mega cap tech.

2:47:49

Now a lot of that got captured by the large tech companies but a ton of it went to new companies as well.

2:47:54

And so my simple framework for AI right now is the more disruptive the technology itself actually is and the more potentially disruptive the business model the more that favors the startup and the less that that favors the incumbent.

2:48:11

So like what was the mobile version of Salesforce. com? Salesforce. com. But if Salesforce.

2:48:17

com if that technology wave instead of being just hey access Salesforce on your mobile device if instead it was like hey we have completely new workflow uh we have a completely new interface a completely new database um and oh by the way a totally different business model.

2:48:34

Maybe you'd have a startup positioned. Yeah that makes sense.

2:48:35

Uh we're talking to David Olivich next.

2:48:37

uh how has uh how does growth underwriting work in defense tech or hard tech or American dynamism?

2:48:46

I imagine it's different because there's more debt, there's more capex.

2:48:47

It's a slightly different business, not zero marginal cost necessarily.

2:48:51

How are you thinking about underwriting the next the hard tech boom?

2:48:55

I I this is a great question and one that we sit around and talk about all the time.

2:48:59

So, u our lives are a lot easier because we have an amazing early stage American dynamism practice and so we see a lot of things.

2:49:06

We're on the cutting edge of the trends.

2:49:07

we we have the right networks.

2:49:07

Um my simple uh rule on that is you can create venture scale amazing outcomes that are even better than the software only businesses but they're probably going to be fewer and further between like they're very very hard to pull off.

2:49:23

So, you know, we saw this at um you know, not quite the degree of difficulty uh as others, but but still quite a hard one at Samsara, uh which is one of our portfolio companies, which was one of the first companies that was sort of a modern hardware plus software uh public company.

2:49:40

Um, and they've demonstrated that they can create an incredible business out of it.

2:49:43

Um, right now the four that I always go to that I talk about that are the the sort of champions uh that I always reference um are uh Andrew, uh SpaceX, Flock Safety, and Whimo.

2:49:55

These are four of our biggest portfolio companies that we've uh you know, led a bunch of rounds in and and in in some cases are some of the largest shareholders.

2:50:02

Um, in the case of Androll, um, you know, it's a model buster because they have made multiple products work and they now have trust and that's a even more defensible position than a software company could ever be in.

2:50:14

Um, so there's 700 billion of defense market cap for them to go get.

2:50:19

Um, and I think they'll go get a lot of it and they might create new market cap too.

2:50:22

Um, you know, SpaceX, um, has, you know, one of the best competitive positions of any company of all time.

2:50:30

I mean, it's like the British East India Company of space. It's incredible.

2:50:32

Um, they have Yeah, I love it.

2:50:36

They have like 20x relative market share.

2:50:39

Like that doesn't happen anywhere.

2:50:40

Um, and then, you know, Flock is doing some crazy stuff.

2:50:43

I always say it has the craziest ROI of any company we invest in because it's like we stopped kidnappings, like you know, recovered lost kids. Yeah, it's crazy.

2:50:53

You know, they solve like greater than 10% of all crime in the United States today. That's just incredible.

2:50:58

And then you know Whimo is the last one which is you know super exciting.

2:51:02

Obviously there's a tremendous amount of R&D and invested capital and difficulty in building that company but it's a magical experience.

2:51:12

Like I uh I took my family from Kentucky in Whimos uh this weekend and they were like oh my god they were very nervous and then they were like okay this is a better driver than you.

2:51:24

So that's uh it's super exciting.

2:51:24

It's growing really fast.

2:51:27

But yeah, the simple thing for me is like they can be better businesses.

2:51:31

Um, but they're much harder to pull off. Yeah, that makes sense.

2:51:35

Well, thank you so much for joining us.

2:51:36

We're going to kick it over to you. This was great. Love back and talk more. Thanks, guys. See you. Cheers.

2:51:42

I don't Is there anybody that deploys more capital than He's up there. He's on the leaderboard.

2:51:48

He's in the conversation. It's funny.

2:51:49

His his his list of investments Yeah.

2:51:52

I thought they were ranked in order of significance, but they're just alphabetical order. Oh, really?

2:51:59

And in many ways, they just get more.

2:52:01

It's like it starts basically at Anderoll and ends at at uh XAI. OX and XAI. Yeah, big big checks.

2:52:08

Well, we're excited to talk to David Ulovich, general partner at Andre on the American Dynamism team. We need an American.

2:52:14

We need this is pretty American.

2:52:17

Feels pretty American, but we need I'm proud to be an American or freeird or something on the soundboard.

2:52:22

Anyway, wake welcome to the stream. David, welcome. Good to have you here. How are you? We're great.

2:52:26

Um, I'd love to talk just jump right into it about uh Army modernization.

2:52:32

We talked to General George.

2:52:34

We had uh Driscoll on as well.

2:52:37

Uh how optimistic are you for early stage startups?

2:52:39

We know it's working at the growth stage, but is this re is this time really different?

2:52:44

We've been beating the drum of American dynamism for a couple years.

2:52:47

There's a lot of startups now.

2:52:49

Is there something changing in the army right now?

2:52:51

changing in the army right now? Yeah, I I think there is something changing and it has to change and you hear the enthusiasm from General George, General Fenton, um Secretary Driscoll, and you hear from everyone else you talk to who's just, you know, not just the people that you haven't spoken with, but the people that are walking up and down

2:53:08

the halls and the people that are working um in the DoD that they know, um for a variety of reasons, both from a we don't have the tools we need, we're not procuring things fast enough, we're not manufacturing things fast enough to support the defense industrial base, um we are attitting our systems that we thought were not attitudable uh as quickly as we would. Um so there's a

2:53:26

Um so there's a whole bunch of reasons.

2:53:28

So I think people recognize there's a need for change.

2:53:30

I think the question is will this be the moment for change?

2:53:32

And if you look at the change that's happening elsewhere in Washington and throughout the government, I think you have no reason uh other than to be optimistic that this is going to be one of those moments.

2:53:44

And you know there's going to be a lot of push back from the primes and from people that um have you know sort of had a had a delicious steak lunch every day for the last 25 30 years but like this is the moment for the early stage startups.

2:53:57

I I think I think you know we we can talk about the early stage startups but I do think it's a moment of change. Yeah.

2:54:02

I mean on American dynamism national interest investing broadly um is there a structural reason why venture capital has led here?

2:54:12

Because you could have seen this boom play out like biotech where companies go public much earlier.

2:54:18

You could have seen a private equity boom where old uh subprimes or primes get turned around taken private lbed and and and revitalized and re-industrialization really could have happened that way but it doesn't feel like it is.

2:54:30

Do I just have a narrow aperture on venture because I'm in this world or is there something structural going on? Ye yes to both.

2:54:36

Um, I think it has to be venture capital first because we're at a technology transformation moment, right?

2:54:44

Whether it's being able to use AI from uh computer vision standpoints and just rapid changes that only startups can really um embrace and and react to and take advantage of quickly.

2:54:55

Um, or just the fact that they can attract the best talent.

2:54:57

You know, look, private equity firms really manage to the to the bottom line.

2:55:01

So, they're not going to pay top dollar for the 10x engineer the way a venture capital uh or venture capital backed startup would.

2:55:06

Um so there's a bunch of reasons why it has to be a ventureback company.

2:55:10

Now that said, I think you're going to see the private equity world quickly follow.

2:55:14

All these small um subprimes and people that contribute to the manufacturing base are going to not be able to keep up with the demand and I think are going to end up getting rolled up.

2:55:26

Um you're going to see people doing even ventureback startups that sort of have a private equity sort of lens like what Jay Malik is doing.

2:55:33

you're going to see these things um start to come about where the private equity model is going to going to make sense.

2:55:38

Y um but even those things might be ventured back just like you see Jay doing with AMA. Yeah.

2:55:42

Um a lot of people think defense when they think American dynamism, but obviously the aperture is much broader.

2:55:49

Can you talk about the interesting fringe of the of the the circle of competence that for what you can put in the American Dynamism bucket as it's grown?

2:55:57

I've seen some energy deals. There's hardware.

2:56:00

There's all sorts of stuff.

2:56:00

But uh what excites you in this kind of like next era of American dynamism?

2:56:05

Yeah, look, we we're going to invest in anything that we as we say supports the national interest.

2:56:10

Um so that defense is a big part of that.

2:56:13

Um public safety is another huge area people don't think about.

2:56:16

I think that uh I think on the venture side, I'm probably the largest public safety investor in the country if if not the world.

2:56:22

Um energy, you know, we have this insatiable thirst for energy.

2:56:26

So whether it's on energy generation um or energy transmission or energy storage those are all huge opportunities uh you know the battery ecosystem that we live in today is entirely dependent on China um that might be at risk down the road.

2:56:39

So we need to have an energy uh storage infrastructure in this country that's not dependent on China.

2:56:45

That creates a whole bunch of supply chain opportunities.

2:56:48

So everything from mining to reprocessing and um production of minerals and turning those things into batteries uh is really important.

2:56:54

So there's so the aperture is wide and then I would include everything from transportation, logistics, education.

2:56:59

Um you know the the tent is big uh welcomes anybody that wants to build for American dynamism but the investments tend to focus on uh defense, energy, space, public safety, those things.

2:57:12

On on energy, I know you're in Radiant. I'm a huge fan of Doug.

2:57:16

what he's building is fantastic and uh but it is small modular reactors.

2:57:22

I think one megawatt reactors designed to replace an a traditional diesel generator.

2:57:27

Um I'm I'm kind of cautiously optimistic that we're going to see someone figure out how to just copy paste Diablo Canyon all over the country.

2:57:38

Uh but it feels like that might need some sort of uh change on the regulatory side.

2:57:43

Are you optimistic about that being the catalyst and will that happen or do we just need an entrepreneur who says, "You know what?

2:57:50

I don't care about the red tape.

2:57:52

I'm gonna go figure it out.

2:57:53

I'm going to go copy paste Diablo Canyon 45 times."

2:57:55

Or is that not even the right strategy?

2:57:58

that not even the right strategy? So look, we we we we are all all all all three of us on this uh on this pod right now have survived the greatest nuclear disaster uh in modern history, which is that after the Vogle 3 and after Vogle 4 reactors came online, we didn't just go

2:58:14

build Voggo 5, 6, 7, 8, 9, take all the workers who knew how to do it, all the supply chain pieces of of the cement makers and the steel manufacturers who knew how to make all the components and pieces that we didn't just take them and and stamp them out across the country. like that was the biggest miss in our in

2:58:27

like that was the biggest miss in our in our lifetimes frankly and it's very depressing.

2:58:31

Um now that said from a from a startup standpoint whether it's radiant and you are correct it's one megawatt.

2:58:38

Um I think there's only two ways to win in nuclear you have to do it at one megawatt or the like thousand you know the gigawatt scale and there's a bunch of reasons why I think radiant will work and why the why the gigawatt scale can work but and one one change on the regulatory front that I think people don't appreciate.

2:58:55

Yes, we need reform in the in the NRC and we need we need the regulators to be much more um productive and positive to approving new reactors and making the fuel supply chain more readily available.

2:59:07

But the one of the things that's held up new reactor designs for so long is that when you make a change to your application or you make a change to your design, you then have to go through this hugely laborious process of figuring out all the different statutes and regulations that your change impacts.

2:59:24

And there's an entire army of consultants that get paid tens of millions of dollars a year to just help navigate all of the the requirements for nuclear regulatory process or fuel, how to transport fuel, all these things.

2:59:36

With AI, you now have the ability to parse through all these regulations so that when you make a change, you can say, hey, you know, this, you know, hey, language model, tell me what are the of the 100,000 pages of regulatory paperwork, what are the things that I need to go update in my application that are affected.

2:59:51

So things that might have taken three months and $10 million of consultants work now maybe just takes a few minutes of of asking the the GPT, hey, how do I update my application that based on these changes or and and that applies to the regulator as well.

3:00:05

The regulator can now go through the applications and if they're using a language model or GPT that's tuned for the regulatory code around whether it's fuel or reactor um or transport, they can actually say, hey, look, how do I evaluate this application?

3:00:18

What are the things I need to watch out for?

3:00:19

And the entire process for nuclear regulatory approval should really be contend condensed dramatically.

3:00:25

And that was never possible before the AI boom.

3:00:27

So people don't always realize while we need nuclear power to capture the AI opportunity, we're also looking to use AI to I think enable the nuclear opportunity in future. That's very cool.

3:00:37

Ty, do you have anything or question?

3:00:37

My main question is uh you know I'm I'm sure sure this is something that you guys talk about internally, but you know you have these categories, right?

3:00:46

you have these categories, right? and let's call it defense uh public public safety uh what's your kind of internal internal dialogue around making net new investments in something like public safety when in many ways I imagine flock safety is like flock safety is the flock

3:01:05

safety of xyz right you know like people I'm sure come to you with with pitches and and uh oftentimes for things that are just inevitably on the flock safety roadmap you know and it's and it's why not just put another, you know, billion dollars into into them is kind of a calculus that I'm sure you guys run. This is like the androll of whatever is

3:01:23

This is like the androll of whatever is going to be Anderoll.

3:01:26

Um, look, sometimes that's true and we just try to put more money into andol.

3:01:29

We've tried to just put more money into flock safety.

3:01:32

We've done that many times.

3:01:32

Um, I think that as we get smarter and learn more about these markets, we recognize that the surface area is just much much more expansive than we realize.

3:01:41

So, take for instance 911 services.

3:01:44

Well, flock safety doesn't do anything with 911 services.

3:01:47

Being a 911 operator is a thankless job. It's an exhausting job.

3:01:52

The turnover rate's very high, and that entire world needs to be totally reimagined with AI where you have AI operators answering the calls, triaging the calls so that the operators can deal with calls more efficiently, first responders can have more information at their fingertips.

3:02:05

Flock safety is not going to do that.

3:02:07

Another example, you read constantly about what's happening at Newark airport with the FA, you know, the FAA air traffic control system breakdown.

3:02:14

Obviously, a ventureback startup can do a much better job than whatever we've had for the last 30 years trying to make um, you know, the air traffic controllers better equipped with, you know, really dated radar systems, not getting the information they need about what's happening down on the ground at the airports.

3:02:29

And so that's just an obvious opportunity.

3:02:32

Well, neither Andrew nor Fox safety is going to do that.

3:02:33

So we just think these surface areas get bigger and bigger the more we look especially the ones that have not been disrupted by modern software. Totally. Yeah, that makes sense.

3:02:44

Uh I last question on on space investing.

3:02:47

Um there's a lot of startups that are predicated on declining launch costs.

3:02:53

Are you seeing the trend in launch cost pricing and dollar per kilogram to orbit on trend or ahead of trend or or are we lagging there?

3:03:03

You you would have to ask my partner Katherine.

3:03:07

I'm not the space expert. I have to confess.

3:03:09

But I think that the cost is going down and you know we're seeing companies that are now in sort of that that next echelon of what happens now that launch is achievable.

3:03:17

What about when when we really expand and and be able to return things to Earth, you know, reliably and quickly?

3:03:22

VAR has made a lot of progress there.

3:03:24

So, when when that really unlocks like then what what what companies come next, I think is a question to to ask and that we're asking.

3:03:31

Yeah, she's been on the show.

3:03:31

We didn't get to answer we didn't get to ask her that, but we'll have to ask her the next time that she's on.

3:03:36

But, thank you so much for taking the time to chat with us. This was really great.

3:03:39

Yeah, great to finally have you on and enjoy the rest of LP day. We'll talk to you soon. Cheers. Yes. Thank you.

3:03:45

We need little round of applause. Applause. Fantastic.

3:03:47

Uh, next up, Anish, who I think you're buddies with.

3:03:51

You've chat, you chatted with a few times.

3:03:53

I'm excited for this conversation.

3:03:54

You can Well, more than a few times, dear friend. Okay. Okay. Yes. Yeah. Yes. Your boys. We're my boy. Your boy.

3:04:03

Uh, I'm excited to chat with him.

3:04:03

Long overdue to come on the show. Yeah. Yeah. Yeah.

3:04:06

We We've been trying to make it happen for a while.

3:04:08

Um, glad we could make it happen today, even if it's just for 15 minutes.

3:04:11

Um, but we'll bring him in.

3:04:13

Um, what are you interested in?

3:04:13

uh talking through what are the interesting uh angles to take with him today.

3:04:20

He's focused on the application layer.

3:04:22

So a lot of the if you see Andrew consumer enterprise is there delineation there both uh both but leaning consumer. Sure.

3:04:33

So um yeah I just want to get into the app layer what he's seeing how he's you know he's a extremely um busy guy and I and I've seen his process. Oh, he's in Arc Boats. Have you seen that? The electric boat.

3:04:48

Yeah, that was like a super that was a super um super random deal that doesn't I mean, he says he's AI apps investing and then an electric boat company, but I've met the founders. Very, very cool company.

3:04:59

Uh anyway, welcome to the stream, Anishh. How you doing? Hey, what's up, guys? How you doing? There we go. What's going on?

3:05:04

Why are we not in person? I want to see the size. I know. We're going to come see.

3:05:09

We uh we just uh we just signed a lease today actually for a new space.

3:05:14

Uh so next time you're in LA, we'll have to uh we'll have to have you over.

3:05:18

And we do have a comically large gong that has not been shown.

3:05:21

Uh but it's about it's about the side it's it's at least I think like seven feet tall.

3:05:25

So you can look forward to hitting that in person.

3:05:30

Man, we used to get to hang out all the time and now we got to do this. I know. I know. I know.

3:05:33

We both we both got busy.

3:05:35

But it's great to see you.

3:05:35

Thanks for making time today.

3:05:37

Uh there's so much so much to talk about.

3:05:40

Uh I was telling John um before uh before you joined a little bit of uh your background, but why don't why don't you introduce yourself um and maybe give a quick overview on on your focus at the firm. Yeah. Yeah, for sure.

3:05:53

It's great to see you, Jordy.

3:05:55

So, I'm focused on AI apps.

3:05:57

I lead our consumer team.

3:05:57

So, we do a lot in consumer. We do a lot in B2B.

3:06:00

I've started two companies.

3:06:00

I sold the first to Google, second to Credit Karma, which is why I've done a bunch of fintech investing as well. And this is great, man.

3:06:08

I built my first company in 2008 and this is the most excited I've been since 2008.

3:06:12

Like everything is working and people that are more technical are winning.

3:06:15

So, it's just like a fun time to be building or be around building.

3:06:19

Did you always have conviction around the app layer?

3:06:21

There was a period over the last few years where people were kind of oscillating between, you know, uh there's no value in the app layer, these are all rappers to back to, uh, you know, just kind of back and forth.

3:06:33

Uh, I'm guessing you never lost faith. I don't know, dude.

3:06:35

I think the rapper thing was such a midthought, you know, like for maybe for like two months rapper thing was a thing.

3:06:42

But give me a am I allowed to swear on here or not?

3:06:44

Uh, our kids are watching.

3:06:46

Yeah, we keep it usually pretty clean.

3:06:48

We're working on we're working on a bleep effect now.

3:06:50

But yeah, go go go swear swear with the people.

3:06:55

I'll save it for a beer when we hang out in person.

3:06:56

But look, if you look at the kind of evolution of that comment in particular, sure there was a minute when it seemed like you had to either spend $100 million to train a foundation model or you had to be a quote unquote rapper.

3:07:07

But as soon as you started to see fine-tuning and then all the things that came after it and then of course open source, that stopped being a consideration.

3:07:13

So anybody who tells me that there's this like rapper concern hasn't thought deeply enough about it, I think. Yeah.

3:07:18

I I also think there's an an effect there which is people that aren't seeing enough real data from companies at the application layer and seeing the growth rate of some of these companies where you're meeting with you're meeting with companies all the time that I'm sure you're that that are growing at at rates that would have shocked you in 2021 and that you're even passing on because it's maybe not even bestin-class now, right?

3:07:41

I don't I you know um and the top decile is not 0 to one in you know 12 months or 0 to2 in 12 months.

3:07:49

It's like 0 to 10 0 to 15.

3:07:49

We've seen 10 to 80 like bananas. What is Yeah. Yeah.

3:07:56

Uh I mean uh talk about open source a little bit.

3:07:59

Are you seeing that impact uh financials in any way?

3:08:01

Um you know obviously tokens can be expensive.

3:08:07

I've we've talked to some early stage founders who are pre-launch and they're like, I need $500,000 of like OpenAI credits to just do the pre-work that I need to do to make my my my startup work.

3:08:19

I haven't heard about that.

3:08:19

I mean, we we were in the free AWS credits.

3:08:22

If you go to YC, you get a bunch.

3:08:24

People would not burn through those.

3:08:26

They're burning through tokens.

3:08:28

Uh have you have you seen that as as a meaningful cost driver at the early stage?

3:08:32

And has open source shaped that at all?

3:08:34

Yeah, there there's actually two things to talk about here.

3:08:36

One is it's actually amazing that there are real costs because it forces companies to make money.

3:08:41

So it forces them to charge real prices right away which forces them to deliver value that substantiates the prices they're charging.

3:08:47

So the business model quality is way better than it was 3 years ago because ironically because the underlying software has gotten more expensive.

3:08:54

So we're seeing way better especially in consumer Jordy you know this right like so much of consumer was this field of dreams investing.

3:08:59

you know, we'll build it and someday we'll have ads, whatever that meant.

3:09:03

And a lot of times didn't work.

3:09:05

And now, like day zero, people are charging for subscriptions and they're making tons of money.

3:09:10

And actually, strong willingness to pay because people have had such magical experiences on the consumer side.

3:09:16

People will be like, "Sure, I'll pay 50 bucks a month for this." Right? Or more, you know?

3:09:21

And this is why the most interesting question right now for builders is what is the thousand a month skew of your product? Right?

3:09:26

It's not can I charge, it's what is the like insanely expensive thing.

3:09:31

And this is why I think for consumers like you're going to have food, you're going to have rent, and then you're going to have software.

3:09:36

That's what the future consumer spend looks like.

3:09:38

On the open source question, John, I think what we're actually seeing is this routing layer emerge where companies are routing to different models based on capabilities and cost.

3:09:47

So in some cases when you need something that an open source model can do or you kind of make the costbenefit trade-off, you do that.

3:09:55

Other times you might route to, you know, claude because you've got a coding thing that you need to generate or a long form writing.

3:10:00

You know, OpenAI's got really good general purpose reasoning models.

3:10:03

So the routing layer is actually a very interesting thing and it only exists because there's so many models and because there's open source and that's changing the considerations for companies like OpenAI that are now moving up the stack. They bought Windsurf.

3:10:16

They're doing you know a bunch of things at the application layer because they're no longer the sole provider of the models.

3:10:21

How are you thinking about the line between consumer proumer and bottoms up adoption in the enterprise?

3:10:29

Because I'm sure you've had this already where you make an investment into what looks like a proumer tool and then suddenly you see some Fortune 500 company adopting it in in you know different patches.

3:10:40

But um yeah, it used to be those early stage startups would be like oh well one person from an atgoogle.

3:10:46

com email signed up so let me just put Google on my landing page.

3:10:51

Now it's like, yeah, they're actually getting meaningful adoption in the enterprise just from the bottoms up perspective.

3:10:55

But yeah, what are you seeing?

3:10:57

I from my mindset, everybody's a consumer.

3:10:59

I think the greatest thing to ever happen to consumer software is giving people expense cards at companies. Oh yeah.

3:11:06

Instead of it buying something, individuals are buying something.

3:11:07

And of course, and you're it's not like you've got this work, it's you know, like work brain and and personal brain like you bring your personal brain to work and you want to use consumer grade software.

3:11:16

And we're seeing products like Korea, KRE, that's one of our best companies.

3:11:20

You know, they're like this this team is like cracked.

3:11:22

They they like live to get every model.

3:11:26

They're so technical and the product has a sort of aesthetic and a sophistication that a regular IT buyer might not understand, but it's getting huge adoption in the enterprise because the people that are in those companies want to be using the best tools and this is what they look like.

3:11:40

How do you think about the gap between what's happening in these like super agile startups?

3:11:43

It's a couple people, they ramp to hundred million in sales really quickly.

3:11:47

They're in the enterprise.

3:11:49

They're making a difference.

3:11:50

Consumer adoption versus like McKenzie is also making like billions of dollars selling like PowerPoint decks on AI adoption.

3:11:57

Is there a world where someone closes that gap and builds like a McKenzie for as a startup or something?

3:12:04

Or is that is that a dynamic that's actually good and maybe Mckenzie's underrated in some weird way?

3:12:11

I don't I mean I would argue that the main thing McKenzie Mackenzie knows how to do is sell to the enterprise. Sure.

3:12:16

There is no product there.

3:12:16

Their product is deep research or worse. Okay.

3:12:21

Everyone's a Mackenzie consultant now. No, you know. Yeah. Yeah. Because Yeah.

3:12:26

Any CEO can first question they ask how should my company be using AI?

3:12:28

And then they go and sign up for a bunch of stuff. Uh how are you thinking?

3:12:35

What what's your timeline around uh ubiquity of ads in consumer AI because we're in this interesting moment where a lot of people are paying for models.

3:12:46

Some people are using free version of models but ads aren't ubiquitous yet but the the movement from uh open AI uh you know coming from Instacart big ads business there.

3:13:00

My take on it I was debating with Jordy was that I love the world where every human being regardless of their economic condition has access to the smartest possible model.

3:13:13

But I'd love to know your take on ads in AI products because it feels like it's coming but we haven't really seen it yet.

3:13:20

really seen it yet. I mean I think that the more spiritual question for us is what is the front door to the internet you know and the front door to the internet has been Google so many years and it's really interesting actually because Google is sort of searchbased and intentionbased and the browse-based web that existed in the '9s we've kind

3:13:37

of moved away from a lot of the browse-based behavior is now in Insta and Tik Tok and these other places we're still in the command line era of AI like you know chat GBT is a command line product essentially and I think it's going evolve in a really different direction and and wherever that terminal state of that direction is will define the monetization model. I I'd be

3:13:55

I I'd be surprised if it's ads, maybe it will be.

3:13:59

Um, so I don't know that we'll ever see ads in the way that we see them today in the AI ecosystem.

3:14:05

What about uh what's happening with the hyperscalers broadly?

3:14:06

Obviously, they're all investing insane amounts of money in capex, but they've been by most reports lagging on the product side.

3:14:14

Is this something where their monopolies are so strong they're just going to wait it out and develop the products internally?

3:14:21

Is there a wave of M&A coming in AI apps?

3:14:25

Um, are they going to try and build or buy or what do you how do you think all this plays out in the big tech world?

3:14:29

I mean, probably all the above.

3:14:31

You know, number one, I think when you have a new technology, the existing companies do a good job of extending their lead in their existing markets with the new technology.

3:14:40

So, I think a bunch of that will happen.

3:14:42

You know, Microsoft will get better at delivering the word processor they've always delivered and Google will get better at delivering the search results they've always delivered.

3:14:50

The real issue for Google is not somebody builds them on search, but that there's a new front door to the internet.

3:14:54

So, the new categories are where the upstarts will dominate.

3:14:55

And that's where I think we'll see a lot of the M&A activity.

3:14:59

But it's probably all the above.

3:15:01

The one most surprising thing for me, not just for sort of incumbents, but often for incumbents, is just that people aren't using the products.

3:15:07

Like I talked to VCs, you should test investors on this.

3:15:11

like give me the five things that you're actually using every day.

3:15:15

And if the answer is chat GPT, come on.

3:15:18

You know, there's such an opportunity to build intuition by actually just using the products and yet so few people do it.

3:15:22

It's like this alpha that's hiding in plain sight. Well, that's great.

3:15:26

Thank you so much for coming on the show.

3:15:27

We'll definitely have you back. Come back on.

3:15:29

It took us too long to do this. You're not in person. I want to ring the gong. Oh, yeah.

3:15:32

Well, we have a soundbar now. We'll talk to you soon. Thank you so much.

3:15:40

And next up we have Mark Andre, the founder of Andre Horowitz coming.

3:15:42

Why is nobody talking about Mark Andre?

3:15:45

Why no one really knows about him?

3:15:47

And I I think what we need is just one more Twitter thread breaking his This was something that probably made us laugh.

3:15:57

It was a top three laugh last month was somebody posted a somebody posted a thread on Twitter. Yeah.

3:16:00

And said, "Why is nobody talking about really good job there?"

3:16:06

Uh I think that was probably entirely chatbt generated but uh shouldn't need much of an introduction. No but big day for him.

3:16:14

So uh we're we're glad to bring him in and uh ask him about open source what he's learning from the dot boom uh American dynamism geopolitics uh raising a huge fund.

3:16:28

There's so many things to talk about.

3:16:28

So excited to have Mark on the show.

3:16:31

excited to go uh deep with him on everything Andre's up to and the firm that he's built.

3:16:36

So, we will welcome into him into the studio. Welcome. How are you doing, Mark?

3:16:40

Hey guys, what's happening?

3:16:44

It's been a great I'm going to hit this real quick. The sound effect board.

3:16:45

Uh thanks so much for joining us.

3:16:48

Um and thanks for wearing us.

3:16:50

Thanks for wearing a suit. Yeah, exactly.

3:16:51

Yeah, you look special occasion. Yeah, special occasion.

3:16:55

This podcast, by the way, congratulations on the podcast.

3:16:57

I just wanted to start out by saying I've been watching.

3:17:00

It's just it's tremendous. Thank you so much.

3:17:02

Well, you know, I have a funny story.

3:17:03

We uh the the moment that I realized that you were maybe paying a little bit of attention to what we were doing and it gave me some conviction that we were on the right track is we we did a reply guy of the week to this guy Baldo and it was like this inside joke and I was like, "Holy [ __ ] Mark just followed Baldo." Very talented guy.

3:17:20

But u but yeah, so it's it's thank you for uh for following along and it's been awesome talking to your whole team today. They're fantastic.

3:17:27

Um, so let's go through some hot topics.

3:17:29

Uh, I want to start with we're in the AI era.

3:17:31

Uh, you obviously live through the dot era.

3:17:33

There's some comparisons.

3:17:36