0:10
[Music] Hey [Music] hey hey hey hey hey hey. [Music] Heat. [Music] Heat. [Music] [Music] Heat. Heat. [Music] Hey, hey, hey.
Search this video
1:14
[Music] [Music] [Music] Power. [Music] Hey. Hey. [Music] Heat. Heat.
2:07
[Music] [Music] [Music] [Music] [Music] [Music] We came to this world. to reach the stars.
3:25
We came to this world to shape our future.
3:36
[Music] to shape my future.
3:49
[Music] We came to feel the new I want to know [Music] the music.
4:38
[Music] We are surrounded by journalists. Hold your position.
4:50
[Music] >> You're watching TVPN.
5:00
Today is Tuesday, July 15th, 2025.
5:02
We are live from the TBPN ultra, the temple of technology, the fortress of finance, the capital of capital.
5:09
We have a great show for you today folks.
5:10
Lots of guests, lots of news.
5:12
The big one is we are doing a postmortem on the wind surf chaos.
5:14
Uh I was saying that uh I liked this post from Pavl Asperu.
5:20
I think he sort of nailed the postmortem, but we can debate this.
5:25
this. uh it I mean if you leave someone for dead but someone else calls in a helicopter that doesn't really change your moral positioning in the situation and Delian says this 100% the Windsurf founders in Google thought they could pull a fast one and that the broader
5:42
community/winsurf employees are somehow are worded enough to uh not catch on that their leave the shellco with cash plan was totally morally bankrupt everyone will remember and so uh I don't I'm I'm perhaps ready to put on the steel man hat and not go quite as far as Delian. I think you might agree with the
6:00
I think you might agree with the Delian and Pavle take a little bit more.
6:05
>> I still think it's important to debate because we want to avoid this kind of situation in the future.
6:09
There's probably going to be more of these deals.
6:11
John John Ludig had some good commentary on this basically saying if you are >> a big tech company it is much easier and faster to do something like this.
6:20
It's not just the FTC sort of like >> climate around antitrust.
6:25
It's just that it's much easier. Yeah. It's much faster.
6:29
>> And often times these companies are just buying the team and the talent.
6:32
So if they can license the tech and get the team, >> it is that's just capitalism doing capitalism.
6:38
In this case, the answer to capitalism doing capitalism was more capitalism >> in the form of Scott Wooh coming in hot uh working uh over the weekend to get a deal done.
6:49
So anyways, fantastic outcome.
6:53
>> Uh Scott looks like a hero.
6:55
>> Scott Wood looks like a hero.
6:55
I think Cognition is going to be stronger for it.
6:59
They got a >> worldass uh you know GTM sales marketing engine for >> what John Ludig said here was big tech's AI talent shortage means with means these M&A choices.
7:11
A you get the talent only but you get it today or B you get the talent plus the business plus the product but months after regulatory scrutiny and he says they will pick every time.
7:24
every time. Thankfully for Windsurf the product lives on but a as the default is extent is existentially dark for startups and so I think what's what's interesting is like we were debating how much of the blame goes to Lena >> and I don't actually know that there's
7:41
all that much on Lina specifically it's >> but it's fun >> it's fun it's fun to point the finger >> it is but I think boogeyman >> I think in general like the FTC has always has always had some sort of approval around big public company buying a small company. They're going to review it.
7:57
They're going to review it.
7:59
Even if they would have approved it, it would they wouldn't Google would not have gotten the talent actually in the door actually working at Google for six months. >> Yeah.
8:08
>> Months after regulatory scrutiny.
8:08
And so that time >> and the other comp here the other comp here that we have is Figma y who was slated to get acquired by Adobe. Took forever.
8:18
Adobe had to pay $1 billion breakout fee.
8:20
Now, Figma is uh going to IPO. Yep.
8:24
>> And I would be shocked if they end the first trading day below the price that Adobe was going to pay.
8:29
And the company is doing better than ever.
8:31
Has launched a bunch of new products and >> um and and so >> not financial advice, but I will give you product advice. Go to figma. com.
8:40
Think bigger, build faster.
8:40
Figma helps design and development teams build great products together.
8:43
Uh can't say anything about the stock, but >> great product. We do love it.
8:46
Um anyway, uh Varunam Ganesha had a good take here.
8:51
Winds surf employees over the weekend and it's the it's the we're so back it's so over. We're so back. It's so over.
8:57
And I think the underrated take on this like yes it's funny and it's a good point but this has a real economic cost.
9:03
Like losing sleep we're sponsored by eight sleep as we put out a video today.
9:08
Um we did sleep is valuable and if you make me lose sleep that does have an economic cost.
9:12
That does have a moral cost. You shouldn't do that.
9:13
And so you should as a founder try and land the plane.
9:20
>> Not just save everyone's life, but actually not have a super rough landing. Yeah.
9:24
Make it a win for everyone, but make it smooth for everyone. Make it smooth.
9:29
>> There was a good post uh Pavle Pavle was on a tear the last few days.
9:31
Uh defender of the founding engineers. >> Yes.
9:36
>> Uh but he said, "I hope these guys smoke Google at AI coding.
9:37
Imagine the chip on your shoulder after this if you're a Windsurf employee." Yep.
9:42
>> And Sandep Shaw over at Windsurf, he's a vice president there, said it's bigger than a chip. We are coming for it. >> Coming for it.
9:51
>> He was left behind in the Roman co. Is that correct? >> Yeah.
9:54
>> He did not go over and he's ready. He's on the go. Yeah, the ghost ship. There we go. >> The ghost ship.
9:59
And Augustus is go off the king and will depuses LFG.
10:01
Lots of lots of support for Sandep uh in the AI coding in the AI idees war it's continuing.
10:12
Um anyway uh Signal has a post here.
10:16
Startup comp used to be confusing because it was complex.
10:18
Now it's confusing because it might be worthless even with a large exit.
10:20
The founders can dip anytime.
10:22
Last chopper out of Saigon vibes except they're flying private while you're holding the bag. Good post.
10:28
Um yeah, my my my take here was that uh Scott Woo looks like the hero, not Verun Mohan, the founder and CEO of Windsurf, who should have gotten an amazing outcome for his team in any other situation.
10:42
Like in any other situation, I say, "Hey, join this fast growing company. We're just post pivot.
10:48
We're second in the market.
10:48
We're doing some cool stuff.
10:51
I'm raising some money and we're growing, but we're only doing 40 80 million ARR and I get you liquidity at 2.
10:58
4 four billion, you should be like, I am ride or die for this guy forever. Yeah.
11:03
And I think that a lot of these a lot of these employees aren't going to feel that way in in in five years, even after all the dust settles, they're going to remember that this was not not handled as well as it should have been.
11:13
And so they're going to be like, I don't want to walk with him. >> Yeah.
11:16
I'm sure when uh when the muzzle uh is re is removed, he'll be able to >> Yeah.
11:22
>> hopefully call every employee one by one. >> That would be good. Sorry. And make up.
11:26
>> And so I still put some of the blame on FTC stuff.
11:27
just FTC stuff broadly, some of the blame on Google comms, some of the blame on Lena Khan and what and and how she changed the FTC.
11:35
>> And the other thing is like this was happening really remember the uh >> uh the the sort of exclusivity period that that OpenAI had on the deal expired and I think this got announced like the very next day.
11:49
So, I'm sure it was in the works, but they definitely were moving quickly.
11:53
And again, it would have been a good outcome if they all ended up at uh at OpenAI, but I also think this is a good fit going forward as well. Yep.
12:03
>> So, I'm excited to see what each branch of the team >> run here has some push back on Signal.
12:09
He says, "This has always been true.
12:09
The idea that uh startup comp is confusing and complex.
12:13
um investors and founders uh screwing over early employees during an M&A is a time-honored tradition in Silicon Valley.
12:22
I don't know how true that is.
12:23
There's a lot of examples where that didn't happen, but of course it it it has happened.
12:27
Uh if anything, the abundance of capital has made things better for early employees.
12:31
Um and so, you know, kind of you know, people people taking both sides.
12:36
Um we didn't get to this post, uh but it was funny.
12:40
Connor saying Windsurf is the soar of startup acquisitions.
12:43
They got acquired by OpenAI, Google, and now Cognition Labs all in one month. You'll love to see it.
12:48
>> Swix has a sixway parlay.
12:48
He said, "I want Mstral founders to Apple, Mrol team to meta, character remain coded perplexity, core to perplexity, ideoggram to Figma, gamut to notion." Read it back. >> Read it back.
13:00
>> Um there was rumors that uh Apple was was um interested in in picking up Mrol. >> Oh, really?
13:08
>> Uh taking a look at it.
13:08
The funny thing though is that if you're France, do you want your national champion >> to you know the the luxury AI uh >> the straw should go to LVMH for sure >> should be home with the arnos didn't we post a joke post around that? I think we did. >> I think we did. >> Yeah.
13:27
I think >> I want fine luxury language models of the >> family.
13:32
I think some of the Mdraw folks were like this is hilarious but also like complete misinformation.
13:35
Uh anyway, um Sundep says, "It's bigger than Chip. We're coming for it."
13:40
And Rob uh says, "Let's talk about a reboot for HBO Silicon Valley."
13:45
Uh and it's funny because when you see the Windsurf tag right in the name, you're like, "Oh, okay. This is hilarious."
13:51
You think this is ridiculous.
13:52
Uh and at least you're having some fun with it.
13:56
>> Can we talk about our our upcoming guest? >> Yeah. Yeah.
13:58
We're going to have uh one of the co-creators of HBO Silicon Valley on the show.
14:02
Uh uh he has been working on on Barry for a couple years kind of out of the Silicon Valley parody world.
14:10
Uh but we're gonna try >> we're gonna try to bring him back.
14:11
We're gonna explain to him >> what's going on >> line by line what's been going on.
14:15
And I think once he really processes it, he'll be interested in making a reboot. >> Yeah.
14:21
>> Um >> well, you can reboot your finances with ramp. Time is money. Save both.
14:24
Easy to use corporate cards, bill payments, accounting, and a whole lot more all in one place. Um I like this post. Yeah, I threw this in.
14:31
Kind July brand ambassador says, "The business the business I stood on was operating at a loss." >> 40k likes. You love it. >> So good. >> Business I stood on.
14:44
>> Perfect inter intersection of thin twit and uh >> business >> and and and Justin Bieber. >> Yeah.
14:50
I remember we put up that uh when when we talked about the Justin Bieber launch.
14:54
You were like, "You're so offline, John."
14:56
And it's like I'm extremely online but just in one narrow bubble of the internet. Yeah.
15:03
>> And so I somehow missed the entire Justin Bieber saga and the meltdown.
15:08
>> Number one album in the world.
15:09
>> It's the new number one album in the world.
15:10
See this is I'm learning this again. >> I think so. >> Wow.
15:13
We listened to some of it.
15:13
We should do some I don't know if we're in the >> We actually listened We were saying we should listen to this on the drive home >> as research >> and uh we made it two songs.
15:20
It wasn't it wasn't really hitten for us, but but congrats to him on the comeback.
15:26
I mean, it seems like a great launch.
15:28
Do you think the uh do you think the viral moment was staged or do you think it was organic and then he just uh leveraged it later?
15:36
>> Seemed like a legitimate crash out. >> Mhm.
15:40
>> I mean, it seems like, you know, >> but it but it was extremely well played.
15:43
>> Even the best of us in technology have gotten frustrated with a journalist from time to time.
15:46
So, yeah, >> who amongst us are above that, right? >> Yeah.
15:50
Um, >> if the misinformation caught me outside of Nou like that.
15:55
>> Yeah, you might you might crash out.
15:57
>> I might crash out >> and then drop an album at number one.
15:57
A new a number one live stream.
16:02
>> Um, and so uh Kyle Jeang I think from the browser company or browser base um has has kind of a hot take browser companies >> kind of a hot take here.
16:10
>> kind of a hot take here. Toy boy verse gigachad Verun Mohan left his company for big tech screwed over employee equity and then of course Scott Woo can solve any math problem imaginable giving winer from employees accelerated vesting
16:23
and cliff's wave >> I like that it's math problems >> Scott basically Scott basically came in and was able to run the playbook that a lot of CEOs have run who was it from uh from Matrix uh >> Ilia >> Ilia Sukar >> uh from pars did the same exact thing with his employees. >> Yep. >> Yep.
16:43
>> Back in the day when he >> with Facebook he sold and he revested which is pretty common and and I I I I >> I would believe that uh that the team that went over to Google is revesting >> to get that multi-billion dollar.
16:55
I don't think it's just a cash and you can just quit immediately because the whole point is a talent acquisition.
17:00
So it's got to vest, right?
17:01
Um but uh yeah uh I mean crazy situation and I I do wonder I do wonder like what what the price would have been to say hey Google I'm it's it's all or nothing what price will you pay now because there's a price to wait for Google would have to pay a price to wait to actually bring over the whole windsurf company right they need FTC approval for that let's call it six months what is Google's cost.
17:33
>> Send our pitch AI doesn't maybe feel like he has six months.
17:37
>> He would miss six months.
17:38
>> If you if you want to if you want a real shot at at meaningful traction in codegen, >> another six months feels like a long time.
17:46
>> So what a billion dollar cost?
17:46
Two billion is it a zero?
17:49
Well, it could it could be the the difference between winning or or having a you know a horse in the race and not >> because, >> you know, plenty of people are leveraging Gemini broadly across their companies.
18:03
>> Haven't heard it a ton from engineers on the show talking about >> using Gemini in their coding stack. >> Yep.
18:11
>> Yapsene was a big fan. >> Yeah.
18:14
uh but uh doesn't come up a ton and they know that there's there's billions of dollars being printed in in codegen and uh I'm sure that Google wants a piece of it.
18:24
Very few very few business lines that that they I imagine they can see a through line of we can get this to a billion dollar run rate and that's kind of the bar at Google. >> Yeah.
18:34
Well, if you need to review some code, head over to graphite.
18:35
dev code review for the age of AI. Graphite on GitHub.
18:41
ship higher quality software faster.
18:43
Um, and speaking of artificial intelligence, I want to go to Tyler and get the update on the new Gwen piece.
18:47
He read it on the treadmill this morning. Is that right?
18:53
>> This morning on the treadmill. Yeah.
18:53
I I realized that there's Google on the treadmill.
18:56
So, I was like, "Oh, I'll read the new >> Wait, what article?"
18:59
>> There's like the treadmill has a has a browser. >> Yeah.
19:02
So, so usually that's the real browser wars.
19:04
>> That's the real browser wars.
19:04
Wait, this comment available? >> Just Google. But yeah, that's good. >> Wow. Break it up.
19:09
break up the treadmill market.
19:12
>> Yeah, >> this is ridiculous. >> It's a monopoly. >> Yeah.
19:14
Okay, so >> Google Google's Chrome is probably paying to be on all the Equinox treadmills >> probably.
19:20
Okay, so uh you you uh so Gor wrote this piece and it kicks off with uh Darkh Patel.
19:26
Uh what was the uh what was the actual inciting um tweet that I saw?
19:33
I saw Darkh posting about it and he was saying uh uh he was saying really interesting new Gor essay LLM daydreaming.
19:41
It's a proposal of how default mode networks for LLMs are an example of missing capabilities for search and novelty.
19:48
By the way, I know it's a big cringe.
19:50
It's a bit cringe to delight in, but if you had told 19-year-old me that a Goran essay would open like this, I would have found it hard to believe.
19:56
And it is a huge milestone for Doresh.
19:58
So, congratulations to Doresh. He deserves it.
20:01
Uh, so Dwarp Patel asks why no LLM has apparently ever made a major breakthrough or unexpected insight.
20:07
You would expect this given the incredible IQ that is on display for 15 minutes as you use an an LLM.
20:14
Um, no matter how vast their knowledge or how high their benchmark scores, while those are by definition extremely rare, contemporary chatbot style LLMs have now been used seriously by tens of millions of people since chatt November in 2022.
20:27
And it does seem like there ought to be at least some examples at this point.
20:31
This is a genuine puzzle.
20:33
When prompted with the right hints, these models can synthesize information in ways that feel tantalizingly close to true insight.
20:39
The raw components of intelligence seem to be present, but they don't. What's missing?
20:45
It's hard to say because there are so many differences between LLMs and human researchers.
20:49
So Tyler, take me through the piece.
20:50
What did Gwyn have to say? >> Yeah.
20:53
So So basically the question is like, okay, why have they not produced new research? >> Yeah.
20:57
um he he basically gives two reasons.
20:59
So so the first is like they don't have continual learning, >> right?
21:03
Which is like >> you train an LLM, you put it to inference, it's like the weights are fixed. They don't change.
21:09
>> There's sure there's like test time inference, but it's not really learning like, you know, between your chats, right? >> Yeah.
21:15
Unless it's in the context window. >> Yeah.
21:17
But there's some kind of like primitive ways to do memory you're starting to see, but it's really like not super great so far.
21:22
Y >> So he kind of there's analogy.
21:23
It's like does a person who has like permanent amnesia have have they ever like produced novel research?
21:30
Like probably not, right? >> Interesting. Yeah.
21:32
This is the PhD amnesiac.
21:34
>> Like we bring them in, they're genius, but every time they go home, they forget everything that they've learned about running our business.
21:40
And could they even figure out how the printer works?
21:44
They would have to start from first principles every single time just to print, you know, a stack of tweets for us.
21:48
And all of a sudden, we'd have a genius who couldn't really add value at our company.
21:52
If you can't print tweets, you're going to have a tough time.
21:55
>> You got to operate the Brother printer to work at this company. Uh anyway, continue. >> Yeah. Okay.
22:00
So, there's continual learning and then there's continual thinking, right?
22:03
Which is kind of like >> um >> the model is only producing tokens when you prompt it, right?
22:10
>> So, it it's not like like an average person is like daydreaming or they're like normally dreaming.
22:15
>> You you kind of are always like passively thinking about stuff.
22:16
So, it's like sometimes you wake up in the morning, >> you just have this like new solution to a problem that you're working on.
22:21
It's like, "Oh, how did that happen?
22:22
I wasn't like thinking while I was >> Yeah.
22:25
How many great ideas have been shower thoughts?" >> Yeah. Exactly. >> It's really true.
22:30
>> So, so, so basically the the solution is he has um this like system where you basically just draw like two totally random like concepts, ideas >> and then you just prompt the model.
22:39
You say >> um >> draw a connection between these two things >> and then for like almost everything it's going to be like total nonsense, right?
22:48
It's like two random things.
22:48
There's not going to be anything.
22:49
But if you do it enough times, you'll eventually get like, "Oh, this is actually like a an interesting connection between these two things. It's like a new concept. It's like a novel idea.
22:57
You can use it for research." Stuff like this. >> Yeah.
23:01
>> So, you're just randomly drawing like like almost words like cow Ethernet cable like no connection there.
23:07
But if you iterate, you might wind up with like cow and Fitbit and then you come up with the idea for that company Halter that's making like a trillion dollars and is doing really well.
23:18
And I never would have thought of that, but if I just like iterated enough to be like, wait, does a cow need a Fitbit? Maybe.
23:23
Okay, is there a business here?
23:25
And you can kind of trace through that. >> Yeah.
23:27
Basically, and then so you have this like you have like two sides, right?
23:30
You have this thing that kind of tries to make some non-obvious connection.
23:33
Then you have this other like half of the model that basically says, is this like an interesting idea? Yeah. Yeah.
23:40
>> And then basically so if it is an interesting idea, it like saves it to this kind of like memory >> bank and then you use that memory bank to to keep pulling new ideas like from that and from kind of everything else.
23:53
>> So you're like building on top of these ideas and then eventually you're going to like slowly very slowly build like these really interesting ideas that are totally novel that you'd only get from basically passing in like millions and billions of >> totally random concepts. >> Yeah. >> Yeah. >> I feel like Yeah.
24:08
One one of the things they he highlights is like this would be incredibly compute intensive, right?
24:16
Like you're talking about like running all the data centers all the time to kind of think about this and it feels like it puts me back on Kerszswall timelines of like 20 I think he says singularity in 2045 and so it's like this is an interesting idea.
24:29
What if it what if it takes a thousand times as much compute as we have now? >> Yeah.
24:34
I mean, you're basically like brute forcing research, which like kind of doesn't make sense in a way, but if you have these models that are like small enough, you can kind of distill them down.
24:42
Maybe there's >> I don't know.
24:43
Think about a scenario where founder exits their company, they do their earnout, they want to start something new.
24:48
Instead of spending two years lost in the woods, you know, think trying to come up with a decent idea, they just spend >> $5 million in five minutes and just like generate like a bunch of different ideas.
25:01
you still then have to like pick an idea and it might seem good on the surface and then you talk to people in the industry and they say well like here's here's kind of the issue with with how you're thinking and then maybe you got to do another run.
25:13
But I think the I think the interesting question is like 5 million feels like the point at which we would say like if someone built this system and they actually said hey we've we've designed the algorithms such that you can you know tell it to go and it will come back
25:29
with a good idea but every time you hit go it you get a $5 million OpenAI bill like people would push that button for sure uh and and it would be venturebacked and stuff but what if it's five billion or what if it's five trillion like there is a point where it just won't no one will push that button. >> Ideas guys would be so thrilled if it
25:46
>> Ideas guys would be so thrilled if it costs like $5 billion to come up with a good idea.
25:52
>> I think it actually does right now.
25:52
I think it might cost more.
25:53
I think the current paradigm is that even if you even if you ran this algorithm and you ran this strategy like the current models are so uh like uh what's the word?
26:05
Like they're they're not great at like compressing and they're not efficient.
26:09
And so you might be able to brute force it.
26:11
that you're brute forcing research, but the level of brute forcing might might like when you math it out might be like in the billions of dollars to get an insight. I don't know.
26:22
>> I think it it also like it doesn't maybe make sense if you're like a random person, you're going to pay $5 million for an idea, but it maybe makes sense if you're a big lab and you need >> like you're hitting a data wall, you need new data.
26:34
>> Uh I think then it kind of makes sense. >> Yeah, it makes sense.
26:36
And then also he also brings up this this thing which is like >> um if you like kind of distill the model even if you like open source it you're not really giving away the new ideas right because >> like say you have some like weird concept that you find it's baked into the model but like you the only way to access that is to like specifically query for it.
26:56
>> So if so you're not actually like giving away these new ideas even if you're um giving out the model.
27:01
>> That's another kind of interesting.
27:01
So you actually have to like inference it enough to pull the insight out of the model even if they're it it's literally the the there's a hundred million dollars in your laptop and your job is to pull it out.
27:11
Like there's a billion dollars in the weights of GPT5 or 4.
27:12
5 but you know to get the insight out of it might cost another billion dollars. >> Yeah.
27:20
It's basically like the the monkeys typing on the you know keyboard until they get Shakespeare. Yep.
27:24
>> It's like the same thing but for like >> research.
27:27
>> Do we have a monkey sound effect yet? >> No. >> No.
27:30
We got to get one of those. Uh, get on it.
27:32
Uh, what what's your reaction to Chris Best, founder of Substack, friend of the show.
27:37
He he said, uh, he's quoting this, "Contemporary chatbot style LLMs have now been used seriously by tens of millions of people since Chat PT, November 2022, and it does not seem like there it does seem like there ought to be at least some examples of novelty at this point."
27:53
And Chris says, "Potential answer, there have been, but the human users hogged the credit."
28:01
So, so maybe these chat bots are spitting out novel ideas, but people are just like, I'm not telling anyone that I got my idea for from chatbt because that's like a bad signal.
28:11
>> Yeah, I think that could be plausible.
28:11
I think it's also just like >> like the way the LM's work is like you you're querying for a specific thing. >> Yeah.
28:19
So like if you have it's more of like you have a hunch of like this could be an idea and then the LLM like really gives it to you but you you never really see like LM just like here's a cool idea right which is what this is kind of >> and this is the >> other the other thing is a lot of a lot of business ideas like good business ideas end up being simple and they don't they don't always feel like oh look at this novel insight. It's just obvious. Yeah.
28:46
Like other people will see the idea and they'll actually be if they're the kind of founder type or maybe they're looking for their next thing, they will legitimately be annoyed because they think, "Oh, I I could have thought of that." Yeah. And so >> Bit for Cows.
28:58
I could have thought of that. >> And for dogs. >> I love that business. >> Um, no.
29:06
So, so that that's the other thing is is at least in in our world in the private markets a lot of a lot of ideas that end up being worth a lot.
29:15
>> It's not necessarily the idea that's worth a lot, but it's >> the having the idea at the right time and then the execution against that idea. >> Yeah.
29:22
>> Uh for for a very long time.
29:24
>> And I I I think that kind of gets to, you know, my silly Kugans eval uh which is tell me a joke.
29:31
That requires novel insight because if you just tell me a joke that's already been out there that I've heard, it's not funny. So, it requires novelty. It requires insight.
29:43
It's actually really hard to come up with a new joke that's never been told before.
29:47
And because it requires putting multiple things together, see realizing something.
29:52
And so I think that um even though I'm not going to chachip regularly and asking for like make a scientific discovery or like give me a genius business idea, I think the tell me a joke is is in the same vein of like hard problem for LLM.
30:07
And I think that if if they get to the point where you can ask it for a novel new humorous joke insight like joke with insight that's actually new, it should be able to do the other things.
30:22
And I think that is some sort of fundamental restraint and it's it's it's what we're talking about when we talk about spiky intelligence.
30:27
It's like very good at certain things, but it's missing this one human capability.
30:31
Uh is there anything else from the post that you think we should highlight, Tyler?
30:36
>> Uh I think that was that was mostly it. >> Okay. Qu is on absolute tear. We love him. >> He's back in the bay.
30:43
>> Oh yeah, he's back in the bay. That's great.
30:45
Um well, we have uh four minutes until our uh first guest of the show.
30:48
So let's cover a little bit of the story.
30:51
Let's cover some ads first.
30:51
Let's tell you about Vanto.
30:53
Automate compliance, manage risk, improve trust continuously.
30:56
Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation whether you're pursuing your first framework or managing a complex program.
31:04
And then should we kick off with some NVIDIA background?
31:08
So when we have Zach on, >> I just want to say congratulations to Vanta for picking up Scott Holden. >> Oh, for sure. Enter the portal. They picked him up.
31:16
>> We love some personel news on this show.
31:18
from Brex over to Vanta to lead marketing efforts at Vanta.
31:22
>> Uh, wait, do we have the Do we have the post? >> The traded post. >> No. Yeah.
31:26
Uh, no, Christina's post, which was hilarious. >> Let's pull it up.
31:32
>> I didn't know she had this in her. It's great. Uh, >> I got it. >> Yeah, you got it.
31:40
>> She says, "Big day at Vanta.
31:40
After a couple of days in the transfer portal, Scott Holden commits to the Purple Lions." >> Purple l.
31:45
Looking forward to working with you, Scott. So, we love to see it.
31:49
>> That's a great photo of him, too. Wow.
31:49
I wonder what the context was for that photo. >> Like a Chad. >> Yeah.
31:55
But like, you know, you see a lot of like LinkedIn like just like head shots.
31:58
This is like the full body shot.
32:00
Really nice soft lighting.
32:00
Teeth looking extremely white. Fantastic. >> Just >> nice. >> Full breakdown. Full break down. Looking diced.
32:09
>> Anyways, let's get into the story before Zack Kukov joins.
32:10
So, uh, news today out of the journal.
32:13
Nvidia can sell AI chip to China again after CEO meets Trump. Yep.
32:18
>> Jensen's CA Hang's case proves persuasive as the US agrees to grant licenses for the H20 chip.
32:23
Uh Nvidia said it has received assurances from the Trump administration that it can sell its H28 artificial intelligence chip in China days after chief executive Jensen Hang met President Trump.
32:35
The administration's move marks a turnabout after the commerce department restricted sales of the chip in April.
32:41
costing costing Nvidia billions of dollars.
32:43
The news came during a visit by Hang to Beijing where he met he was meeting senior officials. I'm very happy.
32:50
He told reporters uh shares in Nvidia rose more than 4% in early trading on Tuesday.
32:58
So uh big uh big turn uh uh of events.
33:02
Uh >> the uh also coincides with gurly girly in China right now.
33:09
uh we got to have him on to to give us a breakdown of what's what's happening on the ground there.
33:15
>> But um yeah, a bit of a surprise.
33:15
Uh this also feels like something that that uh Saxs would have been heavily uh involved with as the AI ZAR. >> Yeah. >> So, um it's good.
33:27
>> So, um it's good. uh you know what whatever people you know a lot of people have have their feelings about Sachs but I I do believe he is a patriot and um would have had a lot of different people around the table yeah uh working on this decision >> so obviously China hawks in shambles
33:45
very upset about this I imagine um not all of them have commented on it but I imagine that with people who take the vanilla view of like sort of like zero sum competition between US and China they're going to be upset because they say, "Hey, just restrict the NVIDIA chips. They won't be able to train AI."
34:00
They won't be able to train AI."
34:02
And for a variety of reasons, uh, AI is is dangerous.
34:05
It's going to be super intelligence.
34:07
We need a democratic, a capitalist super intelligence.
34:09
We don't want a a communist authoritarian super intelligence that's controlled by a single person.
34:16
We want it to be controlled by humanity or controlled by the American voters, something like that.
34:22
On the flip side, >> we want it to be freedom pled >> for sure.
34:25
And I and I agree with that to some degree.
34:27
The question is, does this is there a is there a potential world where selling Nvidia H20s into China actually hurts Huawei more than it hurts America.
34:41
And so it's somehow advantageous for us to get the entire world standardized on the CUDA stack and the Nvidia stack.
34:49
And then and then we are able to have like more leverage over the Chinese AI ecosystem and that gives us more control to basically shift the overall amount of America pillness in AGI generally.
35:06
I don't know there are multiple sides but we're going to talk to Zack Kukov our first guest of the show.
35:12
He's been on multiple times now.
35:15
You know him, you love him.
35:15
Zack, >> he is he was almost the first guest to do a three repeat three days in a row. He made it two.
35:22
But uh but we got bored depending on we got bored of the swamp.
35:26
>> We got bored with the swamp. Sorry. It happens. >> That's all right.
35:30
>> Uh but but trade deal today gets a little bit you know uh if this week gets more chaotic >> on this topic, maybe we'll run the Yeah.
35:38
>> You'll get the >> It's a really perverse incentive.
35:39
Like I'm really hoping for horrible things to happen so I can come back and complete the repeat here. >> Yeah. Yeah.
35:45
take us from the swamp to the TSMC clean room.
35:47
It's the swamp meets clean room segment of the show. Uh so kick us off.
35:53
Uh what how how would you characterize the news exactly what's happening and then we'll go into some analysis. >> Yeah.
36:00
So I think everybody knows the uh H20 chip was specifically designed by Nvidia to meet Biden era restrictions on what chips could be exported to China. Right?
36:10
So if you zoom out, you go back to April, the beginning of Trump 2, what you have is basically a little bit of regulatory uncertainty where Nvidia spent a lot of money producing chips that are specifically designed to go to China.
36:22
And the admin comes back in April at the last minute and says, "Sorry, you can't export the chips.
36:26
We're putting in export controls.
36:28
We're have lensure requirements."
36:30
Nvidia takes a huge like $8 billion revenue impairment basically on the cost of producing and not selling this huge swath of chips.
36:37
By the way, same time, huge domestic market for H20s, lot of demand, even though they're weaker, lot of demand from startups domestically for the same chips.
36:46
Last night, basically between April and last night, you've seen a fullcourt press by Nvidia for the last x number of months where Jensen has been doing a winning campaign, a successful campaign of influence in Washington, trying to make two points exceptionally clear.
37:02
First point is we need to export these chips because the same reasons John just said the stack American dominance.
37:08
We want everyone to standardize so on and so forth.
37:12
We're losing the biggest market besides America.
37:13
And also he's made the point of look we can also comply in a more regulatory aggressive way. Right?
37:20
So there's the chip security act was introduced in Congress in May.
37:22
It's working its way through the house.
37:24
It's moving through the Senate right now.
37:27
Basically, Nvidia's helped shape the act aggressively and it says, "Look, we have to make sure that there are ways to track the diversion of these chips, track where they're going, and and this is still up in the air.
37:37
Is there going to be a kill switch built into these chips?"
37:40
This is all to say last night, huge win for Nvidia.
37:42
Admin comes out and says, "Okay, or really Nvidia comes out and says, admin has told us we are going to be allowed to export these chips to China for the first time with some checks on them, but almost an unfettered way. Markets are up.
37:55
Obviously, Asian markets are up quite a bit.
37:57
Nvidia is up quite a bit on this news.
37:59
Huge economic win for Nvidia.
38:01
Huge question mark for the national security folks who say, "Look, there's some real risks with these chips going out."
38:09
>> Is there a uh horse trade going on?
38:09
Is this part of a larger bargaining discussion where we even if it's not explicitly Nvidia chips for rare earth material minerals, something else might be happening behind the scenes and this just looks like a little bit of a give from the United States and we might be expecting something or maybe we already got something.
38:32
Talk talk to me about kind of the the trade imbalance and just the overall uh deal making that might be happening.
38:41
>> It's a really good point.
38:41
This is the the backdrop of all of this is the trade war between the US and China.
38:45
And as much as it is to you and to me and probably to most of the TBPN folks, the most important question is AI.
38:51
It is one of many questions the admin has to work through.
38:55
Rare earths by the way is a huge one.
38:57
Obviously tariffs on a number of American products are a second one. Right?
39:00
So the big question is look in April the reason we stopped and we actually put in the admin put in these export controls on Nvidia chips to begin with is in large part because of the ratcheting tension with the United States and China and now part of the reason they're reducing some of these is because the Chinese market is opening up.
39:16
We're having great progress.
39:18
Ironically, Japan has been difficult to negotiate with.
39:19
China's been a good negotiator on trade and China really wants these chips to come in because even though they want to wean off Nvidia and go to Huawei as quickly as possible, the H20s are a bridge to get to Huawei's ascend architecture, right?
39:32
To get to a home >> is the meme is the meme that uh we're going to get China dependent on US chips.
39:38
Is it really just a meme?
39:38
like the idea that that the idea that uh they're that that they would just play basically Jensen makes a good point which is like we want them to be dependent on our infrastructure but >> totally >> the Chinese government and industry will likely just say okay great we're happy to depend on these for another few years while we ramp up our own >> Did they promise not to reverse engineer them this time >> yeah that's your question this time. Not this time.
40:11
You know, we've done it we've done it thousands of times now, but this time we'll be different. >> That's right.
40:17
Promises this time Charlie Brown can hit the football for once. That's right. Exactly.
40:21
That's right. Exactly. So I guess uh you know some of the push back was uh somebody yesterday when when this news broke um >> uh Walter Bloomberg you know prominent uh journalist uh who always follows you know >> journalistic standards says Nvidia CEO Hong expects Chinese military won't use
40:41
uh AI chips which is good thing to expect but uh probably unlikely and then somebody uh quoted it and said the increasing willingness of Jensen Hong The boldly play both sides makes Nvidia one of the most serious threats to US n global technological dominance currently running. So very uh dramatic uh but but
40:59
So very uh dramatic uh but but um you know valid point of view for this person to have.
41:07
How do you think uh have you been surprised at how Jensen can effectively go?
41:13
You know, I mean, I wouldn't necessarily play both sides, but you know, this first thing he did when the chip when the the initial H20 ban happened is he flew over there and he started working on a new R&D center in Shanghai.
41:26
Um, so so he he uh is definitely a capitalist.
41:30
Um, but uh but it's been interesting to see.
41:34
The other factor here is like you're running the most valuable company in the world.
41:37
So, >> correct, >> you do have quite a bit of leverage.
41:40
>> you do have quite a bit of leverage. and he almost in some ways he's sort of like holding up the uh you know the entire global economy right a couple bad quarters and everything >> does it does feel like the good ending that I'm hoping for is like this is a
41:56
stabilizing force for the world in general and it's not it's not escalatory and even if it's like okay we're going to be competing more on chat bots and tech companies and stuff it's like this moves like the potential for a Taiwan
42:14
invasion down and like I care a lot more about that than did did Huawei ship a lot I don't know >> yeah I don't know what's I don't know if I I don't know if I agree with that well I'll say a few things you guys had the Apple and China author on right and so a big part of Apple and China was the
42:32
whole shtick of Apple comes in says we're going to teach the Chinese how to do what we do domestically in America and then they do it they get better at it and they no longer need Apple and they a domestic market that can produce in the same way. That's the risk of this
42:42
That's the risk of this export, right?
42:44
Which is that there's no dependence.
42:46
Think Jensen's a really talented operator.
42:47
And by the way, the evidence for that is there was a big letter that went out critiquing his trip to Beijing.
42:52
Only two senators signed it, right?
42:55
It was Banks on the right and Warren on the left.
42:56
And if you think that they didn't try to get other people in Congress to sign that letter, too, you would be shocked at how few people were willing to go on the record on the anti-NVIDA side of this. Right. Interesting.
43:07
The other thing I would say to you is if you think these chips aren't going to get diverted to CCP use, right, to military use, to PLA use, I think that's a little bit of a naive perspective, right?
43:16
It's impossible to prevent them. >> Yes.
43:18
But my question is, >> let's just not use them for the thing that's most critical to national security. >> Yeah.
43:26
>> It's like it would be great.
43:26
It'd be great if if uh it'd be great if they could, you know, set up a local instance of XAI's new uh >> way.
43:36
Maybe that would be really ramp that up in China. But >> okay.
43:40
>> okay. So if uh that I the slight difference that I'm I'm kind of steelmanning here is >> is uh if Apple goes into China and teaches a bunch of CNC engineers how to mill aluminum to make iPhones and they explicitly demand that those manufacturers have other clients like Huawei for smartphones like they are building up an
44:08
internal capacity If you're just selling GPUs that are fully assembled, yes, you could tear them down and re rebuild, but um maybe there is enough lock in in the CUDA ecosystem that you can't just just swap out and and issue, John, is here's another headline from the journal just a couple months ago. Nvidia to set up
44:30
Nvidia to set up research center in Shanghai.
44:31
research center in Shanghai. So they are setting up and investing in R&D locally and all of that what whatever >> whatever they know how to do there whatever they learn how to do there it will all disseminate broadly into the Chinese economy and that is just that's just a fact and and if somebody
44:52
disagrees with that >> I think it's um >> and I guess we are we are seeing like cracks in the CUDA ecosystem lock in right now with AMD and and and ROCM and Rockm Um, and so people and and the GPU and Tranium and you know, Amazon and you know, I'm sure I'm sure Meta's thinking about uh custom silicon and and on device inference at at Apple. So it's
45:14
So it's like yeah, if you can go and train a frontier model on these H20s and then distill it down and run it on Huawei SNS like that does pull their capabilities forward. >> Correct.
45:26
How what is the mooden in in DC around the idea of like AI as being a really impactful military technology right now?
45:38
Because for most people it's knowledge retrieval with chatgp and codegen with cursor cloud code windsurf devon and beyond that it's like okay my B2B SAS company is a little bit more efficient and I have some AI agent stuff in there but like people aren't really saying like >> you know oh it's transforming everything yet.
46:03
They're not showing those examples even though that's kind of the narrative concretely besides like yeah you know prompt defeat America you know like like get me from where we are here to there uh and walk me through the mood in Washington around AI as a military technology specifically.
46:20
I >> think people uh on the military side are probably some of the earliest to realize the value of AI for a couple reasons.
46:29
One um there are already real battlefields today.
46:31
There are theaters of war where automated decisions are being made that are life and death decisions.
46:37
And putting aside the perspective of any of these wars, you know, whether you're looking at Russia, Ukraine, whether you're looking at Israel in the Middle East, there are already real instances today of AI deployed on Frontier live to operate things like drones, to operate things like targeting infrastructure.
46:52
Like that's that's happening today.
46:53
It's not happening in America, but that is happening today.
46:56
The military is very good at keeping a breast of what our compatriots are doing overseas. That's the first thing.
47:03
>> Second thing I would say to you is the big unlock for a lot of this is the decreasing influence of humans in warfare, right?
47:09
Like in 1015 years ago, 20 years ago, if you were an Air Force pilot, that meant you were actually sitting in a plane doing top gun Today, if you're an Air Force pilot pilot, it means you're in Arizona in a cooled warehouse piloting a drone that's like remotely dropping bombs on things.
47:26
That unlock, there's no reason that should just be the case.
47:27
We take them out of the cockpit, put them into the warehouse.
47:30
The next step on that is why have the human piloting at all, right?
47:34
Why not have the autonomous pilot of the drone making those decisions effectively and actually fighting the frontier of war for us where there's no risk at all domestically to our people going into war.
47:45
So, there's a lot of enthusi maybe enthusiare.
47:51
I think the double-edged sword of that is if you're looking at if you're concerned about an invasion of Taiwan, it's far more concerning thinking about a China the PLA army is a couple is not an up-to-date army in many respects.
48:02
It's far more concerning thinking about a PLA army that has access to supercomputer clusters built by H20s or other kinds of American ships H20 that is able to more effectively fight in Taiwan rather than a PLA army still fighting conventional warfare where it's never been able to excel.
48:17
Yeah, I guess I guess my my push back or like question is just around um like I I hear you that AI is being used on the battlefield.
48:30
When I think of AI on the battlefield, I think basic image recognition, look at a satellite image, pick out the missile silos, or if you're a drone, like it's a very very small neural network.
48:40
I'm not I I haven't heard a lot of concrete examples of like, yeah, we're running a large transformer-based AI system to make wartime decisions, but at the same time, like that feels like a year away. >> Yeah. Yeah.
48:53
But who who's going to you know what what government is that is that good PR for any military or government in the world?
49:00
Hey, by the way, we're big we're building the biggest supercluster ever and it's for killing. It's a killing machine.
49:05
Like, who who wants to go out and say that?
49:09
One one question I have is uh is uh the mo is part of why Nvidia's been so effective on lobbying.
49:17
You mentioned not being only being able to get uh if you can only get Elizabeth Warren to sign your doc.
49:23
You're in a against >> uh you're in a rough spot.
49:26
Uh but but I saw so in 2023 they spent about half a million on total on lobbying.
49:32
2024 they spent >> this is Nvidia. >> Yeah.
49:36
They spent 640,000 >> but then in Q1 of this year they spent nearly a million.
49:41
So they clearly ramped it up.
49:44
>> But how much do you think they benefit from just being the largest company in the world and everybody being a Nvidia bag holder some way or another, right?
49:51
>> I mean success success, you know, if you if you're the largest company in the world, people want to be a part of that.
49:56
Also, look, the number of dollars you spend is not always the proxy for lobbying, right?
50:00
Like the the hard thing to understand is Jensen is spending more time than probably any other of his peer CEOs. Yeah. >> Yeah, that's right. >> Yeah.
50:09
His travel budget is probably bigger than his lobbying budget. Right.
50:12
>> That That's exactly correct.
50:12
Why would you need to have 30 registered lobbyists running around when your CEO has a personal relationship? This is the new thing.
50:20
This is the the sort of unique aspect of this admin relative to other admins, which is there is a retail politics relationshipbased component of a lot of this work that did not exist before.
50:31
And if you're Jensen and you can have the great relationship with POTUS yourself, you don't need 30 lobbyists running around.
50:37
You can fly to Mara Lago. >> He's in founder mode.
50:39
And let's not forget Donald Trump Vasheron historic owner, watch guy >> Jensen recently spotted with a $1 million Rishard Mill.
50:48
So, >> God, I've never felt more poor in my life.
50:52
>> If you if you go to shake a hand in Washington and you got a bare wrist, you might be getting booted, but if you pull up with a hitter, >> you're going to the moon.
51:01
>> What are you guys What are you boys rocking today?
51:02
What's the >> I actually forgot to put on my watch. >> Bare wrist today.
51:08
>> But, but I but I typically I typically wear Vashron Constantine >> and >> Oh, very nice. There you go.
51:15
>> And Jordy wears the pig royal oak.
51:15
Uh anyways, anything else we should be looking out for this week?
51:22
>> Only thing I would say to you, look, just John, to your point on who's using actually sophisticated AI, it's not going to be today, at least what's declassified today is not going to be, hey, we blew up this thing.
51:32
It's a lot of cyber warfare, right?
51:34
Like, and as more and more things are about, hey, I want to disrupt the infrastructure of my opponents, their water infrastructure, their electrical infrastructure, their banking and financial infrastructure, that's where LLMs are increasingly being used.
51:46
What you should look for is look do the chips get exported?
51:48
Does this actually happen or does it get flipped again? Right?
51:51
What are the gunks in the works the requirements on lensure that enter?
51:55
And then also what happens on chip security act? Does it make it through?
51:57
Does it actually get through? House great momentum.
51:59
Senate TBD too early to say.
52:02
>> Okay, that's a great take because I feel like right now today I would be more worried about a CCP controlled large language model, transformer-based large model just spamming the United States internet and social networks with a bunch of AI slop that's slightly anti- capitalist than >> which is happening.
52:28
Cinping querying >> please invade Taiwan.
52:33
>> I don't think that the I don't think that the >> Xin Ping is in is in uh >> don't make any mistakes goes please invade Taiwan.
52:39
Don't make any >> don't lose any troops. >> Yeah, make it better. >> Yeah, make it better.
52:43
Um but yeah, I I I I think that that's maybe even even it's more of a clear and present danger today. >> That's right.
52:51
>> Um which by the way, that's Tik Tok. That's Tik Tok today.
52:53
You're opening up Tik Tok.
52:55
It's not AI, but it's the soft power component of China's influence over America. >> Yes.
53:00
And being able to and being able to train that model, which is large and is probably the next version does require a lot of NVIDIA chips, was probably trained on a lot of NVIDIA chips, doesn't have the same quantifiable uh, you know, oh, this many parameters, this many gigawatts, this this many tokens went in.
53:18
Um, but if it gets 1% better and you have control over it, it it is it is a risk and it should and and it should be considered. >> Last question.
53:26
Any any chatter on the ground in uh DC about the Windsurf Google deal or are they going to find out next week on LinkedIn? >> So funny.
53:36
I had a conversation uh over the weekend with somebody who's part of this like Neo Grandian movement, which is to say very sort of Lena Khan orientation.
53:44
And I asked her, I said, "What are people waking up to this? like what's going on?
53:48
Why is nobody talking about it?
53:49
And her only answer to me was stay tuned.
53:51
So my only answer to you is stay tuned. >> Mhm. Ominous.
53:55
>> Uh last question on the Tik Tok sale.
53:59
Poly market has uh Tik Tok sale announced in 2025 at a 35% chance right now.
54:05
And uh Tik Tok banned in 2025.
54:05
I guess it was already banned, but it's just like been limbo.
54:10
Uh but but have you been hearing anything on or do you have any any any uh interesting takes or insight about uh what might happen to Tik Tok since we just landed on this topic randomly. I don't know prep this.
54:22
>> I mean there's no political demand for it.
54:23
There's no constituency who's making the argument of because if you were pro Tik Tok ban, you got what you wanted in theory, not enforceable, but you got what you wanted in theory and you've now taken the win and you've gone home.
54:32
And if you're anti- Tik Tok ban, >> what about what about my American flag toing Mag 7 CEO Mark Zuckerberg?
54:38
Is he not a constituent that wants Tik Tok banned?
54:44
>> I think it's too if he comes out too publicly pushing for that, it's like an antitrust nightmare. It's it's exactly right.
54:51
>> I mean, like the the whole Tik Tok story proves that you potentially can build a new social app if you're willing to spend or at least you could in the rise of vertical video.
55:00
could in that moment create a a meta competitor if you were willing to spend $20 billion, you know, incinerate, you know, incinerate $20 billion or whatever they did to to just spend money to acquire all the users and build your own >> network. That's exactly right.
55:14
And there is nobody on the ground in DC who's using their limited lobbying dollars for the Tik Tok ban.
55:20
What Zuck wants is better data construction, better data center construction, right?
55:23
Fix the energy policy in this country.
55:25
And it's, by the way, it's a it's a zero negative sum system.
55:29
You can only work on so many things during every congressional term, every admin term.
55:35
>> Zero sum, not zero sum. >> It's even worse.
55:37
It's even worse than zero sum.
55:38
It's everybody's got to lose. It's even worse. Zero sum system.
55:42
>> Whoever loses the least wins.
55:44
>> That is unfortunately true in politics. >> Welcome to the swamp.
55:46
Uh Zuck says he's building a 1 gawatt cluster, aiming for a 5 gawatt cluster.
55:54
>> I love I love this should count as a three repeat because we've told you three times now. Last last question. Last question. This is fun though. >> I know. I appreciate it.
56:00
I'll I'll take the three feet.
56:02
>> Um uh uh he said he said 5 gawatt cluster uh in a few years in several years.
56:08
That feels further out than most of the AI hyper accelerationists.
56:11
Uh is that an energy policy thing? What are his clear asks?
56:15
Like what what would what would what would he want to change in Washington around energy policy?
56:21
>> This is a huge energy policy problem.
56:22
Energy policy is is annoying because it happens both at the federal level but also at the state level and even at the local level, right?
56:28
You got things like zoning.
56:29
If you guys read the abundance book, you got things like zoning that impact you at the very local level.
56:32
And the critique on the left right now for a lot of these questions hard, it's a hard one for these folks to work on is gosh, look at all the environmental problems of building the data centers near communities and so on and so forth.
56:43
And I would expect that in an Okaziocortez administration or a Buddhaj administration to be a big problem in, you know, four years.
56:50
What Zuck wants today is make it easier for me to build, right?
56:54
reduce regulation at the state and local level for me to actually construct data centers.
56:57
Let me bring new nuclear reactors online, right?
56:59
Bring new sources of energy online faster in this country and and and set one federal standard to regulate many of the rules such that you don't have different states and different localities who all have patchwork systems because if you need multiple people to coordinate all at once, that's where you get these long timelines and complexity.
57:19
>> It's going to be the delta smelt versus super intelligence.
57:21
That is that is really the battle to watch. >> That is wild.
57:27
>> Uh great having you on, Zack. Thanks for having me.
57:29
>> Great to have you on again very soon.
57:31
>> We'll talk to you soon, Z. Have a great day. >> Take care.
57:33
>> Uh well, if you're looking to pick up a hitter and you're heading to Washington, go to getbzzle. com.
57:40
Bezel concier is available to source you any watch on the planet.
57:43
You can go to your bezel concier and say, "I'm heading to Washington DC.
57:45
I'm talking to the big dog.
57:47
I'm talking to the big man.
57:47
I'm going to Mara Lago and I need a hitter."
57:50
and they will hook you up. Go to getbzzle.
57:52
com >> and don't if if you get a meeting with Jensen, show respect and wear an RM wear an RM. >> Wear an RM.
58:00
I want to talk about the history of Rishard Mill. It's fascinating.
58:04
There's a I found a good thread.
58:06
I also did a Chachi PT deep deep research report.
58:08
I found an interesting >> ad today, by the way.
58:11
Uh >> little little uh note in there.
58:13
Uh, I forget what the line was, but we said it's the cold emails, the the the the warm intros, the deep research. >> The deep research. Yeah.
58:23
>> And the deep research for the show in some instances >> is literally deep research.
58:27
Um, so the watch business here says Rishard Mill has no history, no tradition, and looks like a Happy Meal toy.
58:35
So, how did it become a $ 1.
58:37
7 billion brand in 20 years, selling watches that cost more than a Ferrari? Here's the wild story.
58:43
Um and so um very very interesting story.
58:47
Most the thing that popped out to me was that there is a >> So that does not look like a happy meal toy.
58:52
>> I think it looks fantastic.
58:52
So I will not take that global juggernaut of ultra luxurious uh watchmaking.
58:58
They they call it the billionaire's handshake in elite circles.
59:02
Uh and so uh he was born in the south of France in 1951.
59:06
Rishard Mill is the name of the both the founder and the company.
59:10
Um he became obsessed with uh precision machinery and fast cars at an early age.
59:15
Uh as a teenager, >> he was such a motorsport enthusiast that he would take the train to Monaco for Grand Prix weekends, even sleeping trackside to snag a prime view of racing legends like Jackie Stewart.
59:30
>> That's probably not possible anymore.
59:30
If you try to sleep, >> you'll immediately they're like, "You can't be homeless in Monaco.
59:36
>> Where's Where's your Where's your $50,000 wristband?" >> Yeah, exactly.
59:40
that you >> Yeah, but I mean back in the 50s I mean racing and even F1 was a total like it was just like a bunch of dudes being guys basically.
59:50
It was not it was not it was not like heavily sponsored, heavily organized.
59:54
Like I don't even know if the FIA existed at the time.
59:55
Anyway, um so Jackie Stewart called the Nurburg Ring the green hell because it was so demanding and it stands to this day. It's a great quote.
1:00:04
Um and so in 1967 he is 16 years old and he is uh Richard Mill is there and he witnesses this famous crash.
1:00:14
Ferrari uh driver Lorenzo Bandini's car bursts into flames right in front of him and the incident left a vivid impression of the extremes of engineering and danger in racing.
1:00:22
Um, so he winds up joining the uh in 1974 he joins a small watch maker Finn and then in the 90s he'd become the head of watchmaking at the Parisian jeweler uh Mabusal uh where he successfully launched a watch line for the firm and so he had extensive experience handling luxury jew jewelry and fine chronographs but he was dissatisfied.
1:00:47
Uh, in Mill's own words, working with others meant he was often held back from creating time pieces that could break boundaries.
1:00:54
He felt the watch world was too conservative, too boring, and too self-centered, as he later put it.
1:00:59
Uh, and he longed to push design and engineering far beyond the status quo.
1:01:06
By the late 1990s, with decades of industry know-how under his belt, he was nearing 50 and facing a now or never moment.
1:01:12
He decided to create his own watch brand as a sort of 50th birth 50th birthday present to himself.
1:01:16
One that would realize his wildest ideas without compromise.
1:01:21
So in 1999 he co-founds his company.
1:01:24
Uh he has a trusted friend.
1:01:24
uh they established this uh new firm and uh Ottomar Pay they partner with the high with the R&D firm uh the R&D arm of of AP uh which is Ronald Eye A R&P uh and they get this legendary engineer Julio Papy and Fabris Dash Chanel to help turn sketches into reality.
1:01:46
So AP got a small stake in the venture but otherwise to >> they own about 10%.
1:01:52
>> Yeah, they own about 10%.
1:01:52
So they own like pretty much everything.
1:01:54
Um so the whole idea was to apply the cutting edge materials from Formula 1, aerospace and other advanced fields to high her.
1:02:03
>> To be clear, this is the opposite of a outsider story.
1:02:05
This is an extreme insider leveraging every possible advantage that he has but then winning on a massive scale and effectively doing the impossible. Yeah.
1:02:14
Because for most people, for most people today, if you said, "I want to start a luxury watch brand," you could raise a billion dollars and it would be very difficult to do it.
1:02:24
And that's why Rolex and Ptech and RM and these brands actually end up being very valuable is the amount of money that you would need to incinerate to try to recreate a brand with legacy like that, a brand that could be 200 years old.
1:02:37
Uh the other thing is time, right?
1:02:40
So, it's you can create a luxury brand, but it's not >> that appealing as an investor or founder if it's going to take you a hundred years because you won't be around to see >> to see your you know really achieve your goal.
1:02:52
goal. I was texting with Sean Frank from the Ridge Wallet as CEO of Ridgewallet today this morning and uh and I was kind of explaining to him in in this piece that uh Rishard Mill is considered like an like an overnight success like the
1:03:06
startup like the the the the massive quick winner in in watches and he's like only in only in fine watchmaking is grinding for 20 years to become successful like an overnight success because every other firm is hundreds of years old. >> Yeah. basically. Uh and so um they they >> Yeah. basically.
1:03:22
Uh and so um they they they struggled a ton to actually build the first watch because they wanted to build a three-dimensional object showcasing this extreme technology and they wanted to create a racing machine on the wrist. >> And they did. >> And they did.
1:03:37
Uh and so there were blood and tears in the startup phase, innovations that took a long time to become reliable, resulting in years delay for the launch of the first model.
1:03:45
It took them two years from the company's founding to perfect their inaugural watch, >> which is not even actually that much time.
1:03:51
It's a long time if you think about I raised money and and the seed investor is like so when are you guys launching?
1:03:58
>> You got 12 to 18 months of runway and you're like it's it's going to be two years. >> Yeah.
1:04:02
But I think true R&D true product development can take easily take even if you're moving really really quickly.
1:04:09
>> Open AI Figma plenty of other companies >> was like four years to build a piece of software.
1:04:14
>> Plenty plenty of great companies took a long time to build like and actually deploy the thing.
1:04:18
deploy the thing. But uh it's certainly a narrative violation around like the like the okay I got into YC I'm launching my company in 3 months launch early get feedback you know it's just like it's >> Aurora was a similar story it's it's it probably was 18 months
1:04:33
>> uh and it took you know an incredible amount of back and forth with uh the design firm that we work with and and Charlie who really led the product development um and uh yeah and then you have a certification process and and again I I think good things take time. So 18 months doesn't doesn't really
1:04:50
So 18 months doesn't doesn't really sound crazy at all. >> 24, bro. 24 >> two years.
1:04:54
But he was prepared for this.
1:04:56
He had plunged into the project knowing it would be exhilarating, exciting, challenging, fun, and a lot of hard work.
1:05:01
Finally, in 2001, Rich, the Rishard Mail brand unveiled its first creation, the RM00001 Turbion, and immediately sent sock sent shock waves through the conservative Swiss watch industry.
1:05:12
Only 17 examples were made.
1:05:15
Uh, but that was enough to make history.
1:05:16
It wasn't just the bold tonaushaped case.
1:05:19
I was looking for the word of the shape.
1:05:21
It's the tonau shaped uh and the skele skeletonized dial exposing futuristic guts that grabbed attention.
1:05:28
It was the audacity of the whole package.
1:05:30
It was high complication.
1:05:30
Had a manual turbion, which is the little spinning thing that keeps the time more accurate when your wrist is moving around.
1:05:38
uh packed with te technical innovations, a base plate of reinforced carbon fiber, a torque indicator, and a dynam dynamometer.
1:05:45
Dynamometer to monitor the movement's performance.
1:05:51
Dynamometer sounds like some American dynamism. I love it.
1:05:53
Um >> how how American dynamism pill are you dynamometer? >> Dynamometer. Yeah.
1:06:00
If you don't have a dynamometer on your wrist and you call yourself an American dyn uh they've they've averaged making 2500 watches a year since their founding. >> Wow. That's >> so low volume.
1:06:15
>> Uh but but margin >> of all time. Yeah. Crazy.
1:06:18
Uh and also they uh you know the engineering stuff really comes through.
1:06:23
Most watches have 48 hour power reserve.
1:06:25
RM went with 72 power reserve.
1:06:29
And there are some crazy crazy stories here.
1:06:31
So uh th this is where the conspiracy theory comes in.
1:06:33
So the price tag >> I got to tell you. Yeah, please.
1:06:37
>> Uh one of our one of our dear friends who I will not name uh who is a is a size lord and capital allocator in real time just sent me a picture of his RM on his wrist while a helicopter is taking off in the background.
1:06:54
>> So anyway, beautiful piece.
1:06:56
>> It's a fantastic piece. Okay.
1:06:56
So anyway, uh so RM launched uh and it was $135,000 for the first watch this guy ever launches in 2001. $135,000.
1:07:07
This is twice as expensive as the next most expensive turbion.
1:07:13
So just take the mo Turbion watches are already really expensive.
1:07:16
It's incredibly complicated.
1:07:18
Take the the most expensive one and just double it.
1:07:20
And and so there is a uh so so people are protect fans. They're shocked.
1:07:25
Uh there bluntly asked Rishard Mill why he thought anyone would pay such a sum for his newcomer watch and uh Rishard Mill said very easy because we are not competing with Pate Phipe nor anyone else.
1:07:38
Uh in other words he's playing in the league of his own.
1:07:41
So >> so that that line doesn't typically work on a pitch call >> at least in the software business. >> We're not competing.
1:07:48
So you're you're charging uh you're building this, you know, database company.
1:07:52
You're charging twice as much as a full enterprise snowflake deployment and >> you you you're going to win because you you say you're not competing with them. >> Yes. Yes.
1:08:04
Um and so uh uh orders came in as soon as the first watches went out the door.
1:08:10
Uh but there is this interesting conspiracy theory uh which is that there was a mistake and the reason that they are so expensive is because the uh uh someone uh >> he's doing a print ad. >> Yes.
1:08:24
And so okay so uh a rumor has floated around that Rishard Mills first advertisement which was in the Financial Times baby you know it you love it in the pink sheets.
1:08:35
uh uh in the uh so he buys an advertisement in the Financial Times and the advertising department misprints the price.
1:08:47
>> He wanted they added a zero.
1:08:47
He wanted to sell the RM00001 for $13,500.
1:08:54
They printed in the ad $135,000, 10 times the price.
1:08:58
And what's interesting is that so the ad goes out.
1:09:04
According to Legend, the RM team is like, "You guys botched it.
1:09:06
Like, who's going to buy this?
1:09:08
It's only a $13,000 watch.
1:09:10
You're trying to sell it for$135. This is crazy."
1:09:11
But then they start getting calls. I got to have it. I got to have it. $135.
1:09:15
I I've been I I have $135,000 burning a hole in my pocket right now, and I would love this watch. I want it on my wrist.
1:09:23
And so, uh, Mill received so many inquiries at that price, he decided to stick with the higher number.
1:09:30
And so, this is why it's a conspiracy theory is because Rishard Mill as a brand denies that this ever happened.
1:09:36
Uh, but the legend >> obviously terrible, terrible if it came out that Yeah.
1:09:40
they wanted to make a $13,000 watch and then they accidentally >> charged 135 and then >> they look around, the orders are coming in. Exactly.
1:09:50
like wait we we just you know a thousandx our our our you know >> yeah but this is a classic example of of what's called a vellin good.
1:09:59
So with in economics typically the as price increases uh demand decreases if supply is is constrained.
1:10:08
So um if you raise the price people buy less.
1:10:10
With a vllin good a higher price actually increases demand.
1:10:16
Um, and it's it's reserved for perfect examples of this, like luxury goods, where it is a status symbol.
1:10:20
And now when you see uh the billionaire's handshake going on, you know that's a $100,000 watch at least, potentially much more.
1:10:31
>> Pricing is a uh intimately tied with uh the the way in which people feel about a product, right? >> 100%.
1:10:40
>> And uh you know, for for a good a good example of this is the uh GT4 RS. Yeah.
1:10:45
Yeah, >> it's incredible to drive, but people want it maybe like I know very few people that can honestly say if if price was was no >> Yeah.
1:10:56
>> Uh wasn't part of the equation, would you want a GT3 RS or a GT4 RS?
1:10:58
And there's like an aesthetic component as well.
1:11:03
>> But but even though many people, if they just drive both cars find that the mid-enine >> version is is is fantastic and enjoyable, >> they'll still GT3 for the for the status.
1:11:16
Uh an even stronger example in the car world is probably like the Lamborghini Revventon or the Lamborghini Veneno.
1:11:23
Are you familiar with these?
1:11:23
So these are extremely limited Lamborghini releases and they're built on the existing platforms.
1:11:31
So they take a Huracan uh and then they restyle it and it looks way more extreme, way more and they tune it and they improve it as much as they can.
1:11:41
But these limited Lamborghinis sell for millions of dollars.
1:11:45
And really like you're buying it because it's signaling that you have that much money.
1:11:50
Uh I mean and to some extent a Lamborghini is a is a Vlinin good to begin with, but this is like the most extreme.
1:11:58
And of course a Vlin good is not purely binary.
1:11:59
You can be more vellin than the other.
1:12:01
And more more of the more of the decision-making can be driven by the higher price than than others.
1:12:07
Um anyway, uh Rishard Mills conviction proved infectious.
1:12:10
Despite industry skepticism, the RM00001 found buyers almost immediately.
1:12:16
Orders came in and GQ later observed Rishard Mills very existence knocked some of the stuffiness out of luxury watchmaking by showing that expensive watches don't have to be stayed or classical.
1:12:28
And so, >> yeah, they they've always been controversial, but for what for what RM does, they do it extremely well. >> Yes, >> it is.
1:12:39
They they own their category.
1:12:42
They truly don't have competition. Yeah.
1:12:44
>> Any other watch that I've seen that that tries to emulate what they do.
1:12:46
It does look like a cheap knockoff. >> 100%.
1:12:50
And uh yeah, it is it is truly you know they basically anytime you have a situation where where um insiders come into an industry and have every possible advantage but then absolutely hit like such a grand slam, you just got to appreciate greatness. >> Totally.
1:13:08
And I think that what matters with the story of the actual watches like the RMS is that it's not just a knockoff Nautilus with twice the price.
1:13:20
It It's differentiated on price, yes, but it's also differentiated on shape.
1:13:25
There's no other major watch that looks exactly the tonau shape in my opinion.
1:13:31
The the putting skeletoniz so forward doesn't exist anywhere.
1:13:34
So that's like the fourth the third or fourth point.
1:13:37
And then the last point that's interesting is that every luxury watch brand, they're not telling you this is a rugged watch.
1:13:44
And Rishard Mill was all about ruggedness.
1:13:48
So there's an interesting point in here.
1:13:49
He says, uh, one of his breakthrough strategies in in establishing credibility was to demonstrate the ruggedness of his high-tech creations in a way that bordered on outrageous outrageousness.
1:14:00
um at watch fairs and presentations, he would literally take a prototype and throw it on the ground to prove it wouldn't break.
1:14:06
This stunt wasn't just for shock value.
1:14:08
It was making a point that a Rishard Mill wasn't a delicate museum piece, but a luxury watch that could withstand anything.
1:14:15
The company engineered every component for durability.
1:14:18
Cases were curved for strength and comfort.
1:14:21
Movements were torture tested with massive G-forces and stocks.
1:14:25
As one company monograph described, prototypes were subjected to simply brutal trials, including a device nicknamed the goat's foot, a heavy pendulum hammer that smacks the watch from all angles.
1:14:36
If a part failed, Mills team iterated until it survived.
1:14:39
The result was that these extravagant watches could truly be worn during >> Yeah. Yeah.
1:14:47
This is why people see somebody playing tennis in an RM and they're like, "Oh, I can't believe they would they would they would be, you know, playing, you know, doing some sport with a watch like this."
1:14:56
And it's it's like, "No, actually was designed to do this and it does it very well." >> Yes.
1:15:02
Uh the tennis thing is is a very interesting uh example.
1:15:05
Uh the first Athletic ambassador was F1 driver Philippe Masa who joined the Rishard Mill family in 2004.
1:15:14
>> Related to to Masa Masa.
1:15:14
Yeah, the last name it has two S's. >> He flipped it.
1:15:19
It was so powerful he flipped it.
1:15:20
>> Uh, and so, uh, in practice sessions and races, he wore an a prototype RM chronograph to see how it held up under high G forces and vibration.
1:15:26
So, even if it's not going to destroy the watch, it still might throw off the time.
1:15:30
And actually having a a watch that could hold up.
1:15:34
Maybe it's not in maybe it's not uh, you know, remarkably important for a watch, but it's a cool cool stat. It's novel. It makes it special.
1:15:41
It makes it different than everything else.
1:15:44
Um, but this partnership paid off dramatically.
1:15:45
During a violent crash at the 2004 Hungarian Grand Prix, Ma Moss's car was wrecked, but both the driver and his Rishard Mel RM00006 Turbion walked away unscathed.
1:15:56
The watch's carbon nanopiber movement, proving its toughness in extreme conditions.
1:16:00
Um, and then the other the other tennis example that I wanted to get to was in 2010 after years of experimentation, the brand introduced one of its most famous collaborations, the RM027 Turbion with tennis icon Rafael Nadal.
1:16:15
This watch became legendary for its absurdly low weight under 20 grams with the strap achieved through lithium alloy movement and skeletonized design.
1:16:22
Nadal, a fierce baseline player, put it through its paces in Grand Slam matches.
1:16:26
In fact, Nodal reportedly broke five par five different prototypes of the RM027.
1:16:34
So that's probably like $2 million worth of sales.
1:16:37
>> Although it's not like the prototypes actually cost the the retail price.
1:16:39
But but yeah, >> remember these things have 99% margin apparently if if the rumors to be believed, but uh >> and so uh this allowed them to test it and the engineers perfected it to withstand his powerful swings.
1:16:52
The finalized watch accompanied Nadal to victory at the 2010 US Open.
1:16:56
Visibly strapped to his wrist as he hoisted the trophy.
1:17:02
The site was astonishing.
1:17:02
A $525,000 turbion on a sweat soaked wristband, surviving every smash, every serve.
1:17:09
It sent a clear message that wearing a Rishard Mill was compatible with even the most intense athletic pursuits.
1:17:16
A perfect melding of luxury and performance.
1:17:21
per another perfect example, our our friend uh who sent us this video is actively heli skiing.
1:17:25
So >> really putting putting it through its paces actually >> not at this very moment in time but the recent video >> uh that's hilarious. Okay.
1:17:39
>> Uh and and then of course once uh once Richard Mill became kind of dominant in sports they wound up uh uh you know breaking out slowly.
1:17:46
So in in 2011, this is 10 over a decade after into the buildup of the business, they partnered with Hollywood actress Michelle Yo and co-designed the RM051 turbion uh that featured a diamond set Phoenix motif.
1:18:01
They did a collab with Natalie Portman on the RM1901 with a dramatic spider design.
1:18:05
Uh, and they're just doing all these like anything they can do to create something that's extremely unique, identified with like equestrian polo players, yachtsmen, extreme skiers.
1:18:16
So, just really focused on this brand and willing to do anything even if it means building something from scratch like a piece unique.
1:18:22
Um, uncompromi uncompromising functionality under real world stress, which is just it's a promise that no other brand is even trying to make. >> Yeah. They own their category.
1:18:31
They they found a new category and then they dominate it.
1:18:37
>> Um >> it's absolutely beautiful.
1:18:37
I'm pissed there's no Harvard case study, Harvard Business School case study on the business yet.
1:18:44
>> There are some GQ articles.
1:18:44
There are some Hodinki interviews.
1:18:48
>> I need us I need a proper SWAT analysis. >> For sure. For sure.
1:18:52
>> Guess we should go to Grock 4 heavy and say write a Harvard Business School case study on Rashard Mill. >> Yeah.
1:18:58
>> Please include a SWAT analysis. >> Be good.
1:19:01
>> Don't make any mistakes.
1:19:01
That's kind of what I did with deep research.
1:19:02
I basically asked for that.
1:19:04
But um >> no, it's just such such an incredible story because it's it's effectively it was incredible execution all the way from founding until today. >> Yep.
1:19:17
>> Uh but still doing something that most people would have said was completely impossible. >> Yeah. No discounts ever. 5,000 watches per year.
1:19:25
Only sold in elite boutiques Dubai, Monaco, Beverly Hills.
1:19:28
partners with athletes in rich sports, F1, tennis, golf, >> to be clear. >> Yeah, pretty.
1:19:36
>> There's a certain podcaster that we really want to get want him to start, you know, because he gets really into his shows.
1:19:41
So, having having, you know, an RM, >> it just makes sense because of the weight, like the performance. Yeah.
1:19:48
>> He's putting up five, six, seven hours in the studio.
1:19:52
>> But it gets toy, it gets hate.
1:19:52
They call it a Happy Meal toy.
1:19:54
They call it the rich man's Invicta.
1:19:56
They say it looks like plastic.
1:19:58
It's saying it has a lack of heritage.
1:19:59
It's only 24 years old, but it doesn't matter. RM is here to stay.
1:20:04
It's redefining luxury in real time. Anyway, uh fun story.
1:20:06
If you're planning to build a new watch brand, get on linear. Linear. app.
1:20:11
Linear is a purpose-built tool for planning and building products.
1:20:14
Meet the system for modern software development.
1:20:16
Streamline issues, projects, and product road mapaps.
1:20:20
Start building at linear. app.
1:20:22
Um, >> linear really is the Rashard Mel of product management, you know, platforms.
1:20:28
>> People have been saying it.
1:20:29
>> We've been saying it since about 10 seconds ago.
1:20:32
>> And when you go to sell a watch online, got to use numeral HQ, sales tax on autopilot.
1:20:36
Spend less than five minutes per month on sales tax compliance at numeralhq. com.
1:20:41
Uh, anyway, David Protein, crazy stuff happening there.
1:20:44
I have I still don't know what's going on.
1:20:47
>> I don't really know what's I I have some people texting me with their thesis.
1:20:49
I want to hear yours, but let me kick it off with some basic uh background.
1:20:53
So, David protein announced yesterday.
1:20:56
Our commitment to protein has led us to a strange place. Boiled cod. We're selling it now.
1:21:03
Uh and a week ago, it seems like Peter Rahal, the CEO, friend of the show, uh was teasing this by saying, "What should we make next? We will make what wins."
1:21:12
And he said cookies and cream, salted calmer, caramel, key lime pie, frozen cod, and cod lost by a huge margin, but they made that.
1:21:20
And Ryan Harmon says cod lime pie.
1:21:22
So maybe he's in on this stunt.
1:21:25
I I'm sniffing a stunt here. We'll figure it out.
1:21:27
We'll get to the bottom of this.
1:21:29
So Andy Rosenberg says David teased the boiled cod product on its website months ago because 75% of the calories from David come from protein and the only thing that's higher is boiled cod which has 85% of its calories coming from protein.
1:21:44
Uh the built bar is at 52% the quest bar is at 47% and the bear bells which is delicious but it's more like a candy bar is is something like 40%.
1:21:56
>> And so >> that's your dirty little secret. >> Oh yeah.
1:21:58
If I'm picking between David and Bearbells, it's like, you know, how hard am I going today?
1:22:03
>> Well, let's let's be let's be real.
1:22:03
You you you've been known to have three in a day, so you might do two Davids and then a Bear as a little treat.
1:22:10
>> Also, the Bearbells, there's some that like like they they they get you down the funnel of like this one's like actually a protein bar and then you get like, well, like I tried them all and like the one that's just like pure caramel tastes the best, so I'll stick with that.
1:22:22
And then you read the back of the label for the Bearbells caramel one and you're like, "Wait, this is actually just a candy bar, but it's delicious and it does have some protein and they have a great product.
1:22:30
I think Bearbells is fantastic."
1:22:32
Uh, but there's a time and a place for everything.
1:22:33
Sometimes you got to eat boiled cod apparently.
1:22:35
Um, but there's a big question.
1:22:38
So, they put up an outofhome ad campaign.
1:22:40
Hopefully, they did that on adquick. com.
1:22:45
Out of home advertising made easy and measurable.
1:22:46
Say goodbye to the headaches of out of home advertising.
1:22:47
Only adqu combines technology.
1:22:49
out of home expertise and data to enable efficient, seamless ad buying across the globe.
1:22:52
Uh I don't know if they went with adquick.
1:22:55
They should have, but the ad says boiled cod.
1:22:58
It has a picture of the uh of the cod and it appears to be in New York with a gray uh is that a boxer driving by.
1:23:06
That's some >> I think they're actually selling cod. >> You can just buy it.
1:23:12
>> Are they actually buy it? Okay.
1:23:12
So, uh first off, my reaction. Uh this sounds gross. I don't like this.
1:23:17
I don't care how high protein it is. I'm not into it. So, what's going on? They bought a billboard. Hopefully, Mad Quick.
1:23:23
Uh, and it feels like maybe a stunt of some sort, but they are such a young company.
1:23:28
This might actually confuse people um because people are still just learning about David.
1:23:32
Um but um I was talking to a friend of the show and he was saying to me that uh you know at a certain point uh uh a little inside scoop on boiled cod view from perspective of an earned media earned discussion uh as opposed to paid media ads a segment of TBPN like this one proves that if people are talking about a company that's usually a win especially when the convo is about cod having the highest protein to calorie ratio of any food.
1:24:02
David Bars are second only to boiled cod. Can't beat them. Join them.
1:24:07
And so that's an interesting take is that hey, all earned media is good media.
1:24:12
But if this gets out of control and all of a sudden people think of David Bars as cod flavored and when they go to grab the next David Bar, they're like that could be a risk.
1:24:19
But it seems like >> I order if I order a pack of these will you will eat one on the show.
1:24:23
I will I will test it for sure. >> Okay. We'll work on it. I'm holding you to that.
1:24:28
And then the other question is like it seems like they're in a battle with uh some other brands over their supplier.
1:24:34
David Bar, it seems like they acquired their their supplier.
1:24:37
Uh the key ingredient supplier makes EPG or estrified propoxillated glycol, a fat alternative developed by Epig and made from modified plant-based oil.
1:24:50
It's a controversial ingredient choice because it's less it's less natural than raw meat like boiled cod.
1:24:56
boiled cod seems like the other end of the barbell.
1:24:57
Uh but I need to know more.
1:25:00
Uh and I'm and I'm curious to know how this played out, how this is working, and are they seeing results from this campaign because it's uh it's exciting stuff. It's cool.
1:25:08
Um anyway, uh if you if if if they do wind up buying that uh EPG supplier and they need to get some new clients, they got to get on. >> I did. >> Okay.
1:25:19
Well, then they got to get on ADO customer relationship magic.
1:25:21
Adio is the AI native CRM that builds, scales, and grows your company. to the next level.
1:25:26
You can get started for free.
1:25:26
Um, in other news, >> what else we got?
1:25:31
>> We got a bunch of stuff.
1:25:31
We got Finn AAI, the number one AI agent for customer service. >> You were on it.
1:25:36
>> Number one of performance benchmarks, number one in competitive bakeoffs, number one ranking on G2. Finn. AI.
1:25:39
Uh, high yield Harry says, "This is going to end in tears."
1:25:43
He's quote tweeting Jim Kramer, a new acronym for the meme stocks that just won't quit.
1:25:48
>> Park your money in >> Palunteer, Apploving, Robin Hood, and Coinbase.
1:25:54
the four wolves of this bull market.
1:25:57
>> They are the four wolves of this of this bull market.
1:26:00
>> How's Apploving doing?
1:26:00
I haven't you haven't uh seen people posting charts, but uh >> well, there was all that conspiracy theory that like they weren't making any money, but we talked to Sean Frank and he was like, I'm using AppLoving.
1:26:09
It's like it's maybe not the best, but it's pretty good and I'm actually paying them so it is real.
1:26:14
>> Year to year to date, they're flat, but they're still $118 billion business. >> Okay. Okay.
1:26:19
Well, that's pretty good. Um, interesting.
1:26:22
In other news, >> yeah, their 2024 revenue was 4. 7 billion.
1:26:29
Let's see what they put up this year.
1:26:31
>> We had, uh, there was an interesting post here.
1:26:33
So, this founder that has been beefing with Shan Magcguire, >> uh, creating a bunch of websites about him, uh, emailing his his boss, uh, emailing his LPs, things like that.
1:26:46
has been going around claiming that he's a portfolio uh a Sequoia portfolio founder >> when in reality he got a 10K scout check and then his company shut down.
1:26:56
Um so kind of throwing stones from a uh from a bit of a a glass house there.
1:27:03
Uh but yeah, if you get a scout check from a you know venture scout from a fund that does not qualify as getting an investment from that company. >> Yeah, that's true.
1:27:15
this is a violation of the social contract in Silicon Valley. >> Yep.
1:27:21
>> Um and Sean clearly looked this up and was like, "Wait, did we invested this guy a long time ago?" Dug through it.
1:27:28
And so, >> I mean, the whole thing with the whole thing with Sequoia is like they at this point they have a lot of companies in the portfolio.
1:27:33
But I think if you if you actually look at the stats, it's like >> the like 99% of companies that are like still operating and so he would know he would know if they had done the deal.
1:27:44
So, >> so Sean really wrote him.
1:27:47
>> How much did Sequoia invest in your company?
1:27:48
What partner led the investment and what happened to your company?
1:27:50
Can you please enlighten me?
1:27:51
The the answer is Sequoia invested zero.
1:27:53
Sequoa Scout invested 10K and the company failed. Rough.
1:27:58
Well, don't get in the mud pit if you don't want to get muddy.
1:28:04
>> If you hate mud wrestling, >> if you go in the mudpit of X.
1:28:06
com with Shawn Magcguire, you're going to get muddy.
1:28:11
He's going to he's going to he's going to roast you.
1:28:14
>> Brian Singerman has a new fund. Yes. PT is a big backer.
1:28:17
The TechCrunch is reporting.
1:28:19
So >> interesting strategy.
1:28:20
Did you read this article?
1:28:22
>> I was going to get into it.
1:28:22
So former founders fund GP Brian Singerman and co-founder and managing partner of Quiet Capital.
1:28:27
Lee Lindon are seeking over 500 million for a new fund called GPX. Love the name.
1:28:33
Three people familiar with their strategy told TechCrunch.
1:28:35
So sounds like Techrunch is scooping them.
1:28:39
um which I'm sure they're not happy with.
1:28:42
>> Um but uh it makes sense that these kind of things get out.
1:28:44
So GPX uses a two-prong strategy.
1:28:47
The firm will invest approximately 20% of the capital into funds managed by emerging VCs who are targeting preede and seedstage startups and the remaining capital will go towards partnering with emerging managers uh the same emerging managers you would imagine on leading later stage investments most most likely at series B of their breakout companies.
1:29:05
So often times uh let's say a preede uh focus or or early stage fund does a preede round for a company they end up uh growing really quickly.
1:29:16
quickly. uh GPX can come in and and basically provide the capital in order to help that manager uh actually lead later rounds versus that manager going having to raise an SPV from 50 people >> definely happens and it's kind of like it's harder to win it's harder to
1:29:33
actually win the deal because it it can take a lot of time the capital is not actually fully ready to deploy >> I've been in that situation where the early stage seed fund seed fund is like you know yeah would love to lead this next round just give me six months to like, you know, actually close all the money. I'm going to go around and then
1:29:49
I'm going to go around and then you have this weird like I got to talk to this person, then I got to talk to you and then I got to talk to this person, I got to talk to you and then I got to talk to this person and it's just so much back and forth.
1:29:56
So if you have this fund to funds backer who just lets you stretch one round further uh really really cool concept and I think it could work out really well.
1:30:06
Um yeah uh Brian Singerman great uh great investor, great person.
1:30:12
I always had a fun chat with him when I was at Founders Fund.
1:30:14
excited to see him uh uh launch this fund and kind of um get it going with a new strategy because it seems really cool.
1:30:21
Uh anyway, we should talk about eight sleep.
1:30:27
We launched a >> new ad today. Let's play with them.
1:30:28
They want to play our uh our advertising for eight sleep.
1:30:33
>> This is the big challenge for our production team is just playing videos on the show.
1:30:36
They can do everything else.
1:30:38
>> It's pretty complicated, >> but uh it it's not easy. It's no joke. >> We got this.
1:30:41
So, let's play the ad and then we will tell you more about Eat Sleep.
1:30:44
People see the three-hour streams, the hundreds of guests, the thousands of clips.
1:30:50
For us, it's the work that happens off mic.
1:30:55
>> The cold emails, the warm intros, the deep research. >> Sleep is the edge.
1:31:02
>> Whether it's tracking down a viral poster >> or interviewing a Fortune 500 CEO, >> staying sharp is non-negotiable.
1:31:09
>> Podcasters sleep on eight sleep.
1:31:13
It's so funny that we did the white suits that day because it was just complete chance that we were in white suits that day. >> Complete chance. Um, it's it's funny.
1:31:20
So, some some backstory on this.
1:31:20
Uh, >> Andrea was on X saying, "This is what you get when you're allowed to have creative freedom with advertising.
1:31:27
Can't wait to see more of these." Well, more are coming.
1:31:30
So, uh, you can look out for that.
1:31:32
But the cool thing here with Eight Sleep team, we didn't tell them the concept for the shoot.
1:31:36
We didn't get them notes on the shoot.
1:31:38
All we did was actually send them this final cut that we posted and they were like, "This is awesome. RIP it."
1:31:44
And so, and that's part of part of what we're doing.
1:31:46
I mean, we we we make these for fun and uh it's a good creative process for the team. >> Yeah.
1:31:54
>> And it's But it was it was cool. >> Yeah.
1:31:56
They It's also like an homage to the other eight uh partnership announcement between Charlotte Clair and Eight Sleep that they clearly spent a lot of time concepting and shooting and we were just like let's match that one for one and it'll be a lot of uh it'll be cool and it'll also be funny and so we had a lot of fun shooting that.
1:32:11
So you can go to eight asleep. com to get started.
1:32:14
Uh 5-year warranty, 30 night risk-free trial, free returns, free shipping, >> code tbp. com.
1:32:21
Uh in other news, Reindustrialize is this week >> happening tomorrow. happening tomorrow. Yes.
1:32:26
And they have been on an absolute tear posting interesting things.
1:32:28
Uh they said we exist to reverse this chart.
1:32:30
Uh the top employer in each state back in 1990 you can see the map is almost entirely green because almost every state the top employer in each state was manufacturing.
1:32:42
In 2024 the top employer is healthcare.
1:32:46
And there's a few states that are holding on to manufacturing.
1:32:48
Uh a couple that are Utah's in retail.
1:32:52
Nevada has been in hospitality the entire time. Surprise, surprise. Uh, staying strong.
1:32:57
Uh, Hawaii is the top employer.
1:32:59
Hospitality, that makes sense, is a tourist destination.
1:33:01
But, um, you know, for a long time, America was a manufacturing powering powerhouse.
1:33:05
And this is a really striking chart or visualization of what can happen in 34 years of uh, neglect essentially.
1:33:13
So, uh, if you're headed to re-industrialize, let us know how it goes.
1:33:19
uh we couldn't be there because of travel schedules and whatnot, but uh I hope it's a fantastic event and we're very excited for uh everyone who's over there.
1:33:28
Um well, our next guest is ready to join us.
1:33:32
This is Brian from Cororeweave.
1:33:34
We will be talking about the Neoclouds and uh AI training and >> what it's like being a public company. >> Yeah.
1:33:40
So, good to meet you, Brian. How you doing? >> What's going on? >> Pretty good.
1:33:44
>> Welcome to the show for having me.
1:33:46
>> Congratulations on all the success.
1:33:46
Um, can you uh just take us through like a brief introduction and kind of maybe a little bit of history on the company, but kind of then kind of where you are today because I I I imagine we can jump into a bunch of stuff.
1:33:57
Most people are probably familiar, but >> Yeah, sure.
1:33:59
So, uh, first off, my name is Brian Benturo. I'm one of the founders.
1:34:02
I'm the chief strategy officer of Cororeweave.
1:34:03
I started out as CTO until they found me out.
1:34:06
I've been demoted to chief strategy officer.
1:34:07
So, uh, it's >> promoted. It's got to be more fun. More fun.
1:34:12
>> No, it's it's demoted, dude. Definitely demoted.
1:34:13
Um >> well well I'm just saying like you know dayto day I I imagine you're you're you know you still have the stress of you know running the business and and all that but uh but but maybe you're not on call which would be nice.
1:34:27
>> Nothing's nothing's changed.
1:34:27
They just don't let me write code anymore. >> Yeah.
1:34:30
>> Um which is probably better for everybody.
1:34:32
Uh but you know I'm the chief strategy officer.
1:34:34
I I run um a lot of product strategy, infrastructure strategy.
1:34:38
I work with clients on big uh big deals. Mhm.
1:34:42
>> Um, you know, probably half the company indirectly reports to me.
1:34:46
>> Um, so I'm on the board of directors, etc. , etc.
1:34:48
So, I've been I'm I've been involved in pretty much every aspect of the business.
1:34:53
>> Um, but I'm happy to talk about whatever you guys would like.
1:34:55
Today, >> I want to talk about naming every strategy.
1:34:58
I want to talk about the overarching strategy.
1:34:59
Um, walk me through how you would think about defining the current strategy.
1:35:07
what is the highest priority relationship with hyperscalers, new companies, creating a fluid compute layer, um all the different like government incentives like what what is top of mind for you uh on a day-to-day basis in terms of positioning core against competitors or just broadly solving a problem in the market like how do you actually break down these problems?
1:35:31
problems? So um it's two things right the the first is um serving the customers and partners that got us here right that's the big AI labs it's the hyperscalers it's the people that have an insatiable need for compute infrastructure >> they're the ones that are blocked
1:35:47
>> from delivering to their next customer from growing their user base because they don't have any GPUs >> um that has been um it's been insane to be a part of right is because every time it increases an order of magnitude you're like it can't go again and then it goes again 6 months later. >> Yeah. >> Yeah.
1:36:03
>> Um and you know I was listening to the show before you were talking about uh you know meta's 5 gigawatt data center campuses etc.
1:36:09
Um >> you know it's uh like we're headed there right and I think that right now it's the question of you know everybody's in a race and how quickly you can build that like that large.
1:36:19
Uh but it's also understanding what are the local constraints what are the actual grid constraints what are the supply constraints to be able to get there.
1:36:27
You know you run into things like labor constraints.
1:36:30
you know, there's not enough electricians in the world to go out and build these things and the timelines these people need.
1:36:33
How do you modularize the data center construction?
1:36:38
Um, you know, everybody's kind of solving this as the planes in the air already. >> Yeah. >> Right.
1:36:42
Um, and you know, it it's it's a race and you know, I I always say internally like I can always I can solve any problem with enough time or capital, right?
1:36:51
And a lot of the times you don't get both.
1:36:54
>> Um, so it's choosing, you know, what's the right path to go, you know, how are we going to solve this for them?
1:36:57
we going to solve this for them? Um so that's that's kind of one side right is like okay how do we build to everything that they need how do we understand what their demand's going to be and when you look at um the thing that I think the world misses is that everybody on the hypers scale side and like in our seat
1:37:12
as well like we're looking at it going we need to build x number of gigawatts of data center capacity >> right like we know the demand signals are there but the capital and the speculative capital isn't there to do it >> interesting >> right is I can go out and I can build x gigawatts of capacity for the next two or three right from my balance sheet. But any
1:37:29
But any more than that and the demand doesn't show up and I put my entire company at risk. >> Yeah. >> Right.
1:37:34
So we're we're basically having to haircut the demand signals that we get and only build to a certain extent of it and then by the time we get out there we're in the same constraint position we're in today. >> Got it.
1:37:43
Um talk to me about >> how how do you um real quick how do you think so so it was 2019 that you guys decided to focus more on AI and and cloud and for broadly is that correct roughly? >> Yeah.
1:37:57
So, um, >> and the reason the reason I the reason I asked this is because I saw a company as, you know, there's been a bunch of companies that were, uh, effectively came to the conclusion that you guys did, but maybe six, five, six years later in terms of, hey, we're going to sort of pivot away from, you know, Bitcoin mining into AI because that's where the real growth is.
1:38:19
And it feels like a lot of those company from from an outside looking in uh you know it feels like those companies are going to be at a massive disadvantage.
1:38:28
But I don't know if I have if I have the right read there. >> Yeah.
1:38:33
So we um when we started out in 2017 um we made the decision that we were never going to be as good as the ASIC producer in mining Bitcoin. >> Yeah. >> Right.
1:38:44
And the idea there was, you know, there's always going to be something they don't release.
1:38:48
It actually wound up being true.
1:38:50
Uh they had released like their low power mode, I forget what it was called in 2018 and everybody found out that they were hashing 30% better uh power efficiency.
1:38:58
The world lost their mind about it >> and it was kind of like the okay, we were actually right.
1:39:02
Like we were kind of were paranoid about it, but we never actually knew until then.
1:39:06
>> Um so, you know, we said we're not going to do it as good as them.
1:39:08
Where is there a level playing field?
1:39:09
And we thought there was a level playing field in GPUs.
1:39:13
Y >> and we also thought, okay, if crypto goes to zero, the GPUs are repurposable. They're used for gaming.
1:39:18
They're used for graphics rendering.
1:39:20
Like we could either sell them on the open market or we can build another service with them.
1:39:25
>> Um what we were most surprised by is when we first launched uh our you know let's say first non-crypto product in 2018 2019.
1:39:33
Um it was a rendering service for an open source uh uh 3D uh graphics platform called Blender. Blender.
1:39:43
>> We had like a thous a thousand people sign up in the first day. >> Wow. >> Wow. Yeah.
1:39:46
>> And I was just like, >> "Oh boy, like there's a market here. People need this stuff."
1:39:49
And like we had never expected it to be that large.
1:39:52
And people would come to us and they send us like uh support tickets saying, "Hey, just want to tell you it's so awesome.
1:39:57
I was able to do so much more work because I got access to GPUs I've never had before." >> Yeah. >> Right.
1:40:02
And that was like signal number one.
1:40:04
And we're like, "Okay, we convinced our early investors like we're going to go all in on this thing.
1:40:08
We're going to use crypto to pay for it as we're building out more uh compute."
1:40:11
And uh we set it up so that as we didn't have uh actual cloud or compute workloads, it would turn back to mining >> and the mining would cover all of our fixed and variable expenses and all of the cloud stuff was just upside. >> Yeah. >> Right.
1:40:29
So we were able to show over time this like constant penetration or kind of constant growth of our margin um which was really easy to grow. >> Right.
1:40:38
We were able to lever the cryptocurrency into a massive resource base.
1:40:41
So the timing played a lot a lot into our success there and that we had that permissionless crypto revenue to lean back on. >> Yeah.
1:40:49
but also just having having I think incredible foresight to some degree.
1:40:53
Uh which which uh you're being a little bit more modest about it.
1:40:56
But the fact that companies, you know, even I I saw >> a company uh that was getting pumped on X last week that is still in the process of transitioning from Bitcoin mining to uh you know some something that will look like a Neocloud.
1:41:12
And the idea that you're going to come into this market six years late and really be a a key player uh feels like a long shot.
1:41:22
>> It's um so I think it's even worse than that.
1:41:26
>> And it's even worse it's even worse than that because the talent pool is so small. >> Yeah.
1:41:30
>> Um and to be successful as the technology has gotten more sophisticated and harder to run um you have to be the best in the world, right?
1:41:38
And you know, we have some of our low-level hardware engineers. They're amazing.
1:41:43
Like, I've seen them um, you know, solve problems that nobody even knew existed.
1:41:48
They they identify the problem.
1:41:50
They're able to say, "Okay, this is what this is what the problem is."
1:41:53
They go back to Nvidia and they jointly develop a solution for it, right?
1:41:56
Um, that that's really important as you're building at scale, right?
1:42:00
You know, anybody can run a hundred or 500 or a thousand GPUs, but if you're running hundreds of thousands of these things, like it's really really hard to do, especially at high quality. Mhm. >> All right.
1:42:10
So, um, people think that there's folks that are 12 to 18 to 24 months behind us. Um, good luck.
1:42:19
>> Uh, really quick question.
1:42:19
Uh, I have to ask, uh, what's the story with the wheels behind you? >> Uh, all right.
1:42:25
So, um, there are F1 wind tunnel tires. >> Yeah.
1:42:29
>> Uh, and we just did a So, we're >> uh, we just did a sponsorship with Aston Martin.
1:42:35
Um, and we also and we also executed a uh compute agreement with them.
1:42:40
>> Um, I'm uh so excited about this one.
1:42:44
You know, when I talked before about like the way that we think about our first cohort of clients, it's the AI labs.
1:42:47
The second is the enterprises. >> Yep.
1:42:50
>> And I feel that if we can walk into Aston Martin and help them modern modernize their AI and ML stack and in the most hyperco competitive sport in the world with so much data and so many eyeballs already on it, we can do it for anybody. >> Yeah. Right.
1:43:02
So, so being more concrete there, you have an F1 car.
1:43:05
You need to simulate how air flows over it.
1:43:08
That's a particle simulation and that's perfect for accelerated parallelized computing GPUs.
1:43:15
They're probably doing it on premise at some point.
1:43:17
They're going to take it to the cloud.
1:43:19
They're going to take it to your cloud. >> It It's so much Yes.
1:43:21
But it's so much more than that.
1:43:24
It's not just the CFD workloads.
1:43:26
It's also things like taking radio data from other teams and trying to decipher what their strategies are in real time. >> That's awesome.
1:43:33
>> Uh I mean it's running tire degradation simulations like how does track temperature change with cloud cover?
1:43:38
How does weather come in and actually impact this stuff?
1:43:41
It's helping them make better decisions on track in real time. >> Wow.
1:43:46
>> Um it's like it's the coolest confluence of the technologies that I can imagine and I'm like I'm totally into it. >> That's amazing. Yeah.
1:43:53
translate all of Ferrari's comms from Italian into English. >> Yeah.
1:43:58
Well, I mean, good luck with Ferrari strategy.
1:43:59
So, >> talk about >> uh talk about uh sort of data center buildout timelines.
1:44:06
I think Colossus and and the XAI team were getting a lot of credit for bringing a lot of compute online very quickly, a lot of energy online.
1:44:15
Uh how much was that an outlier or are they just really good at marketing?
1:44:22
Um I I think it's both, >> right?
1:44:26
Um you know, one of the things that I admire about Elon, um is, you know, he sent an email a couple years ago on Thanksgiving and it was like the biggest problem that we have right now now is there's nobody to turn a screwdriver on the line.
1:44:39
I'm going to go turn a screwdriver in the line, right?
1:44:42
>> And he is so good at identifying what the bottleneck is and breaking through it.
1:44:46
And if you're the richest guy in the world and you don't really care about the consequences, you can do that a lot more than you can as a director of a public company is what I've learned.
1:44:54
>> Um is, you know, they're able to go and I think some of what they've done is a bit of the you told me I can't build it, but I already built it. It's already there.
1:45:04
And they're like, "No, you can't build it." No, no, it's there. >> Right.
1:45:08
>> Um and when you're in a race like this, that bravado matters. >> Yep. >> Right. Um I I don't know.
1:45:12
I don't think anybody knows the truth of what's actually there.
1:45:17
Uh but they've definitely been incredible in how they've been able to execute. >> Yeah. Yeah.
1:45:22
I mean the the that data center seems like it's performing very well, but what I learned from Dylan Patel over at semi analysis with the cluster max ranking is that not all clouds are created equal.
1:45:34
And a 100,000 H100s is not perfectly substitutable for another 100,000 H100s over here.
1:45:41
And uh coreweave did extremely well on cluster max.
1:45:48
And I'm interested in pe peeling back a layer deeper into understanding first how do you think you did so well like like what what concretely makes core stand out as a neocloud uh on a on the features level from a customer perspective and then internally what's the secret sauce like is it just great employees? Are they aligned? Are they working harder? Is there key insight?
1:46:13
are you like how do you actually deliver that product at uh at a higher level than you know competitors that Dylan Patel reviewed? >> Yeah.
1:46:22
So so let's let's not get off the XAI thing yet.
1:46:24
I'm going to trans transfer into that.
1:46:27
Um you know if they have 100,000 GPUs and they get 60,000 of them running and they don't care about the other 40,000 of them >> it's kind of enough. >> Got it. >> Right.
1:46:38
And if they're looking at it saying I don't care how much this costs. I have to win. >> Mhm.
1:46:42
and they're in the position to do that.
1:46:44
And if you're Elon Musk, you are right.
1:46:46
That's a powerful position to be in. >> Sure.
1:46:49
>> Now, the clients that we serve um are typically going to be more costconscious >> and are going to look to optimize, you know, their spend and make sure they get the whatever they need out of their investment.
1:47:00
>> Um and what we found is that, you know, these jobs >> and this infrastructure, it it breaks all the time, >> right?
1:47:07
And um when you accept that something's going to break all the time and you design for it, your solution is going to be very different than if you're just kind of um you know square peg in a round hole after the fact saying I have to go deal with these failures. >> Yep. >> Right.
1:47:22
So what we did initially is we built uh with a incredibly high level of observability at the bare metal layer. >> Right.
1:47:28
And the bare metal visibility is really important because things do fail and you know you have uh memory failures, you have thermal failures, you have all these things that impact your jobs.
1:47:38
But if we're able to help our customers identify what's happening, triage it for them and explain to them why something failed, you know, they don't have to go back and say, "Oh, was it my code or was it one of the 700,000 connections in my cluster?" >> Right?
1:47:51
They can immediately point to our software stack that says, "Oh, it was this link flap over here.
1:47:54
We took it out of service."
1:47:57
you restart your job, you'll be totally fine.
1:48:00
>> Um, that's been very, very, very powerful.
1:48:01
And it's not just that visibility layer, it's also the storage and networking um, and the quality control that we put around those things.
1:48:11
Um, you know, I have to give credit to our new C our now CTO, Peter Sanki, who used to work for me until they found me out.
1:48:17
Um he is uh he is incredibly focused on the low-level visibility and quality and drives a lot of that culture throughout the company, right?
1:48:28
Of like >> if one of 500,000 things is going to break, I need to know immediately so I can go in and fix it to make sure my customers are back up and running like 3 minutes later.
1:48:39
>> Um so it's it's building from that low base and understanding what the problems are going to be.
1:48:44
Uh and then culturally, you know, we have people that are here that take so much pride in being first and being best. >> Yeah. >> Right.
1:48:55
And you know, if we're not first to market with a new release, like they're like they're going to cry for a week kind of thing.
1:49:01
>> Um so when we did GB300 a couple weeks ago, or maybe it was just a week ago, like people went out to one of our data centers, like all of our low-level engineers went out there and they lived there for like two weeks to get it done. >> Mhm. >> Right.
1:49:12
like they go and they they work from 8 am to 4:00 a. m.
1:49:13
every day and they love it and it's a cultural thing.
1:49:18
Um, and I think we're really lucky to have that. >> Uh, bit of a tangent.
1:49:22
Were you an H20 buyer at all over the last few months or or not needed? >> No.
1:49:31
>> And and there there was H20 sales happening in the United States from what we heard.
1:49:36
Who are the kind of uh players that are that were were making use of that?
1:49:41
I don't know um if that did happen. I don't know who it is.
1:49:45
>> Um but for us, you know, if we had any power available, we're going to buy the best, you know, the best accelerator we can.
1:49:50
Um you know, there's no reason for us to buy the cut down accelerator if we have power available.
1:49:56
>> Yeah, that makes sense.
1:49:56
You you mentioned that one of the key things is like the insight into what's happening in the data center.
1:50:02
what uh if a specific chip is failing, you want your customer to be able to know.
1:50:07
You want to be able to I mean, we talked to Dylan Patelli who was saying like sometimes you just need to turn it off, turn it back on, or like there's a whole bunch of different things that can happen uh in that feel um very abstract when you're looking at it from the outside, but are very concrete when you're on the inside.
1:50:24
And that's actually what separates great data center performance is just like the handtohand combat that happens on an inside day.
1:50:30
Um my question is like um we've we've talked to some people who are like we're going to put data centers in space or we're going to put data centers like on the moon or >> doesn't seem to align with it with your framework of being like we need to build it in in a way knowing that it's going to break often and we're going to need to fix it and hard to do that if it's it's >> yeah I don't think space is the right place for a data center. Okay.
1:50:53
>> Um but you know to to the point there of sometimes you have to turn it off and turn it back on to solve the problem.
1:50:56
Um we don't find that to be true. >> Okay.
1:50:59
break down is if you if something needs to be turned off, there's a reason why it needs to be turned off. >> Sure.
1:51:05
>> And that like that one comment may be what separates us >> is, you know, our engineering team if they're like, I'm not going to reboot the system.
1:51:12
That doesn't solve problems.
1:51:14
I want to know exactly what happened there because we have to go solve it because if it happens one time, it's going to happen another time.
1:51:18
Like the scale that we run at these like these skeletons come out. >> Yep. >> Right.
1:51:24
Um and you know being making making sure that the data center is on Earth and is acceptable by is accessible by like human technicians is really important. >> Yep.
1:51:32
>> Um because sometimes you have to make judgment calls on like what's actually causing this and um you know how do you fix it? >> Yeah.
1:51:38
Um so the the one kind of I mean the there's a few things that the space data center folks will tell you about like you know free energy or free cooling or something like that.
1:51:49
We we we don't need to go too deep into that.
1:51:51
I guess the question that I'm more interested in is uh how are the recent results from different training runs?
1:51:59
Uh we saw this with Grock 4 that the reinforcement learning post training was 50% of the overall cost.
1:52:06
That's kind of the rumor and and there's been kind of mutterings about hey maybe in the future uh you'll need a whole bunch of distributed compute all over the place generating reinforcement data and doing rollouts and bringing those back and you won't need as much of this hyperconentrated compute in one place.
1:52:29
And so maybe having 10 1 gigawatt data centers is better than one 5 gawatt data center.
1:52:36
It's just power and compute all over the place.
1:52:39
How do you think like the the shape of distribution of uh of compute clusters will evolve over the next few years if you have any insight there?
1:52:50
>> Yeah, so great question.
1:52:50
Um let's let's put it into perspective because you said one 5 gawatt or 10 one gigawatts those are insane by the way like those are huge. >> Okay.
1:52:59
>> Um like even building a gigawatt data center and its impact on the local grid like that's we'll see.
1:53:02
I mean, you talk to people and they're like and they're like, "There's going to be 100 gigawatts data centers in three years."
1:53:08
Like the the crazy AI people get crazy and so I I'm maybe I'm a little bit on, you know, like out in space here literally.
1:53:14
But but bringing back time to Earth. >> Yeah.
1:53:18
So, um people definitely want to build at that scale, >> right?
1:53:23
When you think about 100 gawatt data center or a 10 gigawatt data center, Yeah.
1:53:27
>> the capital required to build these things and to install the compute inside them, like that's country level. Yeah. >> Right.
1:53:34
And you know that I don't think that um the broader market understands the investment required yet to do this. >> Mh. >> Right.
1:53:44
Um but you know back to the question of is it going to be more distributed?
1:53:48
Is it going to be more centralized?
1:53:49
I think that right now while the ability to centralize it exists people would prefer to do that because the optionality of having it all together is you know is worth more than having it distributed. Mhm.
1:54:00
>> Um I think that as you have workloads that become more latency driven, >> right?
1:54:05
So, you know, we're starting to see customers that say, "Hey, I need something that's actually local to a metropolitan area because I'm running >> um you know, aentic AI for one of my customers or for a banking customer.
1:54:14
Um I I need to make sure it's there."
1:54:18
That's a very different story than 12 months ago where people didn't care where inference was because your first response wasn't for like two or three seconds. >> Yeah.
1:54:24
You actually want it on the other side of the world ideally where it's low load. >> Yeah.
1:54:28
then then you're managing to optimize for cost. >> Yeah. >> Right.
1:54:31
But there there's a group of people that are going to optimize for cost.
1:54:34
There's a group of people that are going to optimize for performance.
1:54:35
Um you know, for us, one of the products that we're really excited to develop is um you know, almost like a global load balancer product that helps people choose what they're optimizing for, right?
1:54:47
Um and you know, we've been thinking about that for a long time now.
1:54:52
Uh because like you said, you have to have that compute distributed and it may not be be because you're trying like if you could choose you'd always have it in one place, >> right?
1:55:01
But when you build it distributed, then you have to have the software services that can intelligently say, okay, where am I going to run this?
1:55:08
What am I optimizing for?
1:55:08
Uh where's the data that I need? Right?
1:55:10
And that's where um you know, we've been investing in, you know, we've built a massive global uh network backbone.
1:55:16
We have a ton of our own dark fiber um to be able to move workloads around and have >> What does that mean?
1:55:22
What does that what does that exactly mean? Your own >> dark.
1:55:25
So uh we've leased dark fiber between a bunch of our data centers across the United States and in Europe that allows us to install our own optimal uh optical optical gear um so that we can you know flex up bandwidth as we need to.
1:55:38
Uh you know we have some customers that need things like 64 terabs per second between sites like crazy amounts of bandwidth.
1:55:45
And the idea there is that they're moving synchronization data.
1:55:48
They're synchronizing models between sites. >> Yep.
1:55:52
>> Um so you know that's where you know we have to make these big capital investments knowing what's coming right.
1:55:59
So when our customers start asking for it was like oh yes we already have it there.
1:56:01
Um so you know some of that is just it's the focus that we have and the specialization to know hey like we better start investing in this because a year down the road like we're going to be screwed if we don't. >> Yeah. Yeah.
1:56:13
How how has it uh been, you know, what what are some of the differences been in building out data centers in Europe?
1:56:19
I know you guys have Norway, Sweden, Spain, and the UK. Is that correct?
1:56:25
>> Uh what what's it been like developing uh new data centers there versus uh what you've done in the United States?
1:56:33
>> Um not so different actually, >> right?
1:56:37
Uh you know, everybody was initially saying, "Oh, it's going to be so hard to work in Europe.
1:56:39
uh you know are people going to show up and work as hard?
1:56:43
And um the one of the things that we've done I think that we've done really well in our onboarding process for employees is we've actually brought all the European teams over to work with our tiger teams inside of our data centers.
1:56:55
>> And >> you just blare the Star Spangled Banner just loudly if you're just if you blast music continuously all of a sudden like they have Yeah. moment.
1:57:02
Um, and you know the the culture is infectious and the what's one of the best things about the company is that all of our like the company's so young that there's been no real attrition, right? Nobody's left. >> Yeah.
1:57:16
>> So the people that are running the data center start as started as data center tech a couple years ago.
1:57:19
Like everybody's still fired up.
1:57:20
They still have like this the same drive. They want to be first. They want to be best.
1:57:24
And when you have new employees that walk into that environment and those people are genuinely happy to teach them, >> um, it transfers pretty well, >> right?
1:57:33
So, I think our European investments have gone a lot better than I thought they were going to.
1:57:36
Um, which I attribute largely to our people.
1:57:43
>> A couple years ago, I feel like everyone in tech had to learn about Nvidia and then everyone in tech had to learn about TSMC and then everyone in tech had to learn about ASML and now people are kind of getting familiar with SKHEX.
1:57:52
Uh what do you think the next big company in the semiconductor stack or data center stack or even networking like what is the next big theme or or subcategory of important foundational technology?
1:58:10
foundational technology? uh people are starting to dig into the the rare earth materials and what's going on like what is on the horizon in terms of like underexplored or or misunderstood as important as having a role as we try and build out these 1 gawatt 10 gawatt
1:58:29
data centers what are we going to be hearing about >> that's such a such a good question and the kind of question man I don't even know um I'm not sure I have an answer for you >> yeah you may maybe it's networking gear I mean, you mentioned you're putting optical gear at uh and and just hearing what was it 42 terabs a second. Like
1:58:45
Like that seems like something that could be become very important.
1:58:51
>> Not that you have to call out a single company, just like the the the the overall industry is interesting to be like, oh, you know, everyone's going to be thinking about this industry.
1:59:01
>> Yeah, you know, that's definitely one of them.
1:59:03
Um fiber capacity between metros is going to be an issue.
1:59:06
Um I don't know that it's uh an issue that's insurmountable. >> Yeah.
1:59:13
>> Um there are definitely some gotchas that I can see coming down the the line at people.
1:59:19
>> Um I I don't want to say it because then I'll give them I'll give them warning and since we've already identified what they are and I feel like I'm in a better position. >> Sorry. Yeah.
1:59:26
I don't mean to put you on the spot too much. >> No, it's it's okay.
1:59:29
>> What about electrical What about electrical transformers?
1:59:30
Um I've heard a little bit about like there there sometimes there's a shortage.
1:59:35
they're very complicated.
1:59:36
Like it's this is this like big technical thing that's big piece of hardware.
1:59:41
Uh is that like an important uh thing for people to be aware of going forward or like an important industry to understand?
1:59:49
>> So um the a bit of a tangent off of that, right?
1:59:53
Um everybody talks about how there's no power left in America. >> Yeah. >> Right.
1:59:58
And if you actually look at the data, there's a tremendous amount of power available from base load and load following generation. >> That's great to hear. Thank you.
2:00:06
>> Um, and the the problem exists on peak days. >> Mhm. >> Right.
2:00:10
And like what solves peak load problems is solar.
2:00:14
>> Like solar and batteries will solve peak load problems.
2:00:15
Like you're worried about, you know, 5:00 on the hottest July day when everybody gets home to turn their air conditioning on. >> Sure.
2:00:22
>> Like that's that's the thing.
2:00:24
>> Um, and you know, everybody wants to talk about how uh they want to use demand response as a way to, you know, bring supply back to the power grid from data centers.
2:00:32
Um, you know, this stuff is really hard to bring up and bring down.
2:00:37
It's it's not really intermittent load, >> right?
2:00:40
Um, you know, I think that a lot of this shortage narrative from a power perspective is coming from the fact that everybody and their brother has become a data center developer in the past year.
2:00:51
And there's a lot of people out there squatting on power, right?
2:00:53
Or squatting on power rates.
2:00:55
And I don't think that the uh local regulatory commissions understood what they were signing up for when they started doing load studies for free and giving power allocations.
2:01:04
>> Um you know but it goes back to my point earlier that the biggest constraint is speculative capital to build out >> to meet demand that we expect.
2:01:13
>> Interesting, >> right?
2:01:14
So like yeah transformers are a problem but they're not outside the lead time window.
2:01:19
>> Um you know UPS is everything's a problem if you need it tomorrow. >> Sure.
2:01:22
But if you're able to plan and make the investments in like in a reasonable period of time and you trust your demand forecasts >> and the market is there and believes in it um like this is all fine. >> Yeah. >> Right.
2:01:34
>> Do you still uh I feel like the chart that people have been tracking is like China bringing on you know you know basically like copy and pasting um nuclear reactors.
2:01:43
Do you think that the US should be seriously, you know, there there seems to be enough people uh now from tech to Washington that care about nuclear?
2:01:55
So hope >> it used to be very contrarian.
2:01:56
It was very very edgy to say, oh yeah, nuclear it's like it's it's safe and we should do it and people it's it's still scary.
2:02:04
>> But yeah, in your view, would you like to see more focus on more attention on nuclear or more attention on on solar if you had to pick?
2:02:12
>> Uh I mean so nuclear they solve two different problems. >> Mhm. >> Right.
2:02:15
nuclear builds your base load and your and like it's not nuclear is not load following.
2:02:18
Um it's really hard to bring those things up and down.
2:02:20
They have long outage cycles.
2:02:23
>> Um there it's a base load solution, right?
2:02:25
So as your data center demand is going to increase so much and you're shifting your base load requirements up, that's the solution.
2:02:29
Particularly if you want to retire old coal or you want to retire some of the base load natural gas, um nuclear I believe is the solution there. Mhm.
2:02:39
>> Um but as you're dealing with peak loads, right, you're looking to match that load profile and you know, solar is a great one to do it. >> Yeah.
2:02:46
>> Um you know, wind doesn't necessarily solve that problem during peaking.
2:02:48
Like wind doesn't peak during the afternoon, it peaks overnight. >> Sure.
2:02:52
>> Um so you just have to match like what's the load profile you're trying to match >> and which part of the curve are you trying to solve for?
2:02:58
Is the load profile for uh AI not perfectly matched with nuclear and and is that because of uh inference demand from actual customers?
2:03:09
Like I've heard people use chat GPT during the work week, but they don't use it when they're sleeping.
2:03:13
So there's actually maybe it follows more of like the solar load profile because when I think of training, I think of we're going to run this entire data center for months.
2:03:21
I don't know how far off I am on that, but like that feels like perfect.
2:03:25
Oh, build a nuclear power plant right next to the big data center.
2:03:27
run it continuously, just keep training, and then when you get done with one model, train the next one and just keep it running perfectly matched forever.
2:03:32
Um, but but h how are we actually like like the actual AI workloads, how how do they match to different load profiles? >> Yeah.
2:03:42
So, if you're in a global constraint environment, >> right, you're serving as much load as you can around the clock from anywhere. >> Yeah. Yeah. >> Right.
2:03:50
So, I I don't think that you have any load shape right now.
2:03:51
I think it's just flat load and it's like it's pinned to 100%. Got it. um as >> bullish. >> Yeah. Yeah. It's insane.
2:04:00
>> You're like, "It's stressful. Stop talking."
2:04:02
>> It's it's incredibly stressful, but it's it's a good problem to have. >> Yeah. Yeah.
2:04:06
>> Um I have people yelling at me every day like, "Where's my stuff? Where's my stuff? Where's my stuff?"
2:04:10
And I'm like, "I promise I'm doing everything I can."
2:04:13
>> Um but I I think that as that like as you have the investments where things are more latency sensitive, um you'll see that the load curve may not change.
2:04:23
It may still be pinned at 100, right?
2:04:25
But the types of workloads may be more may be more shaped.
2:04:27
Um I'm not sure how it plays out. >> Right.
2:04:31
I I think that you know the the problem that you're trying to solve there is not necessarily data center load shape.
2:04:37
It's residential and commercial load shape. >> Mhm. >> Right.
2:04:41
Because you're not running like your big commercial buildings in New York aren't running at 3:00 a. m. They're running at 3 p. m. That's the problem. Mhm. >> Makes a ton of sense.
2:04:53
Well, this was really fun.
2:04:55
I hope uh that we see you at the track soon.
2:04:57
One of our other uh one of our sponsors is also an Aston F1 uh sponsor, so hopefully we run into you uh on the paddic.
2:05:07
>> I I'll be there soon actually. >> Fantastic. >> Amazing. Uh great having you on.
2:05:09
Let's make this a regular thing when you guys have news.
2:05:12
Uh it was uh awesome to get all your insights. >> Yeah, this is great. >> Yep. Thanks, guys. Thank you so much. >> Cheers. >> Bye. >> Bye.
2:05:19
And since we're mentioning public.
2:05:20
com, investing for those who take it seriously, multiasset investing, industryleading yields, they're trusted by millions.
2:05:26
Uh, our next guest will be joining us in just a few minutes.
2:05:28
But first, we can tell you about Wander. Find your happy place. Find your happy place.
2:05:36
>> Book a wander with inspiring views, hotel grade amenities, dreamy beds, top, and 247 concier service.
2:05:40
It's a vacation home, but better, folks.
2:05:44
Okay, I got an update from uh a friend uh friend of a friend >> uh which says that uh his friend No, no, no, sorry.
2:05:54
Our friend says that he has a buddy that owns some RM boutiques >> and has not heard of the conspiracy.
2:06:04
>> Really, is interesting.
2:06:04
>> Maybe it was it might have been hallucinated. I don't know.
2:06:07
I need to find a source for that.
2:06:08
Um where where was that actually in here? I will figure it out.
2:06:13
Uh, and we will get to the bottom of the conspiracy theory.
2:06:15
Um, but we will do that later because our next guest is here. It's Tyler Cowan.
2:06:19
How are you doing, Tyler?
2:06:22
>> Which conspiracy theory?
2:06:23
>> So, this conspiracy theory is that you'll actually like this one. >> Okay.
2:06:26
So, >> uh, are you familiar with the high uh, the ultra luxury watch brand Rishard Mill? >> Only by name. >> Okay. >> I don't have one.
2:06:36
>> You may have seen it on a on a sailboat or something like that. >> Yes.
2:06:39
So, Reishard Mill is famous for selling watches at 10 times the price of other luxury watch brands.
2:06:45
So, a Rishard Mill, the first one was $135,000 for a single watch.
2:06:51
And at the time, $13,000 would have been more appropriate for that kind of tier of watch.
2:06:56
And the conspiracy theory is that when Rishard Mill went to promote their new watch, they took out an advertisement in the Financial Times.
2:07:06
They sent over the copy for the advertisement saying, "Please uh please promote Rishard Mill.
2:07:12
Our new watch is is available for purchase."
2:07:14
And the price is $13,500.
2:07:19
And there was a mistake in the newsroom and they added an extra zero and they printed $135,000.
2:07:27
And what happened was instead of getting no offers, they got flooded with offers because people had to have it.
2:07:33
a classic example of a vellin good.
2:07:35
Um, and uh, and the rest is history and to this day the brand denies it.
2:07:40
Uh, but what do you think?
2:07:43
>> Well, it sounds like the opposite of a conspiracy, just an innocent mistake and then it worked out well for the company and bad for the buyers. >> Yes, I believe it.
2:07:50
the the the conspiracy is that the brand now obviously denies it because it would it would not be a good look for them now to to be like, "Well, we actually wanted to price it at 13,000 and then we actually, you know, >> uh now that you're willing to pay 135, well, good deal. Let's do it."
2:08:06
Um, but >> there'll be a deathbed confession on that one.
2:08:11
One way or the other, >> maybe.
2:08:13
>> Well, it's great to have you here.
2:08:13
It's great to have you on video.
2:08:15
Uh, super intelligence is >> Yes. Yes. Yeah. Yeah. Yeah.
2:08:19
We had some technical difficulties last time, but I'm glad. Uh you look great. How are you overall?
2:08:24
Let me take your temperature on uh the last time we talked, you said AGI was around the corner.
2:08:28
Later, two days later, uh 03 released. We now have 03 Pro.
2:08:33
How are you feeling on artificial intelligence progress?
2:08:36
Um overall, do you feel like we're accelerating, decelerating?
2:08:40
Are you happy with the products? How are you using them?
2:08:42
Take me through kind of the state of the union.
2:08:46
Everyone has a different definition of AGI, but the definition I grew up with, which is that it can pass a touring test and do at least as well as expert humans on most endeavors, intellectual endeavors, it clearly passes.
2:09:00
>> So, if you want to call it AGI, you can.
2:09:03
I don't insist on the term.
2:09:03
Maybe it's more misleading than useful.
2:09:07
>> But 03 03 Pro, I use them every day.
2:09:07
I ask them questions about history or music or travel.
2:09:12
My wife and I did a trip through Europe, France, and Spain.
2:09:16
We asked it about our hotels, our restaurants, the churches. We saw everything.
2:09:20
It just keeps on working. It's amazing.
2:09:22
Much better than having, you know, a historian guide with you.
2:09:27
So, to me, it's like AGI. It's better in fact.
2:09:30
Are you What is your prior for hallucination at this point?
2:09:34
because that that Rishard Mill conspiracy theory that I just um relayed to you that comes from an from a chat GPT 03 pro deep research report and uh we're getting text messages that many people hadn't heard of it.
2:09:51
I haven't checked the sourcing yet.
2:09:53
Um so I'm not sure if it's real.
2:09:55
I haven't gone down and done all the factecking.
2:09:57
But just in terms of your prior, how likely would you believe that it's uh that it's a a hallucination?
2:10:06
Well, just ask it, right?
2:10:06
If it's something important, you double check with another AI andor you ask and you find out.
2:10:12
So, hallucinations should not be a practical problem.
2:10:15
They end up one because people are silly >> or they'll ask it questions where there may not be any answer at all.
2:10:20
And that is by far when it's most likely to hallucinate.
2:10:25
They'll just make something up to please you.
2:10:26
But if you know in advance, that's when it's most likely to hallucinate.
2:10:30
Again, you take counter measures.
2:10:32
So, it's a way overrated problem. >> I agree. I agree.
2:10:36
And I just fact checked this and it does come from a 2021 GQ report, which again might be wrong.
2:10:44
>> Hallucinated by the >> GQ never hallucinates, right? >> Take it out.
2:10:48
Gentlemen, do not hallucinate.
2:10:50
>> No, gentlemen do not do not hallucinate.
2:10:53
Um, what do you think of uh Dwaresh Patel's new formulation that there's sort of a a two axis diagram between you and him? Uh, AGI is here. AGI is not here.
2:11:03
The impact of AGI economically will be massive in his in his mind.
2:11:05
Uh and the impact and the impact of AGI will be uh more muted in in his assessment of your assessment.
2:11:14
assessment of your assessment. And I think I I think your general stance which is uh we should hold ourselves accountable to the definition that we set out to achieve you know to achieve you know decades ago when this work started and not constantly just be you
2:11:30
know pushing out uh the the sort of requirements the goal moving the goalposts uh is is important >> and then there's this new definition super intelligence >> yes >> I'm waiting for super duper intelligence right >> giga giga intelligence >> ultra intelligence we already we're already past it. Everyone, folks like
2:11:45
Everyone, folks like us, we know what super intelligence is.
2:11:50
We we we already have it.
2:11:50
We're waiting for the next one.
2:11:52
>> Maybe the best measure is how much of people's time are they spending with this thing and that's rising steadily, right?
2:11:58
So that to me is quite impressive.
2:12:00
Darkeesh understands my views perfectly well.
2:12:04
>> He and I use the term AGI in different ways, so it all ends up being confusing.
2:12:09
I just think the human bottlenecks are so significant that it's not going to be fast progress anytime soon. That's my core view.
2:12:15
It's not based on any kind of bearishness about AI.
2:12:17
It's that I've spent my whole life working in human institutions.
2:12:21
And there are plenty of much simpler developments like don't be a jerk that would boost productivity immensely, but we still haven't gotten around to really implementing. >> Yeah.
2:12:33
Uh, but I mean on the on the topic of super intelligence, I think what people are kind of getting at is that it feels like we have 15minute AGI.
2:12:42
It feels like we have spiky intelligence.
2:12:46
It feels like we have intelligence that can, you know, memorize every book in the English language and yet not tell you a funny joke or or put a novel connection together between two disperate disciplines.
2:12:57
Well, and and potentially more important is this idea of of intelligence with amnesia, right?
2:13:03
It's like uh you know, we've been talking about this on the show and Darkeesh uh talked about it last week, which is this idea of PhD level intelligence that you know can't remember what you told it uh you know yesterday, right?
2:13:15
And so it's difficult to sort of build real momentum and real uh agency.
2:13:22
>> I think it can do all those things.
2:13:22
Now, it might require a little smidgen of work from you.
2:13:27
Like you might need, for one thing, you can turn on memory.
2:13:31
That's maybe not as much memory as you'd like, but you can just reinsert the dialogue you had with it >> and tell it to start from there.
2:13:36
So, okay, it's a slight inconvenience, but my goodness, people, they're so unhappy.
2:13:44
Like, it's incredible what we have.
2:13:44
And with modest effort from you, it does all these things.
2:13:50
>> So, I completely agree.
2:13:50
Uh I I love these systems.
2:13:52
I I enjoy spending time with them even if they're not directly providing economic value.
2:13:56
Uh I find them entertaining.
2:13:59
I find them just interesting. I I learn things. It's fantastic.
2:14:02
fantastic. Um I guess the the the the question is like yes you can stuff the context window but increasingly folks are saying that uh there's some sort of fundamental limitation with the size of the context window that yes you can do
2:14:16
needle in the haststack analysis with a big context window but even if you have you know pages and pages of all your best practices the current systems are more likely to forget some hard one lesson early in working with you than just an average employee. And that feels
2:14:33
And that feels uniquely unhuman. >> I don't know.
2:14:39
I've worked with a lot of employees.
2:14:40
This is one generous way I would put it.
2:14:43
The things are not perfect.
2:14:45
If you compare them to smart humans on intellectual tasks, they're basically ahead.
2:14:50
They can't dribble a basketball.
2:14:52
They cannot do most jobs. That's for sure.
2:14:55
>> It's a big reason why progress will be slow.
2:14:57
If nothing else, those jobs intersect with the physical world. Robots are far away. And there you go.
2:15:04
Um, I think it's a pretty clear prognosis.
2:15:07
I don't really get why anyone disagrees with it.
2:15:11
>> Well, if you need to raise a billion dollars, it's not super convenient truth.
2:15:15
If robots are far away, for example, >> specifically economic growth is far away.
2:15:20
This is the Sachin Sachin Nadella uh uh position which I think he agrees with you on the economic impact that we're currently seeing and potentially the economic impact for the next few years.
2:15:31
And so he is underwriting a very different investment strategy in artificial intelligence.
2:15:37
>> Yeah, it it says a lot in in our view that that that everywhere you look there's just blaring top signals in in the market.
2:15:43
the market. both like literal all-time highs uh but then just other more more indirect signals >> and at the same time Satia is like you know doing layoffs being seemingly you know very pragmatic >> uh and and and I read that as more aligned with your view which is
2:16:01
transformative impactful incredible wonderful m magic but also potentially incremental >> and um you know going to be transformative in the fullness of time but not next quarter Forgive me for not sounding like a complete capitalist, but this is all so important. People should just be doing
2:16:19
People should just be doing it.
2:16:22
>> And to some extent they are.
2:16:22
The labs are working very hard to make progress.
2:16:27
You can debate how much the motive is profit.
2:16:29
But look, these are driven people.
2:16:31
They they get how important it is and they're doing their best. I think that's amazing. >> Yeah.
2:16:37
>> Are they all going to get rich? I'm not predicting that. I don't know. >> Yeah. Yeah.
2:16:40
I I think the the tradeoff that people are debating broadly in the market is if if the next training run that's a little bit of a stretch for the market.
2:16:53
market. We were just talking to the f co-founder of coreweave and he was saying that like the capital markets just aren't really ready to absorb what it means to 10x data center construction and at the same time if AI researchers are saying that the the that by 10xing
2:17:09
data center construction we will lift ourself off of the GDP baseline and we will see breakout economic growth and we will get through some of the problems that you've identified about actually you know restructuring ing the economy and speeding up economic growth. All of
2:17:24
All of a sudden like that question and that that how much should humanity as a whole invest as a portion of GDP that becomes a very important question for the overall market to solve.
2:17:38
>> Well, let's go talk to United Arab Emirates.
2:17:40
They can either buy some more parts of sports teams or do this.
2:17:43
I think they're going to do this.
2:17:47
>> The growth rate, I think, will go up.
2:17:49
I've offered the very speculative estimate of half a percentage point per quarter, >> but that's an enormous sum of money. >> That's massive.
2:17:56
>> Now, you know, is it captured by the capital markets, the shareholders and so on? No.
2:18:00
Just the free time liberated by Chat GPT and the other services already is enormous money.
2:18:05
So, remember Satia had that challenge.
2:18:08
When will it earn $und00 million? >> Mhm.
2:18:11
>> Well, it has many times over.
2:18:11
It just hasn't earned it for him.
2:18:13
So, people get their work done more quickly.
2:18:15
You add up those valuations of time, I'd love to see a study in the number, but it's way way above a hundred million right now.
2:18:24
>> Well, on that point, there was a study last week about does uh does uh AI code generation tools?
2:18:32
Do these tools actually improve developer productivity?
2:18:37
And they did a I believe a double-blinded trial with or or maybe just like a split up trial uh based on talented software engineers working on open source projects kind of trying to work on some of the harder problems in software engineering not just build me a new web page um actually advancing open source projects and they gave half of
2:18:58
the developers uh access to tools like cursor and wind surf and cloud code and and then the other half didn't have these uh AI driven in uh development environments and they said that the AI enabled engineers expected a 20% speed up but they in fact experienced a 20% decrease or 19% decrease in productivity. What is your take on that?
2:19:18
What is your take on that?
2:19:21
Do you think that uh there's a world where we're fooling ourselves into believing that we're actually sped up by these tools or do you think that that's completely ridiculous and it's obvious that this that these tools speed us up? >> That's BS that paper.
2:19:33
first only 16 data points.
2:19:37
Second, only one of those people had experience working with AI already and that person's productivity went up.
2:19:43
It's the other fools who were just trying to get a handle on it at all where the productivity went down.
2:19:47
So, we need to throw that out, get it out of our minds.
2:19:50
Okay, it's obvious from marketplace tests that AI speeds up developer productivity.
2:19:55
Of course, you have to do it right.
2:19:57
You need to learn some things, but it's working.
2:20:01
Overwhelming real world evidence to that effect.
2:20:04
Do you think that there are particular tasks maybe that you've encountered where you've gone to AI and then afterwards realized like well for this particular task not just planning a a vacation or trip or something I was actually slowed down.
2:20:22
>> Sometimes I like giving stupid dogmatic answers. I'll just say no.
2:20:24
I think it's a correct stupid dogmatic answer.
2:20:26
No, it's never happened to me.
2:20:28
I love that because it's definitely happened to me when I've when I've said like generate a generate a a 2 by two chart and I'm like seven prompts deep and I'm like I could have done this in Excel or I could have done this and just taken a screenshot or I could have done this in Photoshop faster.
2:20:42
Uh it it it does it does happen to me.
2:20:45
But I agree on net I I feel sped up and I and I feel very happy to use the tools and I'm I'm enjoying them.
2:20:52
>> I use it to prepare for podcasts. Sure.
2:20:54
>> It cuts my prep time in half. >> Sure. >> No real doubt there.
2:20:58
>> I don't use it for charts. I this tell a human. Maybe the human uses it. I don't even know. I get the chart back. It's great. I love the world. >> I love it.
2:21:06
Uh what is the uh what is the labor economist in you think about the AI talent wars?
2:21:12
Uh has have AI researchers been historically underpaid?
2:21:18
Are we in some sort of odd intellectual property theft ring where we're paying for trade secrets to trans to change hands?
2:21:26
What is your read on kind of all the talent moves that we've seen at staggering numbers by comparison in what's historically happened in Silicon Valley? >> Yeah.
2:21:34
And then specifically, do you think that uh an AI researcher in the year 2030 will be getting a hund00 million signing bonus or do you believe that things will maybe normalize and and uh researchers will go back to being paid well but maybe not more than Fortune uh or MAG7 CEOs? >> Probably normalized.
2:21:55
Now, I'm an NBA basketball fan and what I observe is the top teams need some time to gel.
2:22:02
>> So, you look at the Miami Heat, they got LeBron, Dwayne Wade, Chris Bosch, the first year they lost.
2:22:08
>> You look at, you know, Chicago Bulls, Jordan and Pippen, it took them a few years.
2:22:12
You look at the Lakers, what they had Carl Malone, Gary Payton, was it Shaq and Kobe Bryant? They didn't win a title.
2:22:19
So, it's not as easy as you think.
2:22:22
Not that those aren't great great players, but there's something about how it gels and the process and what part of people's trajectories you get them at that all has to fit together.
2:22:32
I think it's a fascinating experiment.
2:22:35
Uh I'm dying to find out how it's going to work.
2:22:41
>> What do you think of the missionary versus mercenary debate?
2:22:43
uh a prior guest on the show uh highlighted Tom Brady in the sports world taking a lower salary in order to build a better team around him.
2:22:55
Bit of an economic sacrifice for greater performance, longer legacy, ultimately reap the rewards economically.
2:23:02
But uh a lot of people have been saying uh you know you can't buy great missionary talent that will stick with you during the hard times and really advance the frontier of artificial intelligence.
2:23:12
In the case of Brady, you know, he has these contracts for shoes, jerseys, footballs, other things.
2:23:18
He probably made a lot of money from that pay cut by staying more focal.
2:23:23
>> So maybe for some of these AI companies, you know, we need more endorsements, a line of tennis balls or something.
2:23:27
Uh, and then it'll be cheaper to hire talent because the outside market will pay them.
2:23:33
We're not at that point yet.
2:23:33
Uh, maybe we'll get there.
2:23:36
>> I think >> Yeah, I want a gameworn jersey.
2:23:40
Here's what Illyu wore when he, you know, built the key parts of GPT4. >> Exactly.
2:23:45
Um, I mean, we've, we've been putting out these little trading card sports, uh, you know, when when any an AI research, whenever an AI researcher moves from one firm to the other, and uh, I mean, it seems like an insider joke, but there these posts are getting millions and millions of views.
2:24:02
millions and millions of views. like it really is turning into into sports in terms of the attention of the entire tech industry on these trade deals and maybe it's just the big numbers but I think it's also the fact that this is an important technology and people are uh excited about the developments and they
2:24:17
and they want to know who who will win what what product will I be using in five years will it be open AI will it be meta who knows >> all of the above I hope >> probably >> how did you uh how would you try to value the Dutch East India company and the reason I bring that up is um uh uh Nvidia crossed a $4 trillion market cap. Everybody's calling it, you know, one of
2:24:40
Everybody's calling it, you know, one of the most valuable companies ever.
2:24:41
But I think a lot of the people that run the kind of valuation analysis and inflation calculations and things like that believe that the Dutch East India Company was worth somewhere around, you know, 2x of that.
2:24:54
Uh have you spent much time um thinking about uh about that business?
2:25:01
I look at all the numbers in general.
2:25:04
I don't feel I can outg guess the markets.
2:25:06
So the market is a better estimator there than I am.
2:25:09
But look, markets are always wrong.
2:25:11
That's why the price changes all the time.
2:25:12
So the fact that Nvidia is priced so high shows the market is paying attention to AI, which is good.
2:25:20
But the fact that everything else is priced more or less normally, I think is evidence for my thesis of slow takeoff, that it will matter a lot over time, but not that much right away.
2:25:30
>> Yeah, that makes sense.
2:25:30
Um, >> it is interesting something we we just had one of the founders of Core Wee on which is now a $70 billion public company and I was thinking about the advantage that he has as a public company where he doesn't have to raise capital from the private markets that are and and tell the story this like 10-year story.
2:25:48
he can kind of focus on the next quarter and just say like how do we deliver more compute to our customers?
2:25:56
And it's like this interesting change where a lot of the companies today, let's say you're a lab with a $10 billion valuation uh and you need to raise billions of dollars, you kind of need to tell the story around super intelligence.
2:26:08
Meanwhile, if you can just get out in the public markets and have billions of dollars of revenue, you can kind of go back to telling the story of, "Yeah, we're we're hoping to to bring more compute online, more data centers online, uh, and just actually focus on uh on the business."
2:26:24
>> Yeah, that can cut either way.
2:26:24
But look, if you're selling socks today, public market is fine.
2:26:29
But as you well know, fewer and fewer companies want to go that route.
2:26:33
So that does suggest to me the public markets are somewhat broken, overregulated, too much disclosure required, too much bureaucratization.
2:26:41
Shareholders themselves can be kind of nutty.
2:26:44
>> So, you know, it's good to have the two compete, but I'd like to see us make the public markets easier to be in. >> What about Delaware?
2:26:50
We saw Andre Horowitz moving to Nevada.
2:26:52
Um, we've seen some other high-profile exits.
2:26:55
Do you think Delaware will uh try to turn it around?
2:26:59
I know about a pretty meaningful amount of their uh their budget uh comes from uh franchise taxes, things like that.
2:27:08
>> I say get out, send them a message.
2:27:10
They're not feeling enough pain yet.
2:27:12
Let's make it really tough for poor little Delaware. What do they call it? Small wonder.
2:27:15
Well, small wonder that they haven't gotten their regulatory apparatus into better order.
2:27:22
Wow, that's a great thing.
2:27:22
That's uh what uh what about in this I want to talk more about the public market distortions versus private market.
2:27:29
It feels like Elon has kind of run an AB test with Tesla and SpaceX.
2:27:32
They're like within an order of magnitude of each other, maybe within like 2x in terms of valuation and now Elon seems to be building some sort of a Japanese style keretsu between all of his private market and sometimes the public market companies as well.
2:27:47
Um is like would would a hallmark of a well functioning public market be that there is just one ticker for Elon Inc.
2:27:56
and all the companies are under one umbrella or is there a val is there value in the p in the public markets to having like pure plays?
2:28:04
I hear some public markets hedge fund investors are like I just want a pure play on the robo taxi.
2:28:08
I just want a pure play on humanoids.
2:28:09
And I don't know if that's actually like an economically valuable thing for our society to have or what would be best for Elon or the shareholders.
2:28:17
I'm just kind of struggling with all the different stakeholders.
2:28:22
>> I suspect the pure play idea is useful.
2:28:24
So you get a market price signal, what's working, what's not.
2:28:28
>> So Elon obviously is become less popular.
2:28:30
So the value of Tesla goes down.
2:28:32
You see that quite directly.
2:28:32
If they're all bundled, the signal is far more muddled.
2:28:37
>> So I think it's better separate.
2:28:40
So ideally um if it was easy to operate in the public markets we would we would just see SpaceX and Neur and and companies would be going public once they hit like 5 billion or 10 billion something like that.
2:28:54
>> I'm not sure if something like Neurolink would be that's just weird to begin with non-leible. >> Yep.
2:29:00
>> Who are the customers? What are the products?
2:29:01
Yes, they're disabled but after that uh that I would think should be in private markets.
2:29:07
>> Wait, are you sure about that?
2:29:07
because I feel like it's also in in many ways like a classic biotech drug, you know, FDA approvals.
2:29:14
Once it's approved, it goes out to everyone.
2:29:15
It feels like that's one place that the public markets are particularly functioning as they always have where new drug comes goes through phase one, phase two trials, but goes public very early.
2:29:25
We we're not seeing the stripe of biotech hang out in the private markets for two decades.
2:29:32
>> What am I ever going to buy from neuroch?
2:29:33
No insult intended, by the way, but it's not transparent to me. I think it's amazing. I'm all for it.
2:29:40
>> Yeah, >> maybe it has to be not so legible to public shareholders, and Elon wants to do it. That's great.
2:29:46
>> But you might not buy a particular cancer drug.
2:29:49
I would hope that you'd never have to buy any of them.
2:29:51
But that particular drug for some certain indication, some certain illness, could be taken through FDA trials. They could go public.
2:30:00
their stock price could trade on the basis of the FDA results and then when they when when they when they get approved they start selling that drug maybe they get acquired by a bigger drug company but then they live or die by the market size of that particular illness.
2:30:14
>> Well, if it was a cancer drug, sure, I'd buy that.
2:30:16
But a brain computer interface, >> I don't know. I don't want one.
2:30:19
I'm happy with it being there.
2:30:22
And uh again, it seems to me that's probably not a good matter for public shareholders. >> Oh, okay. Yeah.
2:30:28
I'm more just think I mean yeah I I guess if you're thinking about valuing it as like brain computer interface for the masses that's a different story than uh it is a it is a treatment for the for the illness of being a a parapolgic or if you are blind it can cure blindness and that is a medical diagnosis and it is a medical product but obviously Elon's telling a much broader story about that so that's a different thing.
2:30:53
Were you surprised to see Uncle Sam put up a surplus in June or were you expecting that?
2:31:00
>> I wouldn't say I expected it, but you know, month-to-month fluctuations can be so severe.
2:31:06
>> Nothing like that should surprise us.
2:31:07
The debt and deficit, they're still out of control.
2:31:10
>> It's like a It's a startup.
2:31:12
>> It's a startup that that's like, >> yeah, we're we're we're profitable because they have one month free cash one month of free cash flow or one day annualized. Uh >> yeah.
2:31:25
Yeah, that makes that makes sense.
2:31:25
Um, how about stable coins?
2:31:27
There's there's been new regulation.
2:31:29
We have a a issuer that's that's now public.
2:31:32
Uh, there's a massive amount of excitement.
2:31:35
It seems like a lot of the um use cases outside of um you know being you know uh as a um uh value transfer to facilitate investing in the US.
2:31:49
But then the potentially, you know, more interesting use cases internationally.
2:31:52
But but how are you thinking about adoption and and do you believe that um the the US and specifically the traditional dollar will be a major beneficiary of the tech? >> It's been great.
2:32:06
The rest of the world wants access to dollar-based systems.
2:32:10
Our government makes it hard for them.
2:32:12
Stable coins are a workaround.
2:32:15
I'm not sure what their long run fate will be because keep in mind regulations can change and other payments methods can innovate as well, but they're going to pass the Genius Act.
2:32:24
It looks like I'm all for it.
2:32:26
It's not perfect legislation, but it will give people a chance to experiment with these things on a legal transparent basis and let it rip. Right.
2:32:37
>> Why do you say that our government makes it hard?
2:32:39
I always thought it was an authoritarian government that or or or just a foreign government that wants the ability to inflate away their own debt and so they're keeping dollars out of their economy.
2:32:50
They're trying to avoid dollarization.
2:32:53
>> They don't always avoid it successfully.
2:32:55
But you look at the US government know your customer laws for instance.
2:33:01
>> Uh it's a pain in the neck to deal with an American say who wants to open a bank account in Germany. I've tried to do this.
2:33:06
That's because of the US government putting reporting requirements on the German bank.
2:33:09
They don't want your money.
2:33:13
>> So a lot of the problem is USG, not you know Argentina, Russia, whatever.
2:33:16
It's both >> but in fact if people just hold dollars that is usually tolerated >> and uh it's a workaround for both.
2:33:27
>> Uh I want to talk about Google's acquisition of Windsurf.
2:33:30
Uh Google paid a licensing agreement.
2:33:33
They acquired 50 or so engineers from Windsurf.
2:33:35
The CEO and the share and the preferred shareholders got paid out by Google.
2:33:39
Um roughly 400 or something employees were left in a ghost ship in the remain co.
2:33:48
Uh they also had $150 million roughly on the balance sheet and so they were scooped up by Cognition.
2:33:54
Um, I'm not sure how closely you follow the drama, but my question was, is this a paro improvement or a Caldor Hicks improvement for Windsurf stakeholders as a whole?
2:34:06
It raises the cost of capital, and we're going to need more covenants to limit this happening.
2:34:11
Like, say I start a company, raise money, and then just pay myself a huge dividend for all the money.
2:34:17
Well, that's typically against the law because there's a covenant in the contract saying I can't do that.
2:34:22
So, we need new covenants to cover these cases.
2:34:25
Those will be harder to write, but the market will respond.
2:34:29
In the meantime, it's a problem.
2:34:32
Do you believe that the the problem should be solved by employees demanding a new covenant, employees demanding more uh strict employment contracts?
2:34:41
Where will this solution sit?
2:34:47
>> It depends how you're doing the financing, but very often in other settings, it's done by the bond holders.
2:34:52
the bond holders want a covenant that you can't just drain all the resources out of the firm.
2:34:56
The more the firm is based on human capital, the easier it becomes to do that because human capital leaving is not monitorable in the same way that a big cash dividend is monitorable.
2:35:07
>> Uh but we'll figure out ways of doing it.
2:35:09
Maybe it will require arbitration, but this absolutely needs addressing.
2:35:13
Otherwise, you just keep on starting up companies, ship the talent out somewhere else.
2:35:17
In essence, sell it twice.
2:35:17
Bunch of different people are screwed over for the next person who tries it with no intent of doing that.
2:35:23
The cost of capital is too high.
2:35:24
And that's the the longer run problem.
2:35:26
But again, markets are very good at solving problems like that with covenants.
2:35:32
>> So, this is uh this will be solved by the market.
2:35:34
This is not an FTC issue.
2:35:38
>> I would not involve the FTC. I don't like them. I don't trust them.
2:35:41
their previous boss, their current boss. They're both bad news.
2:35:44
Neither knows any economics.
2:35:46
My goodness, they're a nightmare.
2:35:48
Why would you call them on the phone?
2:35:50
They're gonna make it worse.
2:35:54
>> Well, thank you for the for the for the solid breakdown and being very clear about your opinion.
2:35:59
This was fantastic, Tyler.
2:36:01
Thank you so much for coming on the show.
2:36:03
>> Thank you for the update. My >> pleasure. And your wisdom.
2:36:05
>> Good luck with the rest of the day. >> We appreciate it. >> Have a great day. We'll talk to you soon. Bye. Always a great segment.
2:36:12
And we have our next guest already ready to join the show.
2:36:19
Austin, welcome to the stream. Let's play that.
2:36:23
I want to play that intro song a little bit. I like that intro song.
2:36:26
Oh, maybe he's already here.
2:36:28
Maybe you just get Ashton Hall.
2:36:29
>> We got an issue with the soundboard, but we're going to make a sound for you. A handcrafted sound.
2:36:35
>> Kick us off >> with the news and an introduction on yourself.
2:36:39
>> Thanks for having me, guys.
2:36:39
Um, honored to be here everybody.
2:36:41
I'm Austin, uh, co-founder and CEO here at UniFi.
2:36:43
Today, we just announced a $40 million series B led by Battery V. >> Wow. >> Led by who? >> Who? >> Doesn't matter. It does matter. Congratulations. Awesome firms.
2:36:59
Uh, awesome firm leading awesome firms along for the ride. >> Awesome.
2:37:04
uh be a bunch of questions, but uh give give us a quick history on yourself.
2:37:09
You said earlier >> was it the proudest day of your father's life was when you signed an offer to work at RAMP and then the second proudest is today. >> That's true. >> It was. Yes.
2:37:18
It's uh it's been a it's been a big couple years for me with me and my dad.
2:37:22
Um took a lot of years going through college to get to that point, but um >> quick background story.
2:37:27
So originally joined Rant back in 2020 uh on the uh on the growth product team.
2:37:31
uh really attracted to just like the talent density that existed even back in that period of time.
2:37:37
Spent a couple years there uh working on the growth product team with a bunch of incredible folks I know you all know well including Sam Buck um you know folks like um Pville you had on the show the Crosby guys etc.
2:37:50
Um so ton of great folks spent a couple years there and uh spent a lot of time thinking about distribution from an engineeringled perspective and so joined uh saw everything that was happening with open AAI in 2022 and it felt like the nexus of go to market and AI was
2:38:06
starting to happen didn't know that chatbt and things like that were were about to drop but ended up starting Unifi beginning of 2023 uh pretty quickly raised a seed round from a combination of openai thrive emergence and we're off to the races. Um, fast forward two and a half years,
2:38:21
Um, fast forward two and a half years, we're a 50 person team today serving a bunch of incredible companies like Cursor, Perplexity, uh, you name it.
2:38:26
And so, uh, just grateful grateful to be here.
2:38:31
>> Okay, explain the product like I'm a newrad, uh, SDR.
2:38:36
>> So, if you're a newrad SDR, the pitch of the product is you used to have to push a million buttons just to get to the place where you could actually do something that was valuable to talk to a prospect.
2:38:45
We take care of all that on your behalf.
2:38:47
So, example is you might show up in the morning.
2:38:49
You're going to make your list of the 25 people that you want to reach out to based off of, okay, this person started a new job.
2:38:54
I reached out to John in the past.
2:38:56
He's now at this new thing.
2:38:58
He started a new job or he got a new promotion.
2:39:00
We'll make that list for you.
2:39:02
We'll automate all that research.
2:39:04
We'll also deploy agents to go get 50 other data points for you.
2:39:06
So, you can start from the 90 yard line and then you can just be polishing the messaging and the actual work that we've done for you.
2:39:14
But you don't have to you don't have to start by just like manually combing through LinkedIn or manually combing through a database like Crunchbase.
2:39:21
So we'll do all that work for you. That makes sense.
2:39:22
What is the sentiment like among sales leaders uh all the way through kind of the the entry level roles uh STRs BDRs etc. just around AI broadly.
2:39:34
I think the uh there's been a lot of dialogue recently around you know the core challenge with a lot of AI tooling right now is this amnesia.
2:39:44
So the the the the model is really smart but it can't remember sort of lessons from at any point in history.
2:39:53
You know there's some memory but not the sort of hard one lessons that you get from going out and doing a hundred phone calls with customers things like that.
2:40:02
Yeah, that's a really great question.
2:40:04
So, the evolution of the sentiment has changed over the years.
2:40:06
So, 2023 we started the company.
2:40:08
There was a lot of aversion to using AI to do the endtoend job because it just wasn't there. Wasn't remotely there.
2:40:13
Agents could sort of you could see them coming, but we weren't in a place where it felt like that value was tangible.
2:40:18
And so, it was really in early 24 where it started to feel like the world started to believe that even as a seller, the best thing for me to do was to adopt AI to try to automate a bunch of a bunch of what I'm doing.
2:40:27
But to your point, exactly 24, it's been this like version 1.
2:40:32
0 of AI where products are dumb.
2:40:34
They they can do basic research tasks for you.
2:40:36
They can do them over and over again, but they have no context that they've done that task four times in the past.
2:40:41
One of the things that we're really excited about that we launched a couple months ago is something we call our observation model, which is actually this underlying primitive that looks very different than anything else that's existed.
2:40:52
Basically looks like you have a memory associated with every person you've ever talked to.
2:40:57
And that memory is not a database.
2:40:57
It's actually a flexible set of context, just a big text doc basically that is every interaction and every piece of notable information between you and that prospect.
2:41:07
And so as an example, let's say we're talking about Perplexity.
2:41:10
Perplexity has maybe reached out to um I don't know, Ford Ford Automotive in the past and they've got some record of all the sellers that they have and they've contacted them.
2:41:18
They've gone back and forth.
2:41:19
They may like maybe exchange a few ads.
2:41:21
few ads. um all this sort of data has usually existed in a bunch of different silos but we actually built this observation model to join all that together so you actually have one clear set of context and then when Unifi wants to understand hey what should I do next it can ask that set of context what do I know what's the most relevant thing that
2:41:38
I can surface right now and let me go ahead and hit on that and so we really see that as being the evolution of where these AI products are going to go to just make them that much that much smarter and more powerful >> it feels like uh one of the really like just layup use cases for artificial intelligence in a sales context is just data hydration, cleaning up the CRM. But
2:41:54
But I'm interested to hear about how high the walled gardens are in a business context.
2:42:03
Um, we've heard stories about uh Glean getting push back from different uh big tech companies and it's one of those weird things where uh every company says, "Yeah, it's your data."
2:42:15
Until I want to give it to their competitor potentially that's overlapping.
2:42:18
So um what data sources are are are particularly like AI friendly these days and um and what is the shape of the of the garden and the walls of that garden these days?
2:42:31
>> It really depends on I think the overall data strategy for the company and just like what their strategy is built on.
2:42:37
And so to give you one example Salesforce historically their advantage is built on being a fairly open ecosystem.
2:42:42
So Salesforce um the app exchange launched call it like 10 20 years ago at this point it's been building emotes around Salesforce as people have connected tens of applications into the product and that's sort of built the stickiness.
2:42:54
You'd have 25 different products that all integrate through Salesforce that happens through this open ecosystem of connectivity.
2:42:58
My my gut says is that the platforms that are built much more on open-endedness like something like a Salesforce are going to continue to remain open as that's the core value prop of the product and that's the reason why they have a mode itself.
2:43:11
But I think products that have historically been a little bit more walled off and so you know Slack is an example of that will probably continue to stay that way and trend even more in that direction.
2:43:19
Generally speaking, people haven't been super privy to give out Slack information.
2:43:23
They wanted to keep that in especially within an organization and so I imagine we're going to see those things clamp down even further.
2:43:28
And case in point is the is the glean example you brought up.
2:43:32
>> Do you have a thesis around uh sales co-pilot type products?
2:43:36
You can imagine this is a sales call and you're you know running an LLM in real time to to give you the next answer.
2:43:44
I don't think you guys have one of those uh in your suite of products, but I'm curious how you think about that category.
2:43:53
>> It's definitely super powerful.
2:43:53
I mean, you see what's happening with with Chloe right now, right? And what's going up. question.
2:44:01
>> Let me uh let me just real quick uh it's wild though and I think uh we're only at the beginning of that period of time and I think the the way in which people absorb information is going to dramatically change over the course of the next couple years as we get more and more used to these products.
2:44:13
Today, the closest that we get to that in our product is that we've got uh a composer for for emails and for anything you want to do in Unifi.
2:44:22
And next to that, you've got all these AI cont all these AI nuggets and AI research uh tidbits that we've done for you.
2:44:29
Alongside that, uh we're generally speaking believers that the faster your product is, the more real time it is, the more addictive it is to users, the more helpful it is to users in that context, especially in a work context.
2:44:41
And so we're going to keep pushing the product in that direction.
2:44:45
Uh something like a clue is like the far-flung example of that.
2:44:47
We're not quite there yet, but I imagine the world is going to keep moving that direction. >> That makes sense.
2:44:52
>> I want to talk about uh uh you I mean your your your your website's tagline is scale your revenue team's creativity.
2:45:01
And I'm wondering if there's any sort of like just like really creative sales solutions.
2:45:07
Like I was talking to Sam Blonde a while back.
2:45:09
He was working with Parker Conrad for a long time and said like one of the best campaigns he ever did as a sales guy.
2:45:16
He's like a SDR for a long time.
2:45:19
Um was in head of sales was like sending bottles of champagne to people because like they just open it then they have a phone call.
2:45:26
They have this nice thing and so >> it becomes incredibly rude not to respond to the email. >> Exactly.
2:45:31
And so I'm wondering about these like weird squishy like the handwritten note and and in many ways I feel like these creative uh these creative deal closers like the the unexpected text message from a friend who's already using the product is the thing that gets you to actually go and say, "Yep, I'm ready to onboard fully.
2:45:48
I get my company on this new product."
2:45:50
Um, but is it is there are there any are there any like world I feel like a lot of the creativity in in in uh sales is specifically showing the client that I'm not using software for this. I handwritten that note.
2:46:07
I I I sent your I got your address.
2:46:10
I sent you a bottle of champagne.
2:46:11
And so is the goal to more just like remove all the barriers to let the humans do the creative things or do you think that there's a world where we can use AI tools to uh speed up those creative out ofthe-box solutions that might be uniquely tailored to just this one company strategy.
2:46:28
They're the only one that does this weird thing, but it's working and they found this alpha and they stick with it for a long time.
2:46:35
>> Yeah, big fan of of Sam's and uh he's done a bunch of incredible outbound campaigns over the years.
2:46:41
My thinking is that what will happen is what always the thought behind all these things and the reason why it matters this all comes back to taste right I think people in more and more are talking about taste in the world of AI and it's all about knowing what's relevant at the time knowing what's going to resonate with someone personally and being able to do
2:46:56
something that feels really delightful um so it's an example like one thing that we did as part of our fundrais announcement today we put out uh about 25 cameos that we had custom build for everyone we've got one for you guys on the way right now so we'll have it to you later but like things like that that are based off of someone's personal experience and show that you thought about it. Doesn't really matter if you
2:47:14
Doesn't really matter if you use chat GBT in the brainstorming process, right?
2:47:16
What matters is that you got the idea that originated from something special and then the actual end delivery is the same quality you put out as a human, >> what the person feels. >> Yep. >> Exactly. Exactly.
2:47:26
But I think ultimately today AI doesn't feel and so it can't really use that judgment.
2:47:30
There's a reason why like in the world of content and in the world of the stuff that we do here, it's like you just can't AI that.
2:47:36
It just feels bland and generic.
2:47:36
And I think that's going to continue to be true in a sales context.
2:47:39
What you need is a human behind the scenes who's providing that little touch of like what it means to be a person, what it means to be a friend.
2:47:45
And so we'll continue to lean into experiences that make it easier to stay on top of those friendly experiences, but um but just get easier over time to actually execute. >> It's great. Anything else?
2:47:58
>> Uh I got a bunch more questions, but we'll save it for another time. Thank you for joining. Congratulations.
2:48:02
Congratulations on the new raise. >> Get back to work. Get back.
2:48:05
>> I'm sure you guys have a bunch of people to hire to uh deploy that capital. We'll talk to you later.
2:48:10
>> Thanks for having us, guys. >> Thanks for coming on. >> Cheers, Austin.
2:48:13
>> In other news, Apple is reportedly planning to invest $500 million in um a rare earth magnet startup. This is MP Materials.
2:48:23
Uh they're going to build a new recycling facility in Mountain Pass, California.
2:48:27
Currently, they recycle rare earth.
2:48:30
They they take recycled materials and they make rare earth magnets in Texas.
2:48:33
Uh there's a huge Wall Street Journal profile on MP materials and now there's news that they're working with Apple.
2:48:40
My takeaway is that this uh it shows that Tim Cook is actually thinking about long-term supply chain strategy and it's this it's this funny.
2:48:50
>> How much do you think MP Materials is up in the past 5 days? >> I have no idea. >> 85%. >> Let's go. Ring the gong. 500 million. Do it. Hit that gong.
2:48:59
>> I'm gonna warm up the gong properly. >> Oh, yes. Yes. Yes.
2:49:00
So, we've gotten some notes that you must warm up a gong before you hit it.
2:49:04
Also, hit the gold part, not the center. >> Really? >> Yes.
2:49:07
That's what uh that's what somebody said in the comments. There we go. Is that better? It's about the same. >> I don't know.
2:49:13
I'm not I I don't know enough about I'm sure I'm going to be an expert in 20 years when uh we're still doing the show.
2:49:19
Um but I thought this was interesting because, you know, Apple's a huge company.
2:49:23
Tim Cook's a legendary operator and >> correct.
2:49:28
There's just something about being a big company that even if you're seeing, oh, the current thing on X today, everyone's talking about supply chain in China being a risk, it's going to take years to actually build up another supply chain.
2:49:43
So, this kind of tied to the re-industrialization story.
2:49:48
There's like you can start a new company.
2:49:49
You can actually build something like before Apple can really uh like turn the massive uh cruise ship that is the or the battleship or the aircraft carrier that that the company is.
2:50:01
But it seems like it's happening and this seemed like a huge white pill.
2:50:04
So I was excited to talk about it.
2:50:08
>> Um and we should and we should uh dig into MP materials more deeply uh soon.
2:50:13
They're up 22% I guess just the day that this profile came out.
2:50:15
Um, and uh, seems like a very very exciting company.
2:50:18
And honestly, um, when we were talking to the co-founder Cororeweave, I uh, you know, the there's there's this, you know, we all learned about chips and then ASML and then SKHix and memory.
2:50:30
Um, and I feel like rare earth minerals and rare earth magnets are going to be the next thing where everyone's learning about it at the same time.
2:50:40
And a lot of people already have and they've done pieces on it, but I want to learn more.
2:50:45
I could see AP doing a rare earth >> Royal Oak.
2:50:48
>> That would be incredible.
2:50:48
Yeah, like instead of the uh what was meteorite dial? The meteorite dial.
2:50:52
You get the rare earth >> the rare the rare >> the >> um go through some timeline.
2:50:57
I'll be right back and we have our next guest.
2:51:00
We'll welcome Chris into the studio. >> Okay. Yeah. Yeah.
2:51:01
I'll I'll I'll go through some stuff.
2:51:03
I'll I'll just welcome Chris in and uh and get the intro from him.
2:51:06
So, welcome to the studio, Chris. Good to see you. It's been far too long. What has it been? Over a week. >> Over a week, I think. That sounds right. I'm excited to be here. >> I'm glad to have you.
2:51:15
Uh, kick us off with an intro and uh, and a little bit of an overview of the news today.
2:51:22
>> Well, look, before I get started, I should say that if at any point during this interview, Windsurf gets acquired for a fourth time, cut it off and I will uh, there will be no hard feelings.
2:51:29
Uh, quick intro on me, though.
2:51:31
I've been in the valley for 15 years, working with tech companies the whole time and kind of every stage and uh, uh, variation and just about any role that you could think of.
2:51:40
So, I most of my career was as as an investor.
2:51:42
I worked at benchmark working with early stage companies.
2:51:44
CO2 focused on some later stage companies.
2:51:46
Uh before then I founded a healthcare company called Curology and before that I was an investment banker working with big tech companies and helping them uh taking them public and and do M&A transactions and all that kind of stuff. >> Okay.
2:52:00
We'll get into the news in a bit maybe when Jordy gets back.
2:52:02
maybe when Jordy gets back. Um the the question I have is like as an investor as a board member um there is generally a you mentioned wind surf there's generally a duty to be like founder friendly but what do you do when the founders and the employees are misaligned like are you as a board
2:52:20
member are you supposed to vote in favor of what's best for the whole group of founders plus employees as a class are you supposed to d back the founder if they want to do something that's less employee friendly vice versa like this feels like the modern debate that we're having in Silicon Valley is like the social contract is is re-evaluated. How
2:52:36
How have you processed the windsurf news?
2:52:39
Um what how do you think this changes going forward?
2:52:44
What what are the next questions that you think uh venture capitalists and founders need to be talking about to make sure that you know when amazing deal happens everyone's happy about it? >> Yeah.
2:52:56
Well, I think your first goal as an investor, you know, we rarely have a lot of power.
2:52:59
we have sort of in the best case influence.
2:53:00
But I think your first duty is to try to help the founder understand that what's good for the employees and what's good for everybody um is sort of the right thing for everybody over the long term and the ecosystem and your reputation all that stuff is really critical and so I uh I I was delighted to see that everybody got to a very good outcome on this one and I think that's uh it's where it should have landed.
2:53:19
This feels like a 1 plus 1 equals 3 kind of thing. >> Yeah. Yeah.
2:53:22
It seems like uh I mean when the Cognition news broke, we had been talking about like, oh, where could the remain co wind up?
2:53:29
Maybe it lands at like a more legacy tech company like the Humane team wound up at HP and people were kind of memeing about like AI printers and stuff.
2:53:39
And that seemed like, you know, it probably would have been a fine outcome financially, but it would have been like, okay, a lot of those people kind of need to restart their careers and kind of get back into the high aggressive Silicon Valley, high growth startup world that they probably fell in love with.
2:53:54
And that's why they wound up at Windsurf 6 months ago or 10 months ago or 15 months ago.
2:53:57
Um, but with cognition, it's just a complete continuation of that aggressive take over the world, build something new mindset.
2:54:05
So, uh, yeah, I agree with you.
2:54:08
Amazing outcome for everyone involved.
2:54:09
Obviously a tumultuous weekend, but we're glad that the plane landed, even if it was a little bit bumpy. >> Some better news.
2:54:14
We got to talk about the new vehicle.
2:54:16
Uh it broke in Forbes this morning, but we're happy to have you on here today to talk about it.
2:54:20
Give us uh give us a story.
2:54:22
Give us the number, and I'm going to stand up and get ready to to do something back here.
2:54:27
>> So, $175 million debut fun. >> Let's go.
2:54:31
Congratulations, >> you guys.
2:54:33
Hey, is the gong like individually miked?
2:54:34
Cuz that sounds incredible. This is the mic. It's got its own camera.
2:54:40
>> But uh but yeah, we're always working on the fidelity of the gong.
2:54:42
It's really important to uh to to the >> Congratulations. 175 million. >> 175 million.
2:54:48
Stage agnostic, sector agnostic is the idea.
2:54:51
I'll make a very small number of investments.
2:54:53
So sort of 8 to 10 core positions.
2:54:55
A normal venture fund might have 30 or 40 positions or something like that.
2:54:59
And the idea behind that was my favorite part of my career was when I was just making my first few investments that were going to matter a lot for me and because it felt like there was a symmetry between what I cared about and what the founder cared about.
2:55:13
Like for the founder, this company is their whole life. It is everything.
2:55:15
They live and die with it.
2:55:17
And I think one of the problems as you get more senior as an investor, you naturally have more companies that you've invested in.
2:55:21
Every individual company is going to matter a little bit less.
2:55:24
And so with this vehicle, the idea was create a vehicle where structurally everything will matter to me.
2:55:28
It's 8 to 10 core investments.
2:55:30
Every one of them is going to matter.
2:55:32
Uh and I'm going to work really hard on each of them.
2:55:34
Stage agnostic, sector agnostic, still I imagine not an RAIA, not playing in the public markets, not not putting more than 20% of the fund into secondaries. Is that roughly correct?
2:55:47
It's crazy I even have to ask that question, but we are in crazy times.
2:55:51
Is it even possible to start an RAIA at this scale or is that something that only the mega funds are thinking about?
2:56:00
>> No, it's definitely possible.
2:56:00
It just it eats into some of your uh management fees and it requires some various compliance obligations that you have to get used to.
2:56:07
You probably have a second phone and some things that are like you hire some compliance consultants and all that kind of stuff.
2:56:12
I had looked at it because I do like buying secondary from time to time and just decided against doing it.
2:56:17
There's also some additional rules around how you can market uh the the fund and stuff like that.
2:56:21
You have to be a little bit more careful when you're an RA and uh uh so I decided to go the RA route. It's a flexible vehicle.
2:56:30
So I'll do things just kind of of any stage.
2:56:31
I will often not be leading rounds.
2:56:33
I'm willing to lead rounds.
2:56:34
The idea was because I'll be so selective on the company quality.
2:56:36
I think I have to be flexible about everything else.
2:56:40
And so I'll try to just show up in a way that is useful and easy to work with.
2:56:43
How is the concentrated strategy resonating to founders so far?
2:56:49
I imagine it's it's a very real differentiator of just saying, "Hey, you can be one of of 40 bets or you can be one of 10 and I'm going to do everything in my power to to help you win."
2:56:59
Can see how that would resonate.
2:57:02
Um but uh but what's it been like on the ground?
2:57:06
>> I think so far it's working well.
2:57:06
You know, I have founders who text me 10 or 20 times a day.
2:57:09
Um, I also have founders who want to be left alone for the most part and I'm happy to do that also.
2:57:13
But I think I think so far it resonates.
2:57:15
I like, you know, I tell every founder when I make an investment, I just want to be along for the ride and things will be good, things will be bad.
2:57:22
A lot of the time they'll be difficult.
2:57:23
You know, building companies is mostly pain.
2:57:24
Uh, and I will just try to be someone who eases that along the way and helps you navigate the difficult moments and all that kind of stuff.
2:57:33
Um, is there is there an alternate world where you launched this fund prior to Curology and and uh I I think uh it's sort of funny because if you were in that situation, you would have said I had a non-traditional back, you know, uh background for venture capital and it's HBS and go to and benchmark and sort of a traditional background, but you went and did the hard thing and built a big company.
2:57:58
Um but uh but but was it did you have it in mind that I'm going to you know go be a founder, do this and then come back around to venture or or did was it all um just kind of an accident?
2:58:10
>> I think I was always built to be an investor.
2:58:11
Um I enjoy it and there's a lot of sort of features of investing that are well suited to me, but I just think it's kind of a right of passage if you haven't operated inside of a company.
2:58:20
If you don't know what it's like when things are difficult and bad, I think it's kind of difficult to give good advice to a founder.
2:58:24
And so I think it was an important thing for me to develop some empathy.
2:58:28
Like I try to uh understand that this is it's one of the hardest and loneliest jobs in the world.
2:58:34
You'll have weeks where everything is going wrong.
2:58:35
Uh and there's sort of, you know, playbooks for navigating each of these things that go wrong, but I think you have to have been through them before to understand how to do it and who to call and what the footfalls are and that kind of stuff.
2:58:45
And and uh that's sort of how I thought about it.
2:58:48
>> What was the fund raise process like?
2:58:51
175 million for first fund.
2:58:51
Uh it's a it's a big number.
2:58:54
Uh, but I heard from a little birdie that you got some pretty insane LPS in this one.
2:58:58
So, it sounds like you were doing something right.
2:59:02
>> Well, I think that's that.
2:59:02
Well, thank you for that. And I think it's right.
2:59:05
The um I it went relatively quickly.
2:59:05
I was real lucky because I had a few LPs who were just fast and early believers and everything else sort of uh um ended up forming around that.
2:59:14
And so I ended up sort of many times overs subscribed uh and ended up with a group that I'm very very happy about.
2:59:20
I think most of them I'm not allowed to say their name, but the ones that I can say are Adam Street, Common Fund, Northwestern University, Howard Hughes Medical Institute, and a bunch of causes that I'm I'm very excited to be working for here. >> That's really cool. >> It's exciting.
2:59:34
What do you think um how do how do you think about uh investment kind of time horizon?
2:59:39
If you have a a such a concentrated fund, very possible that you'll meet five companies in the next six months that you'll invest in, but it feels maybe unlikely.
2:59:50
Maybe it's more like one a quarter, but but uh it's hard to plan these things.
2:59:56
>> Yeah, when I was raising, you know, everybody says they're going to deploy the fund in three years and then they go they end up going a lot faster than that.
3:00:02
And so maybe this ends up being two years or something like that, but I just don't have that many amazing ideas uh every year.
3:00:08
And so I part of the reason for the way the fund is structured is it's kind of how I like to work that I develop a lot of conviction around something over a period of time.
3:00:16
But then I like to have a bet that matters.
3:00:17
I think the single most painful thing for me is if I I find a company and I sort of am able to persuade the founder I'm able to win the right to invest and it ends up working and it still doesn't matter in a material way for my limited partners.
3:00:30
That's uh that's something that that I think is uh is something I don't want to deal with.
3:00:35
And so I that's the reason for the sort of bigger bets, the concentrated strategy.
3:00:39
And that that requires I think going a little bit slower also because there just aren't that many uh companies that that will meet the bar. >> That makes sense.
3:00:47
>> That makes sense. I have a question about venture markets generally loosely in the context of of wind surf but I'm wondering if it feels like right now the labs are on a buying spree not just open aai and anthropic but uh deep mind and you could see acquisitions happening all over the place and it one framework to
3:01:08
think about it is that they're all just buying call options anyone who's got a unique technical direction or unique technical talent the labs are just happy to pay any price because maybe it's 1% of infinity if you solve super intelligence or maybe it's hey we already have you know Mark Zuckerberg has a massive business in you know
3:01:29
delivering ads if you make that 1% better you've added 10 billion dollars to the market cap so what's a couple billion in in acquisitions going out the door uh totally makes economic sense but my question is is does this change venture capital underwriting at any particular stage age like it feels like you really want to be backing
3:01:52
I guess teams that could be you know aqua hire talent because it gives you this like liquidity option but then maybe the real alpha is finding the >> but at the same time very mission if you look at the capital deployed into wind surf yeah obviously the majority of the capital didn't get a tremendous return despite it being a big headline number
3:02:14
>> is that just because >> a lot of I mean seed and they did >> well I'm just saying like the majority of the capital came later and then I think it was probably at at somewhere around a billion >> okay yeah I don't know but yeah I mean I think on the underwriting question I think in the history of our industry there was always like a hard and fast
3:02:30
rule that you always backed one company in a space there was always going to be one winner >> and then I think as the markets have gotten progressively bigger and you know there's there's like one winner in search for a long time and one winner in operating systems as the markets got bigger and bigger you started to have
3:02:45
like two player markets like you have iOS and you have Android and then they got bigger and you have three player markets and you have you know the cloud AWS and GCP and Azure uh and I think AI is going to be the biggest market of all of these and I think one of the ways that it changes underwriting is that there's probably going to be more
3:03:03
winners in each space than we expect and I think you see this a little bit with Curser and with Windsurf and you see some like really high quality companies I also think it changes underwriting because it feels like M&A is coming back a little bit uh and I think there's sort of some a cause for optimism around that. >> It's great. Anything else? >> It's great. Anything else? >> Makes sense. Uh no.
3:03:18
Uh join any join any boards recently?
3:03:21
I don't know if you can talk talk about that kind of thing. >> Yeah. Yeah.
3:03:25
Join the board of perplexity recently. >> Congratulations. Let's go.
3:03:29
>> I wanted to give you an offramp there in case you weren't ready to talk about it.
3:03:33
Um >> I the product I use 20 times a day and have since it came out for investment research and everything else and one of the the best executing teams I've ever been around.
3:03:41
So I'm very excited to be a part of the effort. >> Fantastic. We love Arvin. >> Yeah. Yeah.
3:03:46
I find myself using it every time we record. Yep.
3:03:49
>> You guys got to try the browser.
3:03:49
If you haven't, >> we we we demoed it live on the stream.
3:03:53
Tyler, our intern, used the Perplexity Comet browser.
3:03:56
Uh had some good feedback for it. It was good. >> Amazing. >> Awesome.
3:04:01
Well, congratulations, I would say, to the full partnership, but you are the full partnership. Is that correct?
3:04:06
Are you gonna be Are you gonna be adding anybody to the team?
3:04:09
>> Just me and the head of operations. >> Very cool. >> Amazing. >> Bet on yourself. Love it. Awesome, Chris. Well, congratulations. Thanks for coming on.
3:04:16
I'm sure we'll have you back. >> Thanks for having me. We'll see you guys. >> Good seeing you. We'll talk to you soon.
3:04:20
>> And we have our next guest already here in the waiting room. Uh Dick Lucas. >> He backgrounds.
3:04:29
>> Video going viral today. There he is. >> Wow. Look at that background. How you doing? >> I'm good. I'm good. How are you guys doing?
3:04:37
>> Looking sharp in the suit.
3:04:38
>> You You look like you changed since the viral video. >> I I know. I know.
3:04:41
Actually, I uh my shirts are uh my shirts are dirty. This one's not.
3:04:45
Um so, this is the same outfit, same pin as well.
3:04:47
So, >> congrats on the uh success of the show.
3:04:51
It's been awesome to watch you guys cook.
3:04:52
So, thank you for fitting me in today. Appreciate it. >> Yeah, it's awesome.
3:04:54
Uh I saw your name for the first time this morning when you started going viral.
3:04:59
Uh great silhouette on the video.
3:05:02
It looked like you're in the Palisades somewhere.
3:05:03
where where you just give give us the full history on yourself and and maybe what what inspired you to get involved uh in politics or mud wrestling as we call it. >> Awesome.
3:05:16
Well, I think it really starts in uh it starts in high school when I got my first computer dig in the dig. com days.
3:05:21
That's when I got really into technology um and got really into software and saw this whole boom taking off and like this stuff's really cool.
3:05:28
Um I didn't think computer science I thought that was only for smart people.
3:05:31
So, I didn't declare my major for CS until sophomore year of college.
3:05:32
Like, it was it was a history minor.
3:05:36
I was like, you know, I'm going to just do the CS thing.
3:05:38
It kicked my ass, but but I but I got through it was you can get a bachelor art of arts and computer science if you can believe that. And that's what I got.
3:05:44
Um, so I I have a long belief in in in technology and then from there um started, uh, doing software stuff.
3:05:52
I was an Android developer for several years, so I was in the code mines.
3:05:55
um started an agency with with a homie uh with Sawyer and now we're um uh yeah, we're cooking on that.
3:06:03
And what inspired this really is is I was creating voter guides.
3:06:05
I was like, okay, California's obviously run run horribly.
3:06:09
If maybe if I just read all what the candidates are saying and go deep on like, you know, who's running for waterboard, who's running for this, who's running for that.
3:06:15
And actually, no one even knows what a state assembly person is.
3:06:17
I'm running for state assembly.
3:06:19
Most people don't even know what that means.
3:06:20
So, I'm like, "Okay, if I go and like look at these candidates, there's got to be people that are out there that are good."
3:06:25
Um, and I'm going through these campaign websites and they don't even they barely even say what they believe. There's no competition.
3:06:33
There's not even debates at these local levels.
3:06:34
And I'm just like, there is a huge opportunity here for someone that is willing to just get themselves in the game, put a couple videos out there, say some common sense stuff, and shake things up.
3:06:45
So, uh, that's that was really the inspiration.
3:06:47
Um and uh uh >> so what tell us what uh tell us what running for state assembly means.
3:06:53
What what is the role that you would like?
3:06:58
What does that look like?
3:06:58
And then what are the what's the process to get there?
3:07:02
>> Okay, so the primary is in June.
3:07:02
It's in June of uh >> June of 26, excuse me, June of this year.
3:07:06
Um and uh excuse me, next year.
3:07:06
So there's 80 assembly members, right? In in California.
3:07:13
Each state has something different, but it's basically like a mini senate in a mini house of representatives, just like the federal government.
3:07:18
So, you have 80 assembly people and you have 40 state senators.
3:07:22
And unlike uh you know, in California there's 80 districts.
3:07:24
I'm in the 51st district, Santa Monica, West Hollywood, um Westwood, uh other parts of Hollywood, Beverly Hills, um and each district has uh assemblymen and then each and then there's also state senators as well.
3:07:36
So, same thing, federal government.
3:07:39
There's stuff that comes out of the Senate um state senate.
3:07:42
other stuff that comes out of the state assembly and uh that's how legislation gets passed.
3:07:45
That's how the machine works.
3:07:47
And right now what we've seen is it's basically one party state.
3:07:49
Now I'm not I think the left has good ideas.
3:07:51
I think the right has good ideas.
3:07:53
I think that there's opportunities on both sides.
3:07:55
The problem is when you lack competition to system, which is what we have in California, it becomes sick. It becomes ill.
3:08:00
I mean, we see this in every single institution or system uh since the dawn of time. You need to compete.
3:08:06
There needs to be pressure on the system or it collapses from within.
3:08:10
And that's what we're seeing in California.
3:08:11
We're starting to see a a a small shift, I'd say.
3:08:14
But you know what the what really broke the camel's back for me was fires happened and people wouldn't even hear about um the wildfires in the Palisades, the eating fire out where where I grew up.
3:08:24
Um people wouldn't even hear they wouldn't even want to listen if you said, "Yo, maybe we could have done some things better."
3:08:30
Like the San uh Sanz reservoir that went dry in the Palisades was down for eight months.
3:08:34
Um I I talked to I talked to a state senator.
3:08:37
I went to a meet up with him.
3:08:38
I was like I was like, "Bro, what happened there?"
3:08:40
He said, "Well, it's unclear how much that would have actually helped."
3:08:42
It's like, "Dude, if it helped 5%, if it helped 10%, if it helped 15%, like you couldn't even really have a discussion about it."
3:08:48
And I said, "Okay, this is the the system itself is is is completely sick where where we had a whole bunch of stuff burned down.
3:08:57
Our mayor in Los Angeles was gone on a trip.
3:08:59
We had fire hydrants go dry and and people even in my network, people didn't even want to talk about it.
3:09:05
So, I guess the only way to talk about you got to run. >> Yeah. No, I love it. Taking action. We We covered the fires.
3:09:10
We John and I are both here in LA.
3:09:12
Uniquely, John lives in Pasadena. I live in Malibu.
3:09:14
We both had to uh get out of town for a while. Uh I missed a show.
3:09:20
John was covering one day.
3:09:22
It was really the only show I met missed this year.
3:09:24
Um but yeah, but the the um my reaction to the fire is it seemed like uh there was a lot of excitement around clearing PCH because that was the thing that was most obviously uh reminder of the the competency of the government.
3:09:42
And so uh but that's only a very small percentage of of the of what the over overall response should should look like.
3:09:50
I'm curious, do you know who your real competition is, uh this year? Yes, I do. Yeah. Dem a Democrat.
3:09:54
He's been in the machine for several years.
3:09:58
And that's the other thing looking at these people what they've done.
3:09:59
A lot of these people with have this audience.
3:10:03
If you are considering running, you should run for office.
3:10:06
You should look at who your local person is and you should run because not you too, all your listeners, but you too as well because you look at their website, you look at their CV, you look what they've done.
3:10:13
A lot of these guys are community organizers.
3:10:14
I still don't know what that means.
3:10:16
And also people they're lawyers or they're people that haven't built anything or created any jobs.
3:10:21
So it's like this all my smartest friends, no one even thinks and I'm sure it's same with you guys.
3:10:26
No one even thinks about going into politics.
3:10:28
So it's like if if you if all the people that are smart or driven in your network aren't going in to politics, who is going into politics?
3:10:35
And well, we see players basically. >> Yeah. No, that makes sense.
3:10:41
>> What's at the top of the stack in terms of lowhanging fruit?
3:10:43
You get in, what do you want to push on?
3:10:46
What do you want to change?
3:10:48
What are the most exciting or like tractable problems that you think you could you >> I think the biggest thing that is resonating right now is uh the cost of housing.
3:10:56
Um you know there's the phrase that the economy is stupid and obviously we need to work on that too but housing in California is 100% the national average.
3:11:02
Same with our uh same with our gas prices.
3:11:04
Um our gas utility prices, gasoline is the highest in the country um as you know.
3:11:09
Uh >> which is not a which is is not uh partially due to the market but also due to just massive taxes that the government has voted to apply to every gallon at the >> for gasoline. Yeah.
3:11:24
>> So you drive across the border, it's it's not like a supply demand thing as much as it is >> or a top eight oil producing state.
3:11:29
it had the highest in the highest gas price in the country. Doesn't make any sense.
3:11:33
There's obviously there's also some good things there. We want clean gas.
3:11:35
So, California puts a lot of, you know, restrictions on refiner refiners on, you know, they want to pay for the top quality clean gas so we don't have we don't have uh, you know, smoky valleys and it doesn't look horrible like it did in the 60s.
3:11:49
So, I get all that, but it's completely out of whack.
3:11:53
>> So, the other but housing, we need to we need to make we need to make it easier to build.
3:11:57
Um uh they did just pass the Dems did just pass something and I tried to understand it.
3:12:01
You can go read the you know leg you can read these bills and like they they talked about it as a big win but it's it's it's difficult to even understand what's going on like in terms of what they got done you know they're making it easier to build now recently as of like July 3rd near um transit areas >> and you go read the bill and it's like 30 pages of is like is this making easier?
3:12:22
Why isn't it just say like you can build here?
3:12:25
If you own the land, you can build here.
3:12:26
Like, this is it's completely out of whack.
3:12:28
So, um, we have a supply problem here.
3:12:30
We need to just make it a whole lot easier to build.
3:12:33
And, you know, there's people talking about freezing the rents.
3:12:34
We can do better than the funny thing is is the people that lie to you will say, "Oh, we're going to freeze the rents for everyone."
3:12:40
That's not even that good.
3:12:40
We can do better than freeze the rents.
3:12:41
We can actually lower if you add supply.
3:12:44
We've seen this in Austin.
3:12:44
We've seen this in other cities.
3:12:45
And my view is there's 50 different experiments going on everywhere in our country.
3:12:48
Every state, you just look at what works.
3:12:52
Where are prices coming down?
3:12:52
Where are businesses going? and you just fall. Yeah.
3:12:55
So, it's like we have the tools right here.
3:12:57
I I I don't I'm not the smartest guy in the world.
3:12:59
I know that I have a lot of respect for people in technology because I think they're epic. I think they're smart.
3:13:03
I think they they raise a standard of living for everyone.
3:13:06
And I'm not going to in, you know, invent the next nuclear reactor, whatever. So, I'm looking at this.
3:13:10
I'm like, this is an opportunity that I can actually, you know, I'm smart enough.
3:13:14
And it's like, we have 50 experiments.
3:13:16
Let's follow what works and let's just do it.
3:13:18
>> How can people support you?
3:13:18
How do people like who who are your constituents?
3:13:22
who will actually be voting for you.
3:13:24
It's It doesn't sound like it's just everyone.
3:13:27
>> And you cover Are you going to be covering the Gundo or is that out of range?
3:13:30
>> No, it's out of reach.
3:13:30
It's out of reach, unfortunately.
3:13:31
Um it' be it'd be an odd to represent those guys, but um dicklucas. com.
3:13:35
Uh my biggest following is on X.
3:13:37
Damn, you got me sweating, boys. I'm fired up. Um uh yeah, dick. com. Stoked on that.
3:13:42
Um uh yeah, I'm not taking any money right now.
3:13:46
We're just going to see what noise we can make.
3:13:47
The people that will vote for me are the people, like I said, in my district.
3:13:50
And I don't I think very few of those people have seen this viral video.
3:13:54
So I want to do two things.
3:13:54
I want to get the word out with you guys and with technology people and the people in my network.
3:13:58
And the second thing is ground game. I got to knock on doors.
3:14:01
I got to I got to talk to people.
3:14:03
Now that I have a website, my video up, I'm going to just go door to door and be like, "Yo, let's let's talk.
3:14:07
Let's I I actually want to hear what what people obviously what's their biggest concern?"
3:14:12
Um >> when you say knock on doors, do you mean literally knock on doors? >> Yes. Yes.
3:14:15
Door knocking is very effective in these local races.
3:14:17
You just have to you just have to do it. Like um yeah.
3:14:21
>> How many how many Yeah.
3:14:21
Do you have to you have to quantify how many doors it is? Break it down, I guess.
3:14:24
But you do have a >> longund there's a thou I think my district's 490,000 people something like that or eligible voters. Something like that.
3:14:31
>> Couple hundred thousand. You do you know Yeah.
3:14:35
I mean you got what 300 a day. Okay.
3:14:37
You can you can through a bunch of those.
3:14:39
I mean it's huge and I'm going to be working while I'm doing this.
3:14:41
So it's like I'm going to do >> and I have a kid coming in November.
3:14:44
So I'm like let's just see what kind of noise we can make.
3:14:47
Um and uh yeah, I'm excited to uh to start knocking on doors and let's just see what we can do.
3:14:52
Technology is the only technology is the only way you get more for less. There is nothing else.
3:14:55
Yeah, you can make processes better, but fundamentally that's the only thing that where you get more for less.
3:15:00
It's the only thing that raises a standard of living for everyone.
3:15:02
So I don't know why people have such an aversion to it.
3:15:03
I'm just trying to be like this isn't left or right.
3:15:06
Let's just let's implement things that make our life better and make everyone more rich and prosperous. >> I love it. >> I love it. Yeah.
3:15:12
California, despite uh you know shooting oursel in the foot over and over and over for decades, it's still amazing place to live.
3:15:19
So uh let's just you know uh stop shooting oursel in the foot. Get some common sense.
3:15:26
>> I think we can I think we can have a bigger economy than China.
3:15:27
I mean straight up we can be let's go second. Everyone Newsome fourth. We're top four. We're top five.
3:15:32
And yeah, we are not because of you. I think we can.
3:15:33
Let's just let's do this. It's possible. Let's do it.
3:15:37
>> Uh have you met Gavin Newsome?
3:15:38
>> You got to get on the phone.
3:15:39
>> No, I have I have not. I have not.
3:15:42
>> You got to reach out to him.
3:15:42
get him get on his podcast cuz he's had all sorts of people. So probably >> I know.
3:15:46
I don't think I'm big time enough for him yet.
3:15:47
But I think now that I've done this, I'll send you I'll send him this link and he'll say, "Okay, I I know it.
3:15:53
I know what Jordan and John are doing." Okay. Yeah. Come on. >> It's time.
3:15:56
Well, thank you for doing this. It is truly a sacrifice.
3:15:58
It is uh it is going to be an absolute slog, but it's worth doing.
3:16:02
And uh somebody's got to step up and do it, and it can start in your district.
3:16:06
And uh we'll have to have you back on as as the campaign gets closer. >> Appreciate it. Appreciate it. Thank you.
3:16:13
Thank you gentlemen so much for having me on.
3:16:14
Really, really do appreciate it. Thank you so much.
3:16:16
>> We'll talk to you soon. Good luck. >> Bye. >> Uh, what a sign.
3:16:18
What a sign of respect joining with a suit with the flag.
3:16:20
I mean, he he checked both boxes.
3:16:22
Lot lot of people bring the suit.
3:16:25
They don't bring the flag. Yeah.
3:16:28
>> And uh he should get a a California flag up there, too. That >> that'd be great. Yeah, the bear.
3:16:30
We love the California flag. >> Uh, what else we got? We got some time.
3:16:34
So, uh, Michael Truel, the founder of Cursor, uh, released a very cinematic kind of like a podcast conversation with Patrick Collison, the founder of Stripe, and it's very cool for a few reasons, but it has the aesthetics of the Johnny IV Sam Alman video.
3:16:55
>> So, you just immediately assumed?
3:16:57
>> Immediately, I was like, "Oh, there's a deal.
3:16:58
They're acquiring Cursor."
3:16:58
And I don't, that's not the news.
3:17:00
I don't even think there's a rumor, but there's something about when you film two CEOs walking into a coffee shop with cinema cameras.
3:17:09
I just permanently think acquisition now thanks to Sam Alman.
3:17:12
Um, but I didn't get a chance to listen to this whole thing. I'm definitely going to.
3:17:18
It's going to be it's like a full hour, 45 minutes or something. Uh, seems very cool.
3:17:22
I love these kind of conversations and I and I think this is like an interesting untapped media product where it's not cursor starting a podcast.
3:17:30
It's just a one-off really cool thing between two CEOs and and they're >> Yeah.
3:17:37
They're not making a commitment of putting out a video like this once a week.
3:17:41
No, it's just like, you know, a couple times a month. >> They're Yeah.
3:17:43
They're just like both technical CEOs.
3:17:45
And so you look at the list of topics, it's like uh writing your first startup in small talk, lisp, chat bots, uh Brett Victor and Dynamic Land, Coinbase's a codebase's big bang moment in MongoDB rewriting Stripe.
3:17:58
How do you, Patrick Collison, use AI?
3:18:01
Uh, changes to GDP to uh slashtotal factor uh productivity unexpected beneficiaries of AI.
3:18:08
So, they're just having like the conversation that they probably have when they just hang out and get dinner together, but they're just recording it with cinema cameras and great audio in a coffee shop and then just putting it on the internet.
3:18:21
And I think this is a very interesting new me new form of like going direct.
3:18:25
It's like it really helps me understand who Patrick is and who Michael is.
3:18:29
And it's and it's in this like they're both insiders.
3:18:31
They can just have the conversation they want to have.
3:18:35
Feel like this is great for both of them for recruiting. I don't know.
3:18:36
I was just like I I haven't watched the full thing, but it just felt like a cool new modality of like content in the startup world.
3:18:45
So, um, aside from the confusion about potential acquisition announcement, it it it seemed it seemed really really cool.
3:18:54
Uh, and and I was excited to see it go out.
3:18:55
So, go give it a watch if you're uh if you're looking to uh to spend 45 minutes hanging out with two uh absolute uh absolute amazing CEOs talking about uh AI and uh coding.
3:19:10
In other news, uh Andrea Horowit's managing partner Scott Capor has been traded to the US Office of Personal Management. >> No, of course not.
3:19:22
Obviously, um this I'm sure had been in the works for a while.
3:19:25
Uh he was announced I think it was announced I think he passed uh confirmation or something but very good news for Scott. Congratulations.
3:19:33
>> Um going into the government is not >> very sexy. It's not high status. >> Yep.
3:19:40
>> Uh but uh like Dick just said, it's uh important for people to step up and make sacrifices to >> serve their city, county, state, country, etc. >> Yeah, we applaud him.
3:19:50
Uh the AI war has opened a new front on Wall Street.
3:19:54
We've talked to a number of founders that are building AI tools for Wall Street.
3:19:57
And today anthropic giant AI giant anthropic announced the launch of Claude for financial services.
3:20:04
Finance is a natural fit for AI given the reams of structured data says Nick.
3:20:09
Uh the industry processes the big platforms are focusing more and more on fintech products.
3:20:15
Enthropic already signed Bridgewwater and S&P Global is here for Enthropic.
3:20:17
Uh and we saw earlier that Devon got hired at Goldman Sachs. You love to see that. >> City. Was it city?
3:20:25
Both and Goldman >> both. And yes.
3:20:27
So, uh so, uh Kaplan uh the president of Cognition, I believe that's his title.
3:20:33
Uh he said like this was the news that we were hoping to launch on Monday.
3:20:36
We had this all queued up that we did this big deal with city.
3:20:40
We're very excited about this, but uh no one saw it because we uh were, you know, acquiring the current thing and and that became the story.
3:20:48
Uh but that was very exciting.
3:20:49
So um all the big AI companies are figuring out how they can plug in to the big Wall Street companies >> and there's some more there's going to be some big uh launches uh and announcements this week from uh some of these labs and we will be covering them as soon as uh they're ready to share more. Yeah.
3:21:08
>> Uh uh TBPN was in the ringer this morning uh talking about >> bakes talking about uh the uh AI trade wars and tying them into sports which we know nothing about, but uh it was cool to see >> uh Katie break that down.
3:21:23
And then in other news, uh Shane Copelan uh over at Poly Market says, "Eight months ago on election night, we were on top of the world after Poly Market called the election.
3:21:34
Eight days later, the FBI broke down my door at 6 am and took all of my computers and phones, looking for anything that could imply foul play.
3:21:42
While traumatic, it etched the story of Poly Market's accuracy and the ensuing resistance into the history of American politics.
3:21:48
And today, I'm happy to announce that this chapter of the story is over.
3:21:52
After cooperating and engaging, we've been cleared of any wrongdoing. Justice prevailed. God bless America.
3:21:58
And of course, Poly Market Probe ended by Department of Justice and win for crypto bets under Trump. So, >> congratulations.
3:22:04
>> Uh, very cool for Shane.
3:22:04
Um, I think we were >> live recording when uh the news broke that he had been raided, >> but we it was we were not live yet.
3:22:15
>> Like, we weren't actually live.
3:22:15
So, we were recording and we got off the show and we looked and we were like, "Wow, that's insane."
3:22:20
>> Well, I have one last post.
3:22:20
Do you have anything else you want to go through, Jordy, or can I close it out?
3:22:22
Uh, a councilman has been charged with petty theft after allegedly removing a Palunteer advertisement from the baggage claim area at Freiedman Memorial Airport.
3:22:34
Uh, Trip Hutchinson pleaded not guilty.
3:22:36
If convicted, he will face up to one year in county jail and a thousand fine.
3:22:40
>> Wait, so did he take this because he wanted it for his garage? >> Game worn.
3:22:42
It's a game worn out of home ad.
3:22:44
And he must be an out ofome ad fan.
3:22:47
It must be a huge Palunteer fan.
3:22:47
And uh I think it's a I think it's probably a just a honest mistake.
3:22:52
Um but uh I I I think all these great out of home ads, they should find a resting place with they should be auctioned off and you should be able to get them in them on our walls.
3:23:03
So if you run a great out of home ad put it up in the airport, you're you're you're taking it down. Send it to us.
3:23:11
>> Throw it up here in the Ultra Dome. >> In the Ultradome.
3:23:13
And we will see you tomorrow in the Ultra Dome.
3:23:15
I cannot wait >> at 11 a. m. Pacific sharp.
3:23:17
Five stars on Apple podcast and Spotify.
3:23:20
And thank you for watching today. Talk to you soon.