0:00
Heat. Hey, Heat. Yeah. Heat. [music] Heat. Heat. N. [music] Heat. Heat. [music] Heat. Heat. Heat. Heat. N. Heat. Heat. N. [music] Hey, hey, hey.
Search this video
3:15
We came to [music] this world to reach the stars.
3:25
We came to this world to see our [music] future.
3:36
We came to this world to reach the stars [music] to shape our future.
3:52
We came [music] to feel the music.
4:23
>> [music] >> I feel [music] sick.
4:40
>> [music] [music] [music] >> Team Deathmatch. >> You're watching TVPN.
5:01
>> Today is Tuesday, December 2nd, 2025.
5:01
We are live from the TBPN Ultradome, the Temple of Technology, the Forest Finance, the capital of capital. ramp. com. Baby, time is money. Save both.
5:12
Needs to use corporate cards, bill pack, accounting, whole lot more. All in one place. >> That's right.
5:19
>> Why is no one talking about Armen Papa?
5:25
>> He's the CEO of Rhinel.
5:25
Uh, and they've been on an absolute tear.
5:28
We're of course going to get to Code Red.
5:30
We're going to talk about OpenAI.
5:31
Uh, but we talk about OpenAI every day basically.
5:36
and I thought it'd be interesting to uh meet the CEO behind the world's fastest growing defense company.
5:40
It's on the cover of the business section of the Wall Street Journal.
5:44
Uh when I think high growth defense companies, I usually think Ander uh or you know, Seronic or there's so many other companies that are growing very fast in defense tech.
5:51
Uh Ryan Mal has been on an absolute tear.
5:56
They're now basically the same size as Loheed Martin and General Dynamics.
5:58
Uh and it was a small company just a few years ago.
6:03
So Ryan Matal, they they make you can see the the gun that they make in that picture.
6:07
Uh they make massive cannons.
6:10
Uh they make artillery shells.
6:12
Uh they've been very important to the Ukraine war.
6:14
Um so in the last three years, they've been on an absolute tear.
6:18
They've gone from roughly 5 billion in market cap three years ago to $80 billion in market cap. We got to ring the gong.
6:25
We got to warm up the gong.
6:27
Ring the 80 billion market cap.
6:32
They've been on a tear, but they had uh and there's been like three there's been basically three key drivers to the growth to the story.
6:40
Uh we'll tell the story in in three acts as briefly as we can.
6:42
Uh and while we while we do, we will say thank you to Gemini 3 Pro.
6:47
Threeact story about Rain Mal.
6:50
Threeact Gemini, Google's most intelligent model yet.
6:53
State-of-the-art reasoning, next level vibe coding, and deep multimodal understanding.
6:56
[music] So, first they had a head start.
6:59
This company, they actually started over a century ago, 1889. Can you believe that? Very, very old.
7:08
So, uh, they spend their first 25 years basically just stacking up ammo for the German Empire.
7:15
This obviously comes to a head in 1914 when World War I breaks out.
7:20
And at the time, the company was one of the largest arms manufacturers.
7:25
Like, they were they were pretty pretty big after 25 years of just stockpiling ammo, growing, growing, growing as a defense company.
7:30
World War World War I breaks out.
7:32
Uh but then after the war, they got to pivot.
7:34
They got to pivot because the Treaty of Versailles forces them to switch to nonmilitary products.
7:40
They say, "Hey, you got to build some cars. >> Make some trains."
7:43
They get fixated on trains. >> And also typewriters.
7:47
>> Not the first dudes to get fixated trains.
7:52
>> Happens to the best of >> But they but they have a good run. They stay in business.
7:54
They keep making trains, locomotives particularly.
7:55
Uh you know, they're making big stuff.
7:58
And then 20 years later, it's uh the mid30s.
8:00
20, it's uh 1935 around there.
8:04
Uh they are uh they're starting to get back into weapons and ammo production. They can't stay away.
8:12
[laughter] Uh oh, who are they rearming? Uh the were mocked.
8:14
Uh and World War II, obviously, it's massive for production. They're printing.
8:21
They're making lots of weapons.
8:22
Uh but by the end of the war, their facilities have basically been destroyed by areas that need to rebuild the company from scratch.
8:28
So after the second war, they get banned from making weapons again until 1950. >> Keeps happening.
8:35
>> And so they have to go back to making typewriters.
8:37
They keep getting relegated to type like you guys no more guns. >> That's enough.
8:42
>> You have to make some typewriters.
8:42
Uh and so they get back into defense tech in the ' 50s60s.
8:47
The German armed forces gets reestablished in 1956.
8:49
Uh, and by 1979, Ryan Matal is making 120 millimeter guns that go on leopard tanks that you've probably seen in that image roughly.
8:59
Uh, and so there's lots of M&A, lots of diversification over the next few decades.
9:02
They expand into automotive and electronics.
9:05
And that kind of brings us to the second act of the story, which is the Ukraine war.
9:08
So, Russia invaded Ukraine on February 24th, 2022 about three years ago. uh Reinhal was around 5.
9:18
around 5.5 billion market cap then uh and 3 days later Olaf uh Scholes the chancellor of German Germany gives what's known as the Zitan Wald Zitenwend speech which is literally translates to turning point so he says this is a turning point Europe has been invaded we
9:37
now have a foreign army on European soil even though Ukraine is not part of NATO it feels like you know Russia is expanding if they keep if they just keep going in the same direction they're eventually going going to be in our hometown. So, we got to do something
9:47
So, we got to do something about it.
9:49
And what does he propose?
9:49
He doesn't just say, "Hey, this is a big deal."
9:52
He says, "No, we're actually going to invest a hundred billion dollars like offbalance sheet from some fund into defense tech.
9:57
We're going to spend more money."
9:59
And then, of course, there's a whole bunch of other initiatives that happen.
10:01
There's the Trump negotiations around how much Europe should pay as a portion of GDP on um on defense.
10:08
But basically, it's this major turning point where Europe goes from spending, you know, sustainment levels.
10:14
Okay, we're going to spend this much every year to we are going to double or triple or you know exponentially grow our spending and it's all going to be net new so you can go and fight for it and that's what Ryan Matal does and so revenue >> was sort of born out of that era.
10:30
>> Heling is like the newer version of Ryan Mattel.
10:32
Ryan Matal is like the old you know roll up.
10:34
It's been around for over 100 years.
10:36
Uh, Heling I think started >> Helling was 2021 >> uh was most recently in the news because they raised I think $600 million from >> Daniel. Yeah.
10:47
>> Sparked uh controversy.
10:47
Of course, a lot of people in uh the Spotify world of music just think that defense tech is def.
10:54
>> Oh, I didn't realize that.
10:54
There was actually backlash. Huh. >> Totally. I didn't see that.
10:57
uh you know, if you're a an artist and you uh you know, believe in in peace at all costs, you're going to >> uh probably be against that.
11:06
But uh >> well, it depends.
11:08
Maybe if you're, you know, uh pod or you're uh you're you're you're some other uh you know, musician that was played during the war on terror, you could be very pro the Helsing investment. Just depends. But yes, I understand. uh overall.
11:23
So, uh revenue's grown 50% uh since 2022 and they are now guiding for sales, I think they do maybe around like 10 billion euros.
11:32
I was kind of going back and forth on euros USD, but uh they're guiding for sales of 58 billion and an operating margin of more than 20% by 2030.
11:41
So, they have like almost AI growth level numbers of everything.
11:47
It feels very similar where there's a there's a structural change in the way their business is going to work. Same thing as Eli Lily. Same story.
11:55
There's a couple of these stocks where there's now sort of a mega trend and they are in position to capture a ton of value as long as they can execute.
12:02
The the big question is, you know, what winds up happening.
12:06
But the third leg of the stool, the third important piece in this story is the current CEO, the man no one is talking about until today, Armen Papa.
12:14
Uh he's been called a white-haired Goliath. I love that.
12:18
CNN randomly threw that up >> a picture.
12:24
>> Just randomly threw that in. >> There he is.
12:26
>> There there's some other photos.
12:26
Um and last year he was targeted in an assassination plot by the Russians. >> What?
12:33
>> So the CNN reported that Russia had made a series of plans to assassinate several defense industry executives all across Europe uh who were supporting the Ukraine's war effort.
12:43
And they also were planning to uh set up fires in different uh there was an IKEA that got lit on fire.
12:51
There were a number of different attacks.
12:52
Um but fortunately, American intelligence discovered the plot and informed Germany in time to stop the attack.
12:59
And now uh the white-haired Goliath is Ryan in the chat says this feels like a paid ad.
13:04
I can assure you it's not.
13:06
John woke up this morning.
13:06
we were at the gym and he's like, "Why is no one talking about Rhymel?"
13:11
>> And uh decided to write about it in the uh in the in the newsletter today.
13:14
So that's >> no of the Wall Street Journal.
13:20
>> Um and so uh >> uh so so they stopped the attack.
13:23
Uh and so Russia clearly sees uh Armen Papinger Paper as a crit as critical to the European defense ecosystem.
13:33
But separately there is a debate over how like where the business goes over the next few years because on the one hand like the NATO inventory requirements are growing a lot.
13:43
That's going to drive a lot of net new demand for military equipment purchases.
13:46
Uh and the market's been historically under supplied.
13:48
But on the flip side, Ryan Matal may or may not be able to absorb as much of the demand as they're planning to.
13:55
They have lots of integration to do between all their different acquisitions.
13:58
Uh and also with a potential potential end to the Ukraine war. >> Yeah.
14:03
It feels like the stock would just immediately trade down on news of a peace deal.
14:08
>> And it has even on rumors of a peace deal. Yeah. Exactly.
14:09
Um >> yeah, it's down 15% over the last month. >> Yeah.
14:14
Um but uh they're they're scaling up.
14:17
in the Wall Street Journal says earlier this year, Armen Paperger uh opened a new factory that will allow his company to produce more of an essential caliber of artillery shell than uh the entire US defense industry combined.
14:27
Uh surrounded by that day by dignitaries including the head of the North Atlantic Treaty Organization NATO, the Ryan Matal CEO is riding a wave of postcold war military spending that is reshaping the global arms trade.
14:41
Ryan Matal is now the world's fastest growing large defense company and a key player in Europe's quest to rearm its home country.
14:47
Germany is shedding its post-war reticence on military spending to lead the charge to capitalize.
14:53
Uh Paper has pushed the once obscure gun barrel maker into almost every part of the battlefield from satellites to warships.
15:01
And that's what people are kind of saying about, oh, there's a lot of acquisitions, there's a lot of new projects, there's a lot of new deals.
15:08
Like, you know, the gun barrels, they've been doing that for 136 years.
15:13
Satellites, they're kind of newer to it.
15:15
Do they have the lineage?
15:15
Do they have the experience?
15:17
Can they stick the landing on those contracts?
15:19
This the money is certainly there, but is the expertise there? That's the big question.
15:24
So, uh, his goal is to create a go-to defense company with the heft and breadth to rival the American giants that have dominated the industry since World War II.
15:33
And if he's writing any software, he's got to get on graphite.
15:37
dev code review for the Age of AI.
15:39
Graphite helps teams on GitHub ship higher quality software faster.
15:41
Uh Rhinel's stock is up 15x since Russia's full-scale invasion of Ukraine in 2022, giving it a market cap of uh 80 billion, roughly on par with US rivals.
15:52
And when he took over the job, he started this job in 2013.
15:57
So he's been CEO of Ryan Matal for 12 years. Uh the company was 1. 6 billion. >> Overnight success.
16:07
>> Overnight success is right.
16:07
Sort of uh like what happened with Lisa Sue, you know, she's 10xed that stock.
16:10
I mean, the funny [laughter] thing is is that Jensen's also 10xed Nvidia in that time, but uh the Lisa Sue story is a little bit more impressive because AMD was really like down in in the dumps and and she has turned that company around fantastically.
16:24
Uh but back to Ryan Mattel.
16:26
Uh this month Ryan Mattel set out ambitions to quintuple sales by the end of the decade uh to the equivalent of roughly 58 billion.
16:33
There's Paperger reflecting on his long tenure at the company told investors that seeing such figures was like a wonderorld. It's a wonder world.
16:43
I love when a executive is speaking a different language and it just doesn't quite translate like is that what we say? >> Kind of get the gist. >> I get the gist. He's happy. I'm happy for him. You know, good job.
16:55
Wait, we had a we had a a buddy of ours who uh actually I'm just gonna I'm gonna name I'm gonna name uh I'm gonna name him.
17:02
I was going to keep him anonymous, but it's just too funny. >> Somebody's propos 2.
17:06
6 is proposing the TBPN X Standard Oil X Ryan Matal collab.
17:13
Uh we so we were talking we were texting with Sean Frank and Conor McDonald at the Ridge yesterday about how their Black Friday >> Cyber Monday went and they shared a bit on it and Sean ends it and says, "Bro, the future is beautiful and I am so happy to be alive with like incredible reaction to a successful Black Friday."
17:34
>> I'm so glad it went well for them.
17:34
I mean the new the Wall Street Journal did report that on a busy Cyber Monday outage at Shopify halts transactions.
17:42
Shopify experienced an outage on Cyber Monday that interrupted transactions for some merchants, but it sounded like Ridge was not affected.
17:48
>> Well, so it was uh that the the real issue is the admin panel went down, which freaked a lot of people out because you're not able to log in and like see what's happening. >> Yeah, of course.
17:57
And and also like even if even if everything's working as as standard, as expected, like that's the day you're just refreshing the admin panel all day, like because you're just like, how much money am I making?
18:06
Like this is really critical, right? >> Yeah.
18:09
Um, so yeah, there were there were some some reports that that a handful of merchants had actually had features on their site go down, but I didn't see any.
18:17
>> I didn't see a ton of people saying like, I lost my Cyber Monday.
18:19
But obviously, um, the the good folks at Shopify will obviously be working extra hard to resolve any of this and and provide a proper postmortem.
18:28
Uh, overall, it does seem like Cyber Monday and Black Friday, Black Friday, Cyber Monday broadly, uh, just went very well.
18:37
Like it just s it just seems like consumer confidence was up, revenue was up, spending was up. >> Harley had a post.
18:42
He said total global Black Friday Cyber Monday sales by Shopify merchants over the last 5 years. 2021 was 6. 3 billion, 2022 was 7. 5 billion, 2023 was 9. 3 billion, 2024 11.
18:52
5 billion and then 2025 14. 6 billion.
18:56
So um combination of uh execution at the at the company level and execution at the uh Shopify level and then obviously the market plays a big role as well.
19:11
>> Yeah, it it it really did seem like uh like things um are just broadly going well or at least okay.
19:17
I think everyone's sort of like nervous with with crypto up and down and uh is there an AI bubble and how big of a bubble what will happen?
19:27
somewhat of a code red going on.
19:29
It >> has been a code red part of the world.
19:31
Uh before we jump into that story, >> profound get your brand mentioned in chatbt reach millions of consumers who use AI to discover new products and brands.
19:39
>> I was going to say um uh kind of the uh Anderoll was covered in the Wall Street Journal.
19:44
The Wall Street Journal has been been doing quite a lot of defense tech coverage.
19:48
Ander they had been talking about their approach of not using government funding for testing purposes which historically >> company would get a contract and then they would work to actually make it and the government was effectively funding R&D.
20:03
>> Anderoll has a more traditional like venture style model where they raise VC dollars.
20:07
They spend that money to >> uh test and develop products.
20:11
And so they gave a quote that was like we do fail a lot, but it the fail the extra context that was necessary was that that's not happening on the on the taxpayers's dime. Yeah.
20:24
And it's part of their, you know, approach of doing rapid iteration uh and sort of like going according to plan.
20:31
Of course, that was taken out of context and turned into a headline that was we do fail a lot.
20:36
uh and not so uh dissimilar from what has happened to openai in the last 24 hours where uh sounds like an internal uh internal meeting uh was leaked.
20:47
We are wondering like what is >> we can yeah let's let's actually let's actually read a little bit more on the on the Wall Street Journal uh and thing because I think it's interesting and I want to go into uh some of the response like how they responded to this.
21:00
First, let me tell you about Cognition.
21:02
The team behind the AI software engineer Devon crush your backlog with your personal AI engineering [applause] team.
21:10
Um, so, uh, Wall Street Journal came out with this story.
21:13
We do fail dot dot dot a lot. It's a very funny quote.
21:15
I actually think it's an awesome quote. We'll get into it.
21:19
I think that they should put on t-shirts and hats.
21:20
Like, I think it's actually a very next campaign.
21:25
>> I think it's a very key cultural I think it's like a don't work at Anderrol type moment.
21:28
It makes a ton of sense in terms of like the the culture of like fail fast.
21:32
This is this is not new in Silicon Valley and and yet it's still being reframed as new.
21:37
It's I'm surprised that there's like alpha here still. But uh let's see.
21:41
Palmer Lucky says the valid reasons for slowness are bureaucratic BS and cowardly executives who cater to snide analysts and public market outlets like WSJ that have nothing to say about years late programs and everything to say about a fire that covered 0.
21:56
00 0002% of our test site.
22:00
I'm not even exaggerating. That's the real number.
22:01
Uh it is exactly what anyone would expect from testing a system that violently blasts lithium powered drones out of the sky.
22:10
This is what weapons development should look like.
22:12
Heck, Camp Pendleton has over 200 fires per year on their training range.
22:17
And that is with fully mature weapon systems.
22:19
Going on and on about this for paragraphs is so pathetic.
22:24
>> Getting close to a fire every single weekday.
22:26
Yeah, they obtained satellite imagery that reveals the damage to the grass on the weapons test range.
22:34
[laughter] Uh the other the other examples of the story are similarly absurd.
22:39
>> John, you're telling me that the grass was damaged at the explosives testing ground?
22:43
[laughter] >> Yeah, it tell you're telling me there was an explosion at the weapons testing facility. >> Yeah, it's wild.
22:51
The uh the other examples in this story are similarly uh silly absurd.
22:56
Oh no, an engine sucked in a piece of fod. Stop the presses.
22:58
Uh autonomous boat behaves exactly as designed and stops moving when it receives a faulty command and are all hit by a pattern of setbacks. It's just so pathetic.
23:07
The type of thing that can only be written and taken seriously by people who have no idea how hardware development actually works.
23:12
And of course, a few other folks in the in the ecosystem uh um chimed in mainly.
23:20
There's a good post here from Blake Scho, founder of Boom Supersonic.
23:22
He says, uh, "If you plan to pass every development test, you'll move slowly and expensively.
23:29
It's optimal to fail many dev tests.
23:32
Selective quote outtake into headline suggests a hatchet job, not an honest report on an attempt to do things differently and better." Um, yeah.
23:39
I I still just think the uh the we do fail a lot is just it's so ripe for a billboard campaign, a t-shirt, a hat or something because uh it if you like the whole thing with Silicon Valley is that you should fail 99 times and succeed once because if you succeed once and fail 99 times, >> it's a million times better.
24:01
It's infinitely better than zero zero failures, zero successes.
24:06
like like you will take a ton of failure for one success.
24:12
And that's the whole that's the whole ethos.
24:13
Um >> that's the American ethos.
24:16
>> That's the Yeah, it's the American ethos.
24:17
It's the technology ethos.
24:17
It's there's a lot there.
24:19
Um anyway, um back to uh Oh, actually, right.
24:23
We we we can wrap up with uh you can go read the Wall Street Journal report on Armen Papager if you if you want.
24:31
We got to figure out how to pronounce his name.
24:33
He did have one uh one fun line in here uh which was uh what did he say?
24:38
He said something like uh uh he said um referring to so now he's uh you know basically the same value as Loheed Martin and General Dynamics and he said uh on the US companies he said they come to me 10 years ago it was a different story and so he's just flexing the fact that like he's now big enough that he he deter he uh he requires like you can go visit him because he's like made it >> always a good sign.
25:05
>> He yeah he he's uh he's he's taking a little victory lap.
25:07
Uh and there's some other funny things in here.
25:09
Uh but you can go and read that.
25:10
Uh let me tell you about linear.
25:12
Meet the system for modern software development.
25:14
Linear streamlines work across the entire development cycle from roadmap to release.
25:19
Um so let's head over to red alert territory.
25:23
Gavin Baker uh responding to the reporting says October 1.
25:29
4 trillion in spending commitments. November rough vibes. and December code red. Life comes at you fast. >> Code red.
25:39
>> Um, it it certainly has felt it certainly has felt fast ever since >> that faithful podcast.
25:47
>> Yes, that was a crazy turning point.
25:49
>> Although there was there was plenty of conversation, you know, prior to that around um >> yeah, >> around uh uh of what the trajectory of OpenAI would actually look like.
26:01
>> Yeah, it's hard to actually understand the full nuance here.
26:03
the full nuance here. Somebody in the in the replies a rational analysis insane at insane analyst what a crazy handle says debt obligations come at you fast and it's like that's not really what's happening here like like the the the code red like leak from this the the
26:21
information reported it was clearly like some sort of all hands that Sam Alman was uh you know holding a town hall with the rest of the open AI team and he's kind of just saying like lock in that's what he should have said never say code >> [laughter] >> You got to say lock in, brothers. Lock Lock in.
26:36
>> Don't say rough vibes.
26:37
>> Don't say rough vibes. >> Code red. >> Say lock in.
26:40
Say we're we're taking that hill.
26:42
We're storming their fortress.
26:42
We will grind Google Gemini team into paste with and we will crush our enemies. We will see them driven.
26:52
>> Hospitals learned this lesson.
26:52
They used to say code red.
26:53
That meant there was a fire in the hospital uh and that you would probably want to figure out a way to get out, right?
27:00
Uh they started is it code blue? >> Yeah.
27:04
Now now they will say code blue.
27:06
So if you hear code blue in a hospital you need to be worried. >> You need to worry.
27:11
But maybe maybe okay steel man. Steel man here.
27:13
Maybe Sam Alman was using code red in the hospital sense. He didn't say code blue.
27:18
If he had said code blue, we should be really worried. >> But he said code red.
27:24
So he's saying it's not that bad.
27:27
>> But don't you think they just retired?
27:29
They I think they just retired. >> Yeah.
27:31
So he's saying I'm I'm using retired phrase.
27:33
I'm not I'm not saying code blue.
27:35
If I if I was saying code >> use code brown, which is a hazardous spill. >> Okay.
27:41
>> Which Gemini 3 spilled on timeline. It's very hazardous. We got a code brown.
27:46
>> Yeah, we got a code brown.
27:47
>> Um >> that's a crazy Is that real or is that some like meme joke? >> No, this is No, no.
27:52
I'm reading the hospital emergency code. >> Okay. Okay.
27:55
Well, anyway, uh let me tell you about Reream.
27:58
One liveream, 30 plus destinations.
28:00
If you want to multiream, go to reream. com. Ching.
28:05
Um, no, I I think I think if you're if you're if you're a CEO who's under incredible scrutiny, like you're Sam Alman, and you have beat reporters at this point who are texting your employees every single day, hey, what's going on? What's on the ground? Give me a quote. What happened? >> Yeah.
28:21
>> Yeah. So to give people to give people context there the beat reporter a beat reporter might reach out to they will actually adopt the strategy of just trying to wear someone down >> where they will send hundreds of messages >> to individual people on on the team just
28:38
over and over and over relentless like email cell phone Instagram DM LinkedIn just like constantly constantly constantly flooding hoping that at some point this person just says like fine like I'll I'll Well, the name beat reporter comes from them trying to beat you down. >> That's the whole point. Is that
28:54
>> That's the whole point. Is that >> true?
28:56
>> That's where it comes from. >> No way. [laughter] >> You're me. Are you messing with me?
29:01
>> Yeah, I'm messing with you. >> Okay. Okay. [laughter] Okay. >> I have no idea.
29:03
But I like the idea of it.
29:05
It's like they just try and beat down new employees.
29:07
>> It's like that >> they try and beat you down.
29:09
>> I got a beat reporter on my team on my tail. >> Yeah. >> Yeah. No.
29:13
Um >> um [laughter] I mean it's certainly what it's what it's become.
29:18
>> There is a little bit of it. No, no.
29:18
I think uh there's beat reporting, there's gum shoe reporting.
29:22
Gum shoe reporting is where you report uh and you're you're actually walking around the town so much that you get gum on your shoes. That's the idea.
29:30
It's like you're on the ground reporting, you're walking around the city, you're getting you're talking to people.
29:34
And then I think like a beat cop and beat reporting is like you're on a beat like it's a drum beat.
29:39
Like every day you report on the same thing.
29:41
And so it's about consistency.
29:43
It's not uh it's not [laughter] What are you laughing at now?
29:48
Uh Ry in the chat, if you work at an AI startup and you aren't drinking Mountain Dew Code Red every day, you aren't going to make it. What if Sam was talking?
29:55
What if he was just saying, "We got to lock in. >> We got to lock in.
29:58
I bought us a bunch of >> Code Red." Yes.
30:01
>> I want you all drinking it every day. >> Yes.
30:04
>> It's time to really focus. >> Yes.
30:06
>> And and of course that snippet got pulled out.
30:08
I just want to know what what what person on the on the OpenAI team thinks it's in their best interest to be in a meeting like that and then just go share inflammatory quotes on said meeting. >> Just leave.
30:22
Just just go make 10 times as much money at a different lab.
30:25
You know, if you if you don't like your employer, just bounce and make more money.
30:30
[laughter] Like why are you why are you sitting there leaking and and just dragging your company down?
30:34
Don't you have stock options?
30:38
Yeah, that's I'm so I'm so confused. >> It's a mole. There's a mole.
30:40
There's someone inside the organization who's working against them or something. I don't know. Seems rough.
30:45
Uh anyway, uh there is some praise for OpenAI on the timeline, which we should get to from none other than Blake Robbins.
30:52
Blake says OpenAI is operating on a different level.
30:57
Play a s that sound cue, Jordy.
31:02
The amount they have shipped in the past few weeks and months is incredible.
31:05
Feels like we are witnessing a generational run. This was on October 6th. >> Okay. This was on October 6.
31:10
Sora was, I think, number one in the charts at that point. >> Yes. >> It's now 21. >> Yes.
31:18
>> Uh Pulse was got some excitement early on, but uh >> I think people are a little bit >> not feeling like as excited. Yep.
31:28
>> Atlas launched and then >> uh it's hard to really gauge what what adoption has been like.
31:33
I I I know some people that that love it.
31:36
Y >> um but uh >> so Eric Sufort on October 6th quote tweeted Blake Robbins and kind of summed it up.
31:45
Uh I think he said indeed impressive, but the scattershot nature raises questions about the company's discipline and ability to support these desperate initiatives.
31:55
Is OpenAI a frontier research lab, a social network operator, a commerce engine, a hardware company?
32:03
because it's hard to do all of that well.
32:05
And then Eric goes back and finds his old >> and they're trying to they're still like very much care about competing in codegen, >> right?
32:16
And so so if you go back if you go back to uh the BG2 interview or just the BG interview, >> um Sam Sam's answer >> to the question of how are you going to support the 1.
32:27
4 trillion of commitments was we're automating science.
32:30
and we're making >> and we're make and we're making like consumer electronics.
32:38
>> And the reason that that that didn't um >> to me that that was kind of like a concerning answer because Google has been doing those things for years. >> Yeah.
32:48
But they've earned the right because they have 25 years >> funding it with massive cash flow. >> Yeah.
32:53
>> Yeah. hundreds of billions of dollars in revenue and and so much cash just to go around and like it's always been this like academic lab and this sort of like environment where they do side projects but they've just they I think before they started any of that they had f
33:08
firmly established themselves as like the go-to um search engine and they >> they were funding that with cash so they were funding these initiatives with cash flow I believe so >> and even though they've been doing it for this long it's not like Sundar is going out there and saying Guys, we're actually going to do it. We're going to
33:23
We're going to do an extra hundred billion next year because we're automating science and we're >> uh we're doing this this new uh consumer electronic device. >> Yeah.
33:33
No, no, it it is it is crazy. Uh let's continue.
33:35
First, let me tell you about Privy.
33:37
Privy makes it easy to build on crypto rail securely, spin up white label wallets, sign transactions, integrate onchain infrastructure, all through one simple API.
33:43
Uh I have I have a plan, and this comes from the chat, of course.
33:47
Uh, if Sam Alman really wants to set the record straight, everyone's saying Code Red, oh, Code Red, it's so bad.
33:54
He needs to come out with a statement.
33:56
We're going to Baja blast Gemini out of the App Store.
33:58
[laughter] If he says, "Our plan is to Baja blast Gemini and Anthropic into the minor leagues of AI research," uh, I think he just wins completely. What do you think?
34:14
I I think people are they're underestimating the the possibility that um code red it was actually read was past tense of read and they're talking about the code that was read by the model.
34:23
[laughter] >> Oh yes, >> what was the code read this scenario?
34:27
>> It might have been re agent model that was I have read the code and and we're ready for the next uh pre-training run.
34:34
Uh I listened to Mark Chen on Ashley Vance's core memory podcast. It's very good. You should go listen.
34:41
Also, Ashley has a new a new YouTube channel for Core Memory Podcasts, so if you want to find it, head over there.
34:48
Um, and uh, and it was interesting.
34:51
Uh, Mark Chen, uh, I really like the way he runs that organization.
34:55
I liked a lot of things he had to say.
34:56
He had some funny funny takes, uh, some funny anecdotes.
34:59
Basically just saying, you know, he's extremely competitive. He doesn't want to lose.
35:03
He's he's he's, you know, going all out right now.
35:05
and that one of the ways he's dealt with the talent wars is to just go to everyone on his team and say, "Hey, uh, I'm not going to match dollar for dollar with Meta."
35:19
Like, if you want to make 10 times as much money, yeah, you're free to leave.
35:23
Like, you can just go, but we are on a mission here.
35:24
We're a team and we think what we're building is so big that in the long term, we will be we will be better and we will be bigger.
35:32
And he also clarified interestingly that although there was a big raid and a lot of people from OpenAI did go to Meta.
35:38
Uh he was saying like there's been there's been he was basically like there's a lot of poaching that's happened from OpenAI generally like whenever someone starts a new lab they always go to OpenAI they're like we need at least one OpenAI guy to know how they do it right makes a lot of sense.
35:54
Uh he also said he didn't lose a single direct report.
35:56
I don't know exactly how many direct reports he has, but he was saying that he didn't lose a single direct report.
36:01
So maybe that's like maybe he's trying to say, "Okay, there were people that were too low down." Yeah. Yeah.
36:06
His his his lieutenant stuck around.
36:08
Uh it was sort of interesting, but he did say that he also uh he he sort of echoed Scholto and said that he believes that pre-training there's still lowhanging fruit there that OpenAI will be doing new pre-training runs that they have seen uh that scaling is holding that there's no plate.
36:29
>> He also said they have models internally that outperform Gemini on benchmarks. >> Yes.
36:35
And obviously he caveed that by saying benchmarks aren't the only thing that matter.
36:41
>> So I do think I mean it's it's worth sharing.
36:44
>> Let's play let's actually play this clip from Ashley Vance here.
36:46
Uh says OpenAI has seen Gemini 3 and is both moved and not.
36:51
We sat down with OpenAI's research chief Mark Chen 90. Is this >> Yeah. Yeah.
36:57
So um to speak to Gemini 3 specifically, you know, it's a pretty good model.
37:01
Um and I think one thing we do is try to build consensus.
37:05
You know, um the benchmarks only tell you so much.
37:10
Um and just looking purely at the benchmarks, you know, we actually felt quite confident.
37:15
Um you know, we have models internally that uh perform at the level of Gemini 3 and we're pretty confident that we will release them soon and we can release successor models that are even better.
37:26
are even better. Um but yeah again kind of the benchmarks only tell you so much and I you know I I think everyone probes uh the models in their own way there there is this math problem I like to give the models uh >> this is funny >> I I think so far none of them has quite
37:42
cracked it even the thinking models um just tell us so yeah I'll wait for that >> is this is this like a secret math problem >> oh no no um well if I announce it here maybe it gets strained on that it's going to get so saturated >> so um to speak to Gemini through specifically you know, it's a pretty good model. Um, and I think
37:57
Um, and I think >> I think this is looping. >> One thing we >> Yeah.
38:01
Um, [laughter] having having a secret math problem that you give every uh model to to assess it uh is uh is is pretty elite.
38:10
Um I I keep reflecting on like like so let's read what Prince is saying here.
38:18
So new new interview with Mark Chen from OpenAI.
38:21
Ashley Vance the interviewer has apparently been spending a lot of time at OpenAI including sitting in on meetings.
38:25
He seems to be writing a book.
38:27
Uh, and he seems to think that OpenAI has made some huge advance in pre-training.
38:30
Pre-training seems like this area where it seems like you've figured something out. You're excited about it.
38:35
You think this is going to be a major advance.
38:36
Mark doesn't spill the beans though.
38:38
He says we think there's a lot of room in pre-training.
38:40
A lot of people say scaling dead is dead.
38:41
We don't think so at all.
38:44
Big question about what that means. Is that scaling RL?
38:47
Is that scaling dollars in? Is it is it Oh, yeah.
38:50
If you if you if you invest a hundred trillion dollars, you can give it one more IQ point.
38:54
It's like, yeah, that would be an example of like scaling holding, but like no one's going to make that tradeoff.
39:00
No one no one is going to be like, yeah, I'm down.
39:04
>> Totally spend the hundred trillion. >> Okay.
39:06
So, what what Sam said in the uh internal Slack memo was >> Oh, it was a Slack memo. >> Yeah. Okay.
39:12
>> Is he was directing more employees to focus on improving features of Chat GBT such as personalizing the chatbot for more than 800 million people.
39:18
Uh and and again we we've seen them like launch more functionality around this.
39:23
Uh I think the theory is that this could be a very like make the product really really sticky.
39:30
Whether or not that's true generally is still unclear.
39:33
It's certainly uh people have have uh uh been very loyal to 40.
39:40
>> Uh Alman also said this is in the information piece.
39:43
>> Did he mention Baja Blast?
39:45
He hasn't he hasn't specifically said Baja Blast, but I think he's kind of alluding to it.
39:51
>> He's warming up to talking about Baja Blast.
39:55
>> Other key priorities covered by the code red include image gen, the image generating AI that allows users to create variety of photos.
40:01
Um, you had included in your newsletter last week that you've been going over to Gemini specifically for Nano Banana. >> Yes.
40:10
>> Um, so I wonder if this is broad.
40:10
Does this does this actually matter? I think it does.
40:14
I think that the image generation functionality like fundamentally what LLMs are doing all what these chatbots are doing is they're they're basically instantiating full web pages.
40:28
They should be able to instantiate anything that you could possibly land on whether it's a video, an image, a blog post with images embedded, an audio format.
40:37
like it should be able to to like not just understand everything and give you the answer, but it should be able to contextualize that answer in any format.
40:47
And so I do think being able to generate images at the top shelf, top tier way.
40:52
The big question we were talking to Tyler was uh should they say, hey, we're just going to use Nano Banana, which is like a crazy thing, but you know, there is a world where they say like, hey, yeah, like we're not going to focus on that.
41:04
We're going to we're actually going to just bend in Nano Banana, but we are going to be the the front door, the aggregator, and we're just going to be the the the uh the the actual >> use runway in the background, right? >> Yeah. Yeah. Yeah.
41:17
Hand it off to a different team potentially. I I don't know.
41:19
It see it seems like that's probably a little bit too close to home.
41:22
Um, but Ben Thompson has had this uh this this claim for a while that potentially uh OpenAI has has a stronghold on the consumer market to the point where if they swapped out the underlying model, they would still acrue tons of the value because people don't really know what model is which.
41:39
Like I think the average user doesn't do it.
41:43
But first, uh Tyler has a something.
41:45
>> Yeah, I mean I think that especially makes sense in the context of images and video because they're just so expensive. Yeah.
41:50
like um I think a nano banana pro image is like I think it's like 10 cents. >> No way.
41:56
>> It's really or Okay, that might be per like a thousand or something.
41:58
But it's still it's still they're really expensive and videos are even more expensive.
42:02
Videos are like really really expensive.
42:03
So I think it makes more sense in that uh scenario because um you would imagine that it's just like so expensive to to vend it yourself.
42:10
It's like you're spending so much resources on that. >> Yeah.
42:15
Um >> we have to look at this.
42:16
I I believe this is nano banana.
42:18
Let me see if I can find this uh this nano banana uh pro image that uh let me see if we can pull this up.
42:25
It's uh it's from John Gregorchuk.
42:28
It says, "Architects are cooked. AI is coming for you. Prepare accordingly."
42:34
Have you seen this, Jordy? >> I did see that. >> You did see this one? >> Yeah.
42:39
>> Did you look at the image closely? >> Uh no. >> Okay.
42:42
So, >> is it all Is it Is it >> It's one of the funniest images I've ever seen.
42:45
So basically th this image which >> it's like a it has a walkway with like a 40 foot drop to the ground.
42:52
>> I mean it's not quite that bad but it's it's close. >> Yeah.
42:55
just I just I I I didn't I I don't buy the the theory that architects are are cooked uh just because you can generate like a floor plan or uh or uh designs for a home just just because the actual process is uh you're dealing with a city basically, right?
43:17
And you're trying to get things permitted.
43:19
It's not it's not like the problem is just making pretty designs, right?
43:23
It's the classic uh let me see uh I'm trying to put this in the in the chat.
43:28
[snorts] Uh it it's the classic like uh you know is the radiologist's job just to look at images and detect cancer?
43:36
No, it's way more than that.
43:38
Okay, so this is the image and I was actually crying laughing because the uh the the the the tagline is architects are cooked AI is coming for you prepare accordingly.
43:49
And you see this and it's like this AI generated image and it looks like remarkable like >> looks like a floor plan.
43:55
>> It looks like a floor plan. It looks amazing. Like it looks like okay.
43:56
Yeah, that's like all the lines are straight.
44:00
We used to be in the era of like any any text would be typoed and there would just be crazy lines everywhere.
44:04
Uh but you zoom in and it's like one of the funniest layouts ever because you realize that it's just it's just one massive [laughter] room with with like three or four. Okay.
44:18
So, first off, okay, so you come in through the two-car garage.
44:20
Then there's a powder room.
44:22
So, so first off, there's this mud room mudroom and laundry with two bathtubs in it. Scroll up to the right. Okay, scroll. Yeah, right there.
44:32
So, why do you have two bathtubs next to your coat closet, right? >> In the mud room. >> In the mud room.
44:36
And then, and also like you can't go normally you come out of garage, you go straight into the mudroom, but here you have to go into the main area, which is the gallery hall, and then you go from there into there.
44:46
And so scroll to the left a little bit so we can see the full. >> What is the code?
44:49
So then you go there's a powder room and then there's the >> coat and two toilets.
44:54
[laughter] >> There are two toilets next to each other.
44:58
Remember we were touring that facility and it had two it had two co two bathrooms right next to each other with no line next to it or >> Yeah. Yeah. Yeah.
45:05
We were in the in the in the office in with in the crazy office that had the machine.
45:09
Uh, one of the bathrooms just had it was like a it was like meant to be a private bathroom.
45:16
Uh, and it just had two toilets there. We were like, "What? Why the two toilets?" [laughter] >> Yeah.
45:20
So, it's like So, so you come in through your main foyer, then there's a master bathroom, then there's a coat bathroom with two more [laughter] toilets, and then there's a huge walk-in closet [gasps] with which isn't even directly attached to the anything else.
45:35
So, you have to like go through this corridor to get to the rest.
45:37
And so this master suite has three toilets.
45:40
But then it gets better, dude. It gets better.
45:43
So go over to the top right hand side because this is So look at bedroom number two.
45:49
It's just like off the center.
45:51
Then bedroom number three is there.
45:53
>> Then there's a Jack and Jill bath. Then scroll down.
45:55
Then there's nothing three sinks. >> Three sinks. [laughter] Three sinks. No toilets.
46:00
And then there's another bed bathroom.
46:03
And then there's a [laughter] third a third bathroom with a toilet.
46:09
[laughter] >> This is >> with you have five sinks.
46:13
>> This is You might not like it, John, but this is this is this is architecture at its best.
46:19
>> Yeah, you have you have [laughter] five You have five five sinks next to your two bedrooms, which And then also bedroom two doesn't have a doesn't have a bed >> anything. It just connects.
46:29
It opens into the gourmet kitchen.
46:33
>> [laughter] >> But then if you scroll down, if you scroll down, you can see that there's like this huge walk guest suite.
46:38
What is the huge walk guest suite?
46:40
And then you have this like massive dining room.
46:42
That just makes no sense.
46:45
And then down at the bottom to to to kick it off, uh there's uh of course like the great room that's directly tied into the kitchen with there's just the most open floor plan you can possibly imagine.
46:54
And then if you scroll down, you'll see that there's like just these windows that like like all of a sudden >> Trevor in the chat says bathroom scaling.
47:04
>> Why why are all of a sudden the doors like vertical instead of this is supposed to be a top down image and now I'm looking at these doors and they're like present.
47:12
>> What did the comments say?
47:12
Do the comments say like, "Hey buddy, why why' you put, you know, three sinks in that one bathroom?"
47:19
>> Well, everyone everyone gets that it's a joke.
47:21
Um >> Oh, it was meant to be it was meant to be a joke. >> Yeah. Yeah. Yeah. >> Okay. Yeah. Yeah. Yeah.
47:25
Uh this this John guy like totally thinks it's so funny and uh and is just like joking around and and so everyone's just like nightmare fuel like this is crazy and uh John's making the same joke.
47:35
It's super convenient off the open floor plan.
47:37
No kitchen toilet like you [laughter] know people like and then people just joking about all the different stuff. Uh and uh I don't know.
47:46
I mean you know is is is AI going to help with you know architectural design? Of course.
47:51
Uh, is it is Nano Banana going to randomly oneshot like the perfect floor plan? No. Also, no.
47:58
Uh, but you know, of course there's there's stuff that's that's the funniest image.
48:02
[laughter] >> It's so funny. >> So funny.
48:06
Uh, anyway, I was I was actually dying laughing at at this thing. Um >> Um, okay.
48:11
Back back to the code red.
48:13
Vance uh automate compliance and security AI that powers everything from evidence collection to in continuous and continuous monitoring to security reviews and vendor risk.
48:22
Uh yes >> DD dos is adding fuel to the fire.
48:26
>> He says this is why OpenAI is in code red.
48:29
Uh in the two weeks since the Gemini launch chat GBT unique daily active users a 7-day average are down 6%.
48:35
He is sharing to be clear web traffic data.
48:40
Ah, these these traffic sources are so rough.
48:43
I just feel like people use apps like like the web traffic is probably a good proxy.
48:47
It's probably a decent proxy.
48:49
Um, but even then I I just I don't know how how high intent those users are because it's like do you think you're being tracked by similar web that effectively?
49:00
Like I would hope that I don't have that much spyware on my Chrome browser that it knows exactly where I am.
49:07
Maybe it does, but I would think that, you know, OpenAI and Gemini and Google would be like, "Yeah, we're not we're not letting you put a pixel on our site.
49:15
How is >> Do you know do do >> we should we should have someone from Similar Web on the show explain it to us?"
49:23
Like, tell us your sources. Yeah. How do you tell us?
49:25
How do you actually calculate all this stuff?
49:26
Um because I mean, you could just poll people.
49:29
You could just ask a million people, hey, what are you using? Right.
49:33
I don't think that's how this works, but uh >> I wonder I wonder if the any of the the Chrome extensions sell your data.
49:38
I'm sure I'm sure >> a number of them. Uh >> yeah.
49:44
Yeah, I have this uh >> I have this Chrome extension installed right now. TBPN timeline viewer.
49:47
It was vibe coded by someone sitting over there.
49:52
He doesn't even know what programming language it was written in.
49:55
[laughter] >> This is not true.
49:58
>> Did I Did I ever tell this story on the on the show? >> I don't know.
50:01
>> So [laughter] Tyler doesn't want to tell it.
50:02
So Tyler gives us this uh we use a Chrome plugin to to like track the show when we're sharing posts between us.
50:09
And uh this Chrome plugin he like vibe coded it and he sends it over and I unpack it to install it.
50:17
And I'm like why are there like Node modules here?
50:19
Like uh like that's usually for like Node.
50:21
js JavaScript on the back end.
50:22
Uh and he's just like what are you talking [laughter] about?
50:26
And I was like you don't know that you're using Node. js.
50:29
>> It's a good extension sir >> because I think it was just so >> I trust Claude.
50:32
I trust cloud to make the right decision.
50:35
>> I [laughter] don't even specify what programming language it uses, [gasps] which is like pretty sick.
50:38
It's actually extremely bullish for cloud in cloud code. It's really good.
50:42
Uh anyway, the part of the code red of course is uh that uh OpenAI Sora app has fallen out of the top 20 most downloaded apps in the United States on both the App Store and Google Play.
50:54
And so uh things are things are falling.
50:58
I actually opened up Sora today.
51:00
I looked at it and there was some cool stuff happening.
51:04
This is a little bit of a hot take.
51:06
Like it was not it there was still a lot of slop which I would define as like the uh you know it's a POV video of a bus driver with a bunch of cats on the bus and it's like cute and funny or like you know it's a chipmunk water skiing like that type of stuff.
51:22
>> Why you were late for the gym today? >> No.
51:23
Uh but I I was sneaking a peek at Sora while I was driving.
51:28
And if I was like, if I die and crash because I'm looking at sloth, this would be extremely depressing.
51:33
Um, but >> at a stop sign. >> Yeah, I stopped.
51:36
But, uh, the there there was there was one cool uh one which was like more like pixel art actually and it was interesting because you remember the OpenAI Super Bowl ad >> like if you if you prompt Sora to make that type of content, it actually is really cool and you can remix it in a very interesting way.
51:55
And so, uh, like Sam had taken, somebody else had done like a bunch of geometric shapes, uh, pulsing to like electronic music and then Sam was able to take and say make it orchestral music and make them pastel colors and he was able to like remix off of that.
52:13
And that felt like, okay, maybe we're getting into territory.
52:15
Very odd that Sunno and Sora are so close in names.
52:20
I don't know how that happened.
52:21
Maybe they should team up or something.
52:23
wouldn't be the first time OpenAI has uh named something.
52:26
>> Well, who who Oh, yeah. >> similar. Yeah. >> IO. I don't know. Uh but uh I don't know.
52:32
I I I was I was I was seeing like I I don't think it's fully over, you know.
52:37
I think it's like it's in a uh it might be in just a trough of disillusionment. You don't know.
52:41
This could be this could be >> the trough is in and a trough of disillusionment.
52:46
[laughter] >> The trough is in the trough. It's entirely possible.
52:49
But clearly uh the vibes are rough and people are uh taking shots terminally.
52:54
Online engineer says just put the ads in the chat little bro in the chatbot.
52:59
Um because Sam Alman says OpenAI is making a very aggressive infrastructure bet with new partnerships. >> Okay.
53:06
But to be fair, this clip was from uh interview that was like at least a couple months ago. >> Yes. Yes. Yes.
53:12
And also uh you can do both.
53:15
What is interesting is that uh it's it's maybe >> it sounds like they're delaying it sounds like they're delaying ads ads which is which feels odd because uh I personally was I'm I'm maybe the only person that's really excited about ads in chatbt.
53:27
Um I think it's a good thing for the business.
53:30
I think it makes a ton of sense and um and I was excited to see where that rolls out.
53:33
I hope that they don't delay it.
53:34
I think that that that's where they should be running.
53:36
Um, if that if they really are losing ground to to Google and Gemini and like the Gemini app so quickly, I'm shocked because it feels like the Gemini 3 news like the launch went well.
53:47
People were excited about the model.
53:48
The model card looked good.
53:50
The benchmarks look good, but you still have to be pretty tuned in to understand the nuances of the model one way or another.
53:58
Like it's just not like the big model smell and the vibes.
54:02
Like your average AI user doesn't care if the model responds with it's not just this, it's that.
54:12
Like most people clearly that's why it wound up getting RL into the model.
54:17
Most people are like, "Wow, contrastive parallelism. This is epic. I love it. Thank you."
54:23
Like this is really >> contrastive parallelism. >> Yeah.
54:27
Antithetical parallelism.
54:27
like I've never this is like a big big big phrase big word like this is amazing.
54:34
Um, and so I I'm I'm shocked that there would be such a I'm not I'm not shocked by like a vibe shift in on X and in Teapot with regard to how people have been skeptical of the OpenAI financing and so they've been looking for a a a crack to show and Gemini coming out and and leaprogging a little bit even if it's just on some obscure benchmark that the end user might not even care about.
55:01
Um, I was really interested in uh I I understand that like Axe would jump on that narrative.
55:07
Uh, but I'm surprised to see if it's true this idea that like there's actually some sort of consumer shift and I mean it seems like with the red alert comments like maybe maybe it is uh maybe it is.
55:19
Do you think they have to uh explain the funding gap at this point or can we all just agree that maybe maybe everyone got a little too excited? >> Yeah, I don't know. I don't know.
55:32
Um I I feel like everyone's sort of repriced everything already with the Oracle roundtpping and just this idea that uh you know some of the equity investments like they are circular but it's basically just like a discount on their purchases and uh you know these things probably aren't as binding as as we think and so um I feel like the the open AI is going to blow up the economy narrative.
55:58
I feel like that was really oversold and is is much it it should be fading in my opinion, but I don't know.
56:07
Um, uh, Buco Capital Bloke has been, uh, digging into the funding hole.
56:13
>> Apparently, Chad GBT is also down right now. >> I just tested it.
56:16
It's not down for me, but the chat says it's down.
56:20
>> Well, X is saying it's down. >> Oh, really? Oh, wow.
56:22
Uh, time to Baja blast those servers back online, brother.
56:28
[laughter] It's time to rock.
56:30
Uh, we need We need a pump-up speech that doesn't include any negative phrases that can be taken out of context.
56:37
We need we need to be Baja blasting. We have to Baja blast.
56:40
We have to baja blast our way to the top of the app store.
56:45
Sora team, I need you to baja blast.
56:49
>> It's time to >> It's time to baja blast to the top of the app store.
56:52
You have to Baja blast at the top of the app store.
56:55
You have to and uh we're gonna have to Baja blast some uh some funding into this company because apparently there's a $270 billion funding hole here.
57:05
Uh this is from a a podcast between uh Ron John who writes at read margins and and Alex Caneritz at big technology.
57:15
They did a podcast together and uh here's the quote from Buco Capital Bloke says squaring the total.
57:21
It leaves OpenAI in a 270 billion 27 billion funding hole. The math doesn't work.
57:28
Maybe OpenAI should release to the world.
57:30
Here's how the math can work because I haven't seen anyone state how this can actually work.
57:35
And so even if you get there, OpenAI does fall $27 billion short of the money it needs to continue funding its commitments. Right?
57:40
So it has in 203 in 2030 Open AAI free cash flow will be about 287 billion. That's like insane.
57:50
Uh that's [laughter] if if this is I this feels like silly to me because uh if you if you're in a situation where you have 287 billion of free cash flow like you can't raise more debt on that.
58:03
Like I I I feel like math tends to work out when you go from a nonprofit to a $300 billion cash flow a year in 10 years.
58:13
Like it just everything just forms in front of you.
58:18
Like the you like yes, you are building the bridge as you're driving, but like that tends to happen when you're on that much of a tear.
58:23
The bigger question is like can they actually free cash with their $287 billion in 2030?
58:26
Um, >> so Amazon's uh free cash flow uh for 2024 was 38 billion and let's see what Google's >> Yeah, this is like a >> 72.
58:42
So saying that that uh they're going to do three times more >> than uh Google and Amazon.
58:52
>> So this is the HSBC report is modeling 386 billion in annual enterprise AI revenue by 2030. enterprise AI revenue. Huh.
59:01
That's these are just huge numbers.
59:04
It's it's almost not worth analyzing.
59:04
Um I I I still think the biggest the the biggest thing is just uh understanding how how significant how tied up are these uh are these contracts.
59:15
Well, um let me tell you about fall the generative media platform for developers.
59:21
[applause] Develop and fine-tune models with serverless GPUs and on demand clusters.
59:25
So, um, what else is going on?
59:29
We should read through, uh, Ben Thompson's latest piece because he's provided a lot more context on Google, Nvidia, and OpenAI with a post called Google, Nvidia, and OpenAI.
59:42
And we we we thank Ben Thompson for uh, always having an even keel.
59:44
Uh, highly recommend subscribing to strategy.
59:47
Uh, it's a fantastic publication if you're not subscribed already.
59:51
Um, and uh, and he's a former guest of the show.
59:53
Uh so let's read through this his latest uh Monday uh piece.
59:58
He says, "A common explanation as to why Star Wars was such a hit and continues to resonate nearly half a century on from its release with everyone except Jordy Hayes who hasn't seen it because he hasn't seen any movies."
1:00:12
>> I've seen Star Wars, John.
1:00:13
>> How many Star Warses have you seen?
1:00:15
>> I have to have seen all of them except some of the more been some recent >> All of them except some of them. Well, no.
1:00:21
The the the more the more recent ones are the like hasn't there been like a new Star Wars in the last >> How many How many Star Wars are there?
1:00:27
>> There's There was is there six >> six? There's six Star Warses.
1:00:30
That's how many movies they've made?
1:00:33
>> Six like real Star Wars. >> There's six. There's six.
1:00:38
>> I'm gonna I'm gonna Hasn't there been like six?
1:00:41
>> Everyone calls it the Septilogy.
1:00:41
Yeah, there's there's six.
1:00:43
[laughter] >> Wait, aren't there six? >> There's nine. There's three trilogies.
1:00:50
There's the There's the original trilogy, the prequel trilogy, and then the sequel trilogy.
1:00:53
And then there's also two spin-offs. >> Okay.
1:00:56
So, I didn't watch I didn't watch any of like the the the like new ones.
1:01:00
>> You But you watch the prequel. >> Rise of Skywalker.
1:01:03
>> Talking about George Lucas directed. >> Yeah. >> Okay. Okay.
1:01:05
So, so, so he's a Lucas head. >> I'm a Yeah, exactly. >> Okay.
1:01:09
So, you've seen A New Hope, you've seen Strikes Back, >> real Star Wars, >> you've seen Return of the Jedi, and then you've seen Phantom Menace and uh Revenge of the Sith and Return of the something.
1:01:19
I can't actually I actually don't know that much about Star Wars, but anyway, you should know enough to follow along with this analogy from Ben Thompson.
1:01:27
He says, "You have Luke bored on Tatooine, called to adventure by a mysterious message born by R2-D2 that he initially refuses refusing refusal of the call."
1:01:38
This is the classic uh this is the classic um hero's journey.
1:01:43
Uh so he refuses the call.
1:01:46
A mentor in Obi-Wan Kenobi leads him to the threshold of leading Tatooine and faces tests while finding new enemies and allies. He enters the cave.
1:01:55
The Death Star escapes after the ordeal of Obi-Wan's death.
1:01:57
Spoiler alert, Ben, what are you doing, brother?
1:02:00
What if somebody hasn't seen it and they don't know that Obi-Wan dies? It's crazy.
1:02:06
>> Oh, >> and carries the battle station plans to the rebels while preparing for the road back to the Death Star.
1:02:11
He trusts the Force in his final test and uh and returns transformed.
1:02:18
And when you zoom out to the original trilogy, it's simply an expanded version of this of the story.
1:02:23
Uh this time, however, the ordeal uh is an entire second movie, The Empire Strikes Back.
1:02:28
The heroes of the AI story over the last three years have been two companies, OpenAI and Nvidia.
1:02:32
The first startup is called uh the first is a startup called with the release of Chachib to be the next great consumer tech company.
1:02:42
The other was best known as a gaming chip company characterized by boom and bust cycles driven by their visionary and endlessly optimistic founder transformed into the most essential infrastructure provider for the AI revolution over the last few weeks.
1:02:56
However, both have entered the cave. They're in the cave.
1:02:59
This is the cave of disillusionment [laughter] and are facing their greatest ordeal.
1:03:05
The Google empire is very much striking back.
1:03:09
And I believe uh didn't Anen uh over at A16Z uh coin that the the Empire Strikes Back >> formerly A16Z went independent. >> Oh, he's independent. Yeah. >> Oh, I had no idea.
1:03:22
>> So, is this what um Sam meant when he tweeted the picture of the Death Star?
1:03:26
[laughter] >> I I feel like we never really figured out what he meant by that.
1:03:29
>> That was before GBT, I think. >> Yes.
1:03:32
>> I think he wanted the 2025 vague post of the year. >> No award.
1:03:36
It's so vague that even after the release, he's still in No.
1:03:37
I I I I I [laughter] I think this is it. I think this is it.
1:03:41
It's it's Google's the empire and and he's launching the thing that >> it still didn't make sense at the time because um like open was like clearly in the lead of the models >> like 2. 5 I think I think 2.
1:03:51
5 was the best job >> cash flow you know who has more soldiers who has more researchers who has more TPUs right like like they you know it'd be fair to characterize um it'd be fair to characterize Google as the empire the whole time >> I guess the founding of Open eye then.
1:04:12
Uh, Google was definitely the Death Star, right? >> Interesting.
1:04:15
>> That's isn't that the kind of >> origin story of OpenAI? >> Yeah. Yeah. Yeah, it was.
1:04:19
They were they were worried about that.
1:04:21
Um, anyway, I enjoy the uh Star Wars based analogies almost as much as I enjoy numeral. com compliance handled.
1:04:29
Numero worries about sales tax and VIT compliance so you can focus on growth. So, Google strikes back.
1:04:36
The first Google blow was Gemini 3, which scored better than OpenAI's state-of-the-art model on a host of benchmarks, even if actual realworld usage was a bit more.
1:04:44
And even Gemini 3's biggest advantage is its sheer size and the vast amount of compute that went into creating it.
1:04:51
This is notable because OpenAI has had difficulty creating the next generation of models beyond the GPT4 level of size and complexity.
1:05:01
What has carried the company is a genuine breakthrough in reasoning that produces better results in many cases, but at the cost of time and money. Time and money. Throw in a ramp.
1:05:12
com ad right in the middle of this jackery article. I love it.
1:05:15
Gemini 3's success seemed like good news for Nvidia, who I listed, Ben listed as a winner from the release.
1:05:23
Quote, "This is maybe the most interesting one.
1:05:26
Nvidia, who reports earnings later today, is one is on one hand a loser because the best model in the world was not trained on their chips, pro proving once and for all that it is possible to be competitive without paying Nvidia's premiums.
1:05:39
On the other hand, there are two reasons for Nvidia's optimism.
1:05:40
The first is that everyone needs to respond to Gemini and they need to respond now, not at some future date when their chips are good enough.
1:05:48
Did you know that Sundar was uh appar like people were claiming that he had used the phrase code red and he back in 2022 at the chat GBT launch when hard was >> and so >> yes I I I I I'm remembering that now I missed it.
1:06:06
Sundar came out and said he didn't use that exact term, >> but uh there there was reporting that he did.
1:06:14
And I heard a >> a rumor that he also said that he wanted to baja blast [laughter] Sam Alman >> out of the atmosphere >> out of San Francisco, out of the atmosphere [gasps] with a Death Star laser. [laughter] >> Yeah.
1:06:29
No, >> I want to I want to try to find more historical examples >> of Baja blasting, folks. We got a Baja blast.
1:06:36
You got a Baja blast sometimes.
1:06:39
Um, so Google started its work on TPUs a decade ago.
1:06:42
Everyone else is better off sticking with Nvidia, at least if they want to catch up.
1:06:45
Secondly, and repeat relatedly, Gemini reaffirms that the most important factor in catching up or moving ahead is more compute.
1:06:52
This analysis, however, missed one important point.
1:06:56
What if Google sold its TPUs as an as an alternative to Nvidia?
1:06:59
We're going to talk to Tay Kim, author of The Nvidia Way about that.
1:07:06
>> He's going to tell all.
1:07:07
>> He's going to tell all. He's breaking a silence.
1:07:08
Um, so, uh, that's exactly what the search giant is doing.
1:07:11
First with a deal with Anthropic, then a rumored deal with Meta, and third with a second wave of Neoclouds, many of which started as crypto miners and are leveraging their access to power to move into AI.
1:07:21
So, a lot of those NeoClouds, they they found a bunch of power and they don't really have the right chips yet or maybe they're upgrading their chips.
1:07:29
They might be in a new cycle and so TPU could be at the top of the menu for them.
1:07:34
>> Uh, s suddenly it is Nvidia that is in the crosshairs with fresh questions about their long-term growth, particularly at their sky-high margins.
1:07:42
If there were in fact a legitimate competitor to their chips, this does, needless to say, raise the pressure on OpenAI's next pre-training run on Nvidia's Blackwell chips.
1:07:50
Uh the base model still matters and OpenAI needs a better one.
1:07:54
Um and [snorts] Nvidia needs evidence that it can be created on their chips.
1:07:58
Um what is interesting to consider is which company is more at risk from Google and why.
1:08:02
On one hand, Nvidia is making tons of money and if Blackwell is good, Vera Rubin promises to be even better.
1:08:09
Moreover, while Meta might be a natural Google partner, the other hyperscalers are not.
1:08:17
>> [clears throat] >> They're not going to be selling you, you know, is we're going to have the CEO of Amazon Web Services, Matt Garmin, on the show in just >> AWS announced a new chip, >> and I don't think AWS is going to be buying TPU anytime soon, but we will be asking him that question. Exactly.
1:08:29
Uh, and I want to get to the bottom of it.
1:08:33
So, OpenAI, meanwhile, is losing more money than ever and is spread thinner than ever, even as the startup agrees to buy ever more compute with revenue that doesn't exist yet. >> That's the worst. words.
1:08:47
>> And yet, despite all that, and while still being quite bullish on Nvidia, I still like OpenAI chances more. >> Whoa.
1:08:55
>> Oh, Ben Thompson likes likes OpenAI's chances. Indeed.
1:08:58
If anything, my biggest concern is that I seem to like OpenAI's chances better than OpenAI itself. Whoa. Nvidia's Moes.
1:09:08
Wait, he wrote this before he wrote this before the Red Alert memo. Interesting.
1:09:11
He's really got a crystal ball over there.
1:09:14
Um, so Nvidia's modes, if you go back a year or two, you might make the case that Nvidia had three modes relative to TPUs.
1:09:21
Senior per superior performance, significant more flexibility due to GPUs being more general purpose than TPUs and CUDA and the associated developer ecosystem surrounding it.
1:09:31
OpenAI meanwhile had the best model, extensive usage of their API and the massive number of consumers using chatbt.
1:09:37
The questions then is what happens if the first differentiator for each company goes away?
1:09:42
That in a nutshell is the question that's been raised over the last 2 weeks.
1:09:45
Does NVIDIA preserve its advantages if TPUs are as good as GPUs?
1:09:49
And is OpenAI viable in the long run if they don't have the unquestioned best model.
1:09:54
So Nvidia's flexibility advantage is a real thing.
1:10:00
It's not an accident that the fungeability of GPUs across workloads was focused on as a justification for increased capital expenditures by both Microsoft and Meta.
1:10:09
TPUs are more specialized at the hardware level and more difficult to program for at the software level.
1:10:14
Uh to that end, to the extent that customers care about flexibility, then Nvidia remains an obvious choice.
1:10:20
The interesting thing about the flexibility is that isn't SSI a big TPU buyer?
1:10:25
I I feel like they I feel like SSI was maybe going big on TPU and I think of SSI is very much like we're going to experiment.
1:10:36
We need maximum flexibility by default.
1:10:37
I would assume that they're a heavy consumer of of GPU because they want as much flexibility as possible.
1:10:44
But maybe the nature of Ilia's research is flexibility within that that is still afforded within the TPU ecosystem.
1:10:52
>> They're using TPUs through Google Cloud. >> Yeah.
1:10:55
But there's been no >> Oh, yeah.
1:10:57
They're not buying them, but still that implies more flexibility because you can just turn it on or off. >> Yeah.
1:11:02
I'm talking about the actual like like the TPU is a is a is an ASIC.
1:11:04
It has it has like literally like less features than the GPU.
1:11:08
Like a gaming GPU and a like the TPU doesn't I think it doesn't support like FP4, right, or something like that.
1:11:16
There's some there's some type of math that is harder to do on a TPU because it's making tradeoffs, right?
1:11:22
And uh and so even though I don't understand it I fully I understand that uh that you know TPUs are more specialized at the hardware level.
1:11:31
And so if you were to be in like the era the era of research maybe you would want something that's less specialized because you'd be like I'm going back to exploring all sorts of different types of math that aren't necessarily >> again if you're if you're buying TPUs through the cloud you're effectively just buying cloud services. >> Yeah.
1:11:48
you do have more flexibility because you can say, "Hey, we're we want to use more of this or we want to use less."
1:11:53
You're not like buying a bunch of servers and and chips that >> on the scale on the scale thing that makes perfect sense. >> Yeah.
1:12:00
I think also like historically TPUs have definitely been more restrictive just because of like the software was just not as good um or it was closed source or whatever and then you you know yesterday Don Patel was talking about how Google is slowly trying to open source more and more stuff for the TPU. Yeah.
1:12:14
>> So you would imagine that in the future um it should be you know much easier to use TPUs.
1:12:18
Generally >> Sean in the chat says flexibility in terms of hey I have this new architecture.
1:12:22
Do I need to write kernel code from scratch or is there a nice CUDA CUDA module I can use just time right flexibility is engineering work from Nvidia. >> Yeah. Yeah.
1:12:30
No that's a really good point.
1:12:32
Uh so CUDA meanwhile has been a critical source of NVIDIA lockin both because of the low-level access it gives developers but also because there is a developer network effect.
1:12:40
Dylan Patel was talking about this.
1:12:41
you're just more likely to be able to hire low-level engineers if your stack is on Nvidia.
1:12:47
The challenge for Nvidia, however, is the big is that the big company effect could play out with CUDA in the opposite way to the flexibility argument.
1:12:53
While big companies like the hyperscalers have the diversity of workloads to benefit from the flexibility of GPUs, they also have the wherewithal to build alter an alternative software stack.
1:13:03
alternative software stack. that they did that they did not do so for so for such a long time is a function of it simply not being worth the time and trouble when capital expenditure plans reach the hundreds of billions of dollars however what is worth the time and trouble changes a useful analogy
1:13:22
here is the rise of AMD in the data center that rise has not occurred in on premises installations or the government which is still dominated by Intel rather large hyperscalers found it worth their time and effort to rewrite extremely low-level software to be truly agnostic between AMD and Intel, allowing the formers lead in performance to win the battle. Uh, and so AMD better
1:13:43
Uh, and so AMD better performance, better efficiency per dollar, but didn't have the best software.
1:13:52
And now but now because there's so much uh on the line so many so the spending amount is so high companies will go and work around all the bugs develop new software that allows them to take advantage of AMD's better performance.
1:14:05
Uh in this case the challenge Nvidia faces is that its market is a relatively small number of highly concentrated customers with the resources mostly as yet unutilized to break down the CUDA wall as they already did in terms of Intel's differentiation.
1:14:21
It's clear that Nvidia has been concerned about this for a long time.
1:14:24
This is from Nvidia waves and moes which he written at the absolute top of the Nvidia hype cycle after the 2024 introduction of Blackwell.
1:14:32
This article takes full circ says in the before times i. e.
1:14:42
before the release of chat gbt Nvidia was building quite the free software moat around its GPUs.
1:14:49
The challenge is that it wasn't entirely clear who was going to use all of that software.
1:14:54
Today, meanwhile, the use cases for those GPUs is very clear and those use cases are happening at much higher level at a much higher level than CUDA frameworks i. e. on top of models.
1:15:05
That combined with the massive incentives towards finding cheaper alternatives to Nvidia means both the pressure uh to and the possibility of escaping CUDA is higher than it ever has been.
1:15:17
Even if it is still distant for low-level work, particularly when it comes to training, Nvidia has already started responding.
1:15:23
I think that one way to understand DGX cloud is that is Nvidia's temp attempt to capture the same market that is still buying Intel server chips in a world where AMD chips are better because they have already standardized on them.
1:15:36
N IM's are another attempt to build lockin.
1:15:39
In the meantime though, it remains noteworthy that Nvidia appears not to be taking as much margin with Blackwell as many have expected.
1:15:46
The question as to whether they will have to give back more in future generations will depend on not just their chips performance but also on redigging a software mode increasingly threatened by the very wave that made GTC such a spectacle.
1:16:00
So Blackwell margins are doing just fine.
1:16:03
I should note is he's he's back to the original article, the modern article.
1:16:06
Uh as they should in a world where everyone is starved for compute.
1:16:11
Indeed, that may make this entire debate somewhat pointless.
1:16:15
Implicit in the assumption that GPUs might take share from G from TPUs might take share from GPUs.
1:16:19
Is that for one to win the other must lose the real decision maker may be TSMC which makes both chips and is positioned to be the real break on the AI bubble. Interesting.
1:16:32
So, uh Chat GPT and Moes resiliency.
1:16:36
That's >> I can read through this one.
1:16:37
uh chatbt in contrast to Nvidia sells into two much larger markets.
1:16:41
The first is developers using their API and according to open AAI anyways this market is much stickier and reticent to change which makes sense.
1:16:50
developers using a particular model's API are seeking to make a good product.
1:16:53
And while everyone talks about the importance of avoiding lock in, most companies are going to see more gains from building onoid lock, expanding from what they already [laughter] >> always Baja blast.
1:17:02
Uh and for a lot of companies that is opening eye one, I I would caveat here.
1:17:07
We were we were talking to a founder yesterday uh who um off the show who was saying he um immediately uh uh as soon as Gemini 3 launched spent like 12 hours uh 12 hours like uh just moving moving over uh to to Gemini from from OpenAI.
1:17:25
So depending on the product I I don't know that >> uh API is always going to be super sticky.
1:17:31
Um I say winning business one app uh by one app by one will be a lot harder for Google than simply making a spreadsheet presentation to the top of a company about upfront costs and total cost of ownership.
1:17:44
Still API costs will matter and here Google almost certainly has a structural advantage.
1:17:47
The biggest market of all however is consumer Google's bread and butter.
1:17:50
What makes Google so dominant in search impervious to both competition and regulation is that billions of consumers choose to use Google every day multiple times a day in fact.
1:17:59
Yes, Google helps them helps this process along with its payments to its friends, but that's downstream from its control of demand, not the driver.
1:18:05
What is paradoxical to many about this reality is that the seeming fragility of Google's position competition really is a click away.
1:18:13
It is in fact its source of strength.
1:18:15
Um, and then there's a excerpt from >> We can skip this one and continue at the bottom.
1:18:21
The CEO of a hyperscaler can issue a decree to work around CUDA.
1:18:22
uh an app developer can decide that Google's cost structure is worth the pain of changing the model undergirling their app.
1:18:31
Uh changing the habits of 800 million people who use chatbt every week however is a battle that can only be fought by individ individual by individual.
1:18:40
This is chatbt's true difference from Nvidia in their fight against Google. >> Yeah.
1:18:47
>> And so this I think is the most important takeaway is just Ben Thompson creates aggregation theory.
1:18:52
this idea of like it's so important to aggregate demand in the modern internet world.
1:19:01
It's potentially the only thing you can do.
1:19:03
You can't really monopolize supply.
1:19:05
It's very hard to monopolize supply, but monopolizing demand is something that happens.
1:19:09
>> Um and and and the the strength of habits is significant.
1:19:13
Like we're watching this stuff every single day.
1:19:16
So we can t take the time to okay, yeah, we should test out this other, you know, model.
1:19:23
we should daily drive this app.
1:19:26
But for a lot of people, if they've if they have a map that's installed and they've been using it for a year, they're never changing.
1:19:31
>> Uh even if the model is slightly better over there, they're just not even going to hear about it because they're just like, "This is the thing that I use to plan my vacations or this."
1:19:38
>> And the thing the thing that I've heard come up multiple times is people that when Gemini 3 launched, they switched to Gemini 3 on desktop, but they stayed using Chat GBT on mobile. >> Yep.
1:19:47
And I mean to to be completely transparent like like the uh the the Gemini mobile app has is really really struggling to stay connected.
1:19:55
There's something in the when you fire off a prompt, it doesn't like save it locally and then cache that and then send it off, inference it and then come back it like unless you keep the app open like it will uh it will just give you like a server disconnected error.
1:20:11
Like I've gotten like dozens of these um and and that's going to be a real Like I think it should be something they should be they should be able to fix in like a weekend, but you know hopefully it's soon.
1:20:22
Um but for a lot of people >> Logan. >> Yeah.
1:20:25
Well, they they'll get it. Um >> they'll get it.
1:20:29
>> But uh uh back to the moat and the map ad uh the moat map and advertising.
1:20:32
I this is I think a broader point.
1:20:35
The na the naive approach to moes focuses on the cost of switching.
1:20:39
In fact, however, the more important correlation to the strength of a moat is the number of unique purchasers to users.
1:20:46
The strength of the moat is increased by the number of buyers.
1:20:55
Okay, so that you can see where this is going with like Nvidia has five buyers and chat GBT has a billion buyers essentially >> and once it has 20 million, how many Yeah, advertisers might be even more, right?
1:21:07
So this is certainly one of the simpler charts I've ever made because it's literally just one line.
1:21:10
Um but it's not the first in the moat genre.
1:21:13
So he talks about the moat map.
1:21:16
I argued that you could map large tech companies across two spectrums.
1:21:20
The degree of supplier differentiation from Facebook where the supplier is completely commoditized just your friend on Facebook to uh to Microsoft and Apple where the suppliers are somewhat more controlled.
1:21:32
Um yeah there's the >> what a chart.
1:21:34
the the more the more unique buyers of your product you have, the the the stronger your moat because it's hard to con because you have to convince each one of them.
1:21:42
Um and then second, the extent to which a company's network effects were externalized.
1:21:50
Internalized network effects are Facebook again and then externalized is Microsoft.
1:21:54
And so putting this together gave the moat map.
1:21:56
So who has the uh the network effect versus the suppliers map.
1:22:01
If we scroll down there, you can see them.
1:22:03
Uh what you see in the upper right are platforms.
1:22:05
The lower left are aggregators.
1:22:07
Platforms like the app store enable differentiated suppliers which allows them to profitably take a cut of purchase purchases driven by those differentiated suppliers.
1:22:15
Aggregators meanwhile have totally commoditized their suppliers but have done so in the service of maximizing attention which they can monetize through advertising.
1:22:22
It's the bottom left that I'm describing with the simplistic graph above.
1:22:26
The way to commoditize suppliers and internalize network effects is by having a huge number of unique users.
1:22:30
And by extension, the best way to monetize that user base and to achieve a massive user base in the first place is through advertising.
1:22:38
It's so obvious the bottom left is where ChachiBT sits.
1:22:40
[laughter] >> I wonder I wonder what he thinks then about uh about them potentially kind of >> delaying ads >> way and delaying ads.
1:22:49
>> Probably punch in the air.
1:22:49
Him and Eric Seford are probably no.
1:22:53
And I'm I'm right there with them. I completely agree. Boo. Launch the ads product. Launch the ads product. Get it out. Come on. Don't delay that.
1:23:03
That's the most important thing.
1:23:05
Um, [laughter] so, uh, at one point it didn't seem possible to commoditize content more than Google or Facebook did.
1:23:13
But that's exactly what LLMs do.
1:23:15
The answers are a statistical synthesis of all the knowledge the model makers can get their hands on and are completely unique to every individual.
1:23:24
At the same time, every individual's user usage should at least in theory make the model better over time.
1:23:32
It follows then that Chacht should obviously have an advertising model.
1:23:35
This isn't just a function of needing to make money.
1:23:37
Advertising would make CHP a better product.
1:23:39
It would have more users using it more, providing more feedback, capturing purchase signals not from affiliate links, but from personalized ads would create a uh would create a richer understanding of individual users, enabling better responses.
1:23:52
And as an added bonus and one that is very pertinent to this article, it would dramatically deepen OpenAI's moat.
1:23:58
Yeah, I I keep going back to >> uh this idea that OpenAI needs personalized ads like like Instagram.
1:24:07
Like that's what Sam said when he was interviewed multiple times on his ad strategy.
1:24:10
He was like, you know, like ads can be bad, but this these Instagram ads are pretty good.
1:24:14
You know, he's he's and people are, oh, he's like read he's backtracking.
1:24:18
It's like, no, that's fine.
1:24:19
uh get like like do the proper business model like please implement the correct business model. Uh I'm happy about that.
1:24:26
Um, but the interesting thing is that Instagram does not surface ads when you necessarily when you search for something like if you're if you're on a video for like a Ferrari like you don't just immediately get an ad as your next thing for like Ferrari of Hollywood like or Ferrari of Beverly Hills like no you get an ad for the toaster that you were about to check out on.
1:24:50
I actually do like half half half the ads I get on Meta are local dealerships in LA. >> Yes.
1:24:59
But but importantly not when you're like searching. It's not tied to search.
1:25:04
And so uh Chatbt can do the same thing where they can clearly show you um >> uh where they can clearly show you something that you are about to check out.
1:25:14
You're shopping for Christmas for this thing.
1:25:16
You're searching for the Roman Empire.
1:25:18
let's show you the ad for the thing that you're shopping that you're shopping for like right next to it. It's fine.
1:25:23
Uh and so I I think that can work very well.
1:25:25
Anyway, uh let's go to Google's advantages.
1:25:27
It's not out of the question that Google can win the fight for consumer attention.
1:25:29
The company has a clear lead in image and video generation, which is one of the uh one of the reasons why I wrote about the YouTube tip of the Google spear.
1:25:36
I mean, Google's advantage in data is insane.
1:25:40
Like YouTube, so massive.
1:25:40
And that's got to be just a compound and comp.
1:25:43
I mean, we're uploading another three hours of video to YouTube today. You're welcome.
1:25:49
[laughter] >> This one's for you, son.
1:25:51
>> This one's for you, Dennis.
1:25:51
Um, but the the the flip side is like they also see the entire internet because the way the Google bot scrapes like the Google searching in Gemini is such a killer feature.
1:26:02
Like it's such a it's such a killer feature if they can keep that on and they can actually surface that like and and the AI search results are obviously going to get good.
1:26:10
They're going to figure out how to surface it.
1:26:11
I I think I'm still pretty optimistic.
1:26:12
But let's see what Ben Thompson has to say.
1:26:16
And let's also tell you about Adio.
1:26:16
The AI native CRM Adio builds scales and grows your company to the next level.
1:26:20
So Google is obviously capable of monetizing users even if they hadn't turned on ads in Gemini yet.
1:26:26
It's also pointing out as Eric Sufer did in a recent strate collabing.
1:26:33
We'd love to see it that Google started monetizing search less than two years after its public launch.
1:26:39
It is search revenue far more than venture capital money that has undergurtded all of Google's innovation over the years.
1:26:44
And it is what makes them such a behemoth today.
1:26:46
In in that light, OpenAI's refusal to launch and iterate on an ads product for ChatBT, now three years old, is a dereliction of business duty. He's calling him out. >> Whoa.
1:26:57
>> Particularly as the company signs deals for over a trillion dollars of compute. What are you doing, Sam? Get the ads out.
1:27:02
Put the ads in in the chatbot. We love it.
1:27:05
>> Just put the ads in the chatbot >> and go, you got to Baja blast some ads into that app. You got to.
1:27:09
[laughter] I want ads in Chad GBT. Please.
1:27:12
On on the flip side, it means Google has the resources to take on Chachi GPT's consumer lead with a World War I style war of attrition. Ryan Mal call back.
1:27:26
Uh, OpenAI's lead should be unassailable, but the company's insistence on monetizing solely via subscriptions with a downgraded degraded user experience for most users and price elasticity challenges in terms of revenue maximization is very much opening the door to a company that actually cares about making money.
1:27:41
To put it another way, the long-term threat to Nvidia from TPUs margin uh from TPUs is margin dilution.
1:27:48
The challenge of physical products is that you do actually have to charge people who buy them which invites potentially unfavorable.
1:27:55
>> I always so yeah who who both Gemini and and Chatbt will have ads eventually. Yes. Right. >> You can bet on that.
1:28:04
>> Uh who will ramp ad revenue faster?
1:28:08
It's hard not to bet on Gemini, even with a smaller user base, because they have the ad network.
1:28:13
They have all the they have all the customer relationships already.
1:28:18
You can they can just say like, "Hey, here's a popup.
1:28:19
Do you want here's here's $10,000 of free ad credits. Try it out." Right? It's like native.
1:28:26
>> It will already be in AdWords.
1:28:26
you the the example that I use is like how you know Zuck was able to take Instagram which had a lot of users, plug it into the meta ads platform and just scale revenue like crazy and then do it again with with reals.
1:28:39
>> And so uh yeah, very very clear that they both will have ads.
1:28:41
Um and uh and again if if Gemini can really ramp that quickly, they could again like I I do feel like we're moving towards a world where you like every consumer will be able to get the best LLM for free, >> right?
1:29:00
I don't I don't I don't necessarily believe that uh every American will be paying for an LLM in 5 years.
1:29:06
And so, uh, if Google can get there first and then keep Gemini on the frontier and deliver the best free, fastest text and >> image model, that's going to be very, very difficult to compete with.
1:29:24
And so, again, like getting getting to ads faster, uh, it feels like it makes more sense. >> I like that point. I have a rebuttal.
1:29:31
First, I'm going to tell you about Figma.
1:29:32
Think bigger, build faster.
1:29:32
Figma helps design and development teams build great products together.
1:29:36
So, uh, getting to ads first is an advantage. That's your take.
1:29:41
Uh, I like it, but there is a little bit of a risk with launching ads first because you could forever be branded with you're the ads one.
1:29:51
And we saw this when uh Arvin from Perplexity came on the show and he mentioned this idea of ads in LLM queries and we all agreed on the on the discussion.
1:30:02
We all agreed that ads were going to come to AI tools because that is the way to get the most people using them and make intelligence free.
1:30:12
and you got in a, you know, a debate with Mark Cuban over this, for example, um, and, uh, >> and it seemed very logical, but there is the there is the fact that the first major chat app to put ads in their app is going to be a massive news cycle.
1:30:32
It's just going to be like national news. >> Sam adman. >> Exactly.
1:30:36
It'll be OpenAI has ads now or it'll be Gemini has ads now.
1:30:39
And so you don't necess if you like it's much easier to be the second mover there because it's going to be less of a news cycle.
1:30:47
And so you kind of do want to there is there is a little bit of advantage to being the second mover there, right?
1:30:53
Because you're you're you're going to get sort of branded as like oh that's the ad supported one and the other one can add ads and people be like oh yeah like I guess but that's like standard.
1:31:02
You know there's going to be like a backlash and people will be like oh no I don't like this company blah blah blah blah blah.
1:31:06
Like >> yeah, I mean the harder the harder thing is just how do you do it, right?
1:31:09
S like I do feel like it's different. You're going to an LLM.
1:31:13
People are get going to an LLM for advice and recommendations.
1:31:18
That's different than going to Google and searching and seeing ads at the top and there's plenty of surface area.
1:31:23
I >> No, no, I'm not I'm not saying there's no surface area, but I'm just saying like the right way to do ads in LLM is not clear yet. >> Um yeah. Yeah.
1:31:32
I mean they will need to do some experimentation but I mean just starting with um you know like Google already has retargeting information.
1:31:39
I I actually went to uh Gemini with a question and it clearly knew everything about me and it said you're the host of TVPN and you've start you founded these companies and it had it it already knew that probably just because I authenticated with something else.
1:31:52
I don't know but uh it it should know okay we could retarget you with this. Let's put this in.
1:31:59
Um, and and the same thing with uh with OpenAI, like with chatb, there's plenty of spaces where it's like you're waiting for it to uh to give you the answer.
1:32:09
Okay, hey, we're generating you the image.
1:32:10
Why don't you show me other images of ads right there?
1:32:12
Uh there's tons of surface area.
1:32:14
I I agree that there will be a whole bunch of iterations on like what the ideal ad looks like.
1:32:20
Um but yeah, you could clearly get out.
1:32:23
So, let's go back to Ben Thompson. Close this out.
1:32:25
says, "The reason to be more optimistic about OpenAI is that an advertising model flips this on its head.
1:32:30
Because users don't pay, there is no ceiling on how much you can make from them, which by extension means that the bigger you get, the better your margins have the potential to be, and thus the total size of your investments.
1:32:42
Again, however, the problem is that the advertising model doesn't exist yet."
1:32:45
So, he started this article recounting the hero's journey in part to make the easy leap to The Empire Strikes Back.
1:32:53
However, there is a personal angle as well.
1:32:54
The hero of this site has been aggregation theory and the belief that controlling demand trumps everything else.
1:33:02
There's there Google was my ultimate protagonist.
1:33:04
Moreover, I do believe in the innovation and velocity that comes from a founder-ledd company like Nvidia.
1:33:11
And I do still worry about Google's bureaucracy and disruption potential making the company less nimble and aggressive than open AI.
1:33:18
More than anything though, I believe in the market power and defensibility of 800 million users, which is why I think Chachbt still has a meaningful moat.
1:33:25
At the same time, I understand why the market is freaking out about Google.
1:33:31
Their structural advantage their their structural advantages in everything from monetization to data to infrastructure to R&D is so substantial that you understand why OpenAI's founding was motivated by the fear of Google winning AI.
1:33:45
It's very easy to imagine an outcome where Google's inputs simply matter more than anything else.
1:33:50
Which is to say, one of my most important theories is being put to the ultimate test.
1:33:55
Which perhaps is why I'm so frustrated at OpenAI's avoidance of advertising.
1:33:59
Google is now my antagonist.
1:34:01
Google has already done this once.
1:34:04
Search was the ultimate example of a company winning an open market with nothing more than a better product.
1:34:09
Aggregators win new markets by being better.
1:34:11
The open question now is whether one that has already reached scale can be dethroned by the overwhelming application of resources, especially when its inherent advantages are diminished by refusing to adopt an aggregator's optimal business model.
1:34:25
I'm nervous and excited to see how far aggregation theory really goes. Fascinating. >> It's his baby.
1:34:32
>> Yeah, it's uh it is I I agree it is the correct it is the correct framing.
1:34:36
Um, it'll just be very interesting to see.
1:34:41
Uh, I I I really wonder um who's gonna who's going to take the leap first?
1:34:44
Who is going to uh who's going to jump and uh and put ads in in in the in the app first.
1:34:51
It feels like Google should do it.
1:34:54
It feels like Google will will be able to do it. >> Yeah.
1:34:56
Nobody's going to be like, "What?
1:34:59
>> Google is putting ads in a product?" >> Yeah.
1:35:02
>> It won't be that surprising.
1:35:04
>> Um >> so they should probably move faster.
1:35:06
>> We have some uh breaking news.
1:35:07
>> What's the breaking news?
1:35:07
Jason Frerieded is joining the show at 2 p. m. Surprise guest.
1:35:10
He's launching Fizzy today.
1:35:13
>> Canban as it should be, not as it has been.
1:35:17
>> Uh I will wait uh we'll wait to talk about this till he joins in an hour and 20 minutes.
1:35:23
So little surprise g uh guest appearance from alleged >> calling out his competitors directly.
1:35:27
I love when founders do that.
1:35:29
[laughter] >> Founder mode. >> Founder. >> Founder mode. Yeah. Very.
1:35:34
>> Uh, should we talk about John Gandrea leaving the company? >> He's out. I quit. >> I quit.
1:35:39
[laughter] >> You like that one?
1:35:41
I think that's probably been used before.
1:35:43
Sub headline, but >> Amar sub.
1:35:48
>> So, of course, Mark German has the scoop, I believe. Um, >> the Germinator.
1:35:53
>> Germinator's at it again.
1:35:53
He says, "Apple AI chief John Gandrea is leaving the company.
1:35:58
Amomar Subramana from Microsoft is joined to lead AI under Craig Federigi.
1:36:05
And so uh we should dig in a little bit to this history.
1:36:11
So uh Swix has a little bit of a uh of a deep dive here.
1:36:15
He says uh Aman brings a wealth of experience to Apple. He's quoting here.
1:36:20
uh having most recently served as uh CVP of AI at Microsoft and previously spent 16 years at Google where he was head of engineering for Google Gemini.
1:36:29
Wait, oh, I guess at the end of that because Gemini is not 16 years old.
1:36:33
Uh this is bearing the lead.
1:36:35
He joined Microsoft AI four months ago. Wow, what a crazy turn.
1:36:40
>> LinkedIn says six months ago, but but who's who's counting? >> That's pretty fast.
1:36:44
And so, >> but this makes this makes sense considering Apple is partnering with Gemini and not a lot of people are going to be in a better position to help integrate that into Siri >> than Amar. >> Yeah.
1:36:57
I mean, I don't know, maybe there's something to uh, you know, just having a taste of all the different [laughter] big tech companies. Oh, yeah.
1:37:04
I've I've been at Microsoft. I know how they work. I've been at Google. I know how that works.
1:37:07
I'm I'm ready to ready to rock over here.
1:37:09
Uh, German does need to come on back on ASAP. I agree. Reg he was fantastic.
1:37:15
>> We'll get him in person hopefully before. >> So German continues.
1:37:18
He says uh strange hire for a number of reasons, but it's hard to argue that the Apple job is a bad one.
1:37:23
[laughter] Anything is an important point.
1:37:27
So the bar is as low as it comes.
1:37:29
Easy to lay up on the resume.
1:37:29
So that'll be it'll be fun to see.
1:37:31
be it'll be fun to see. I'm I'm personally just excited to actually test drive what Gemini how it works in Siri how how seamless that is because if it really is just raise press the button get Gemini and it's linked up properly
1:37:49
and it doesn't have timeouts and it gets back to you pretty quickly like that's going to be a pretty powerful experience that's that's definitely going to cut down on chat GBT app usage for iPhone users I would imagine >> underrated threat >> I would Think so. Like the re like the
1:38:01
Like the re like the there are so many moments where >> people are counting out Apple. >> It's not Yeah.
1:38:11
I don't I don't even know that I don't even know that Apple will benefit massively from this.
1:38:15
It's not like they're going to sell twice as many iPhones. They're already so big.
1:38:18
It's not like they're going to charge.
1:38:21
>> I don't think it's necessarily like especially bullish for Apple.
1:38:23
It's it's an underrated threat for OpenAI.
1:38:26
There's a lot of queries that I'll hit >> open chat GBT on mobile that are >> not even like super economic, but just a lot of my usage around like, hey, just trying to learn about something or or research a product, etc. Yeah.
1:38:41
>> And if that's just like again, one tap and you're in there. >> Yeah.
1:38:45
I mean, yeah, the the original promise of Siri was, you know, not just, hey, what's the weather today?
1:38:51
>> Uh, but really asking anything.
1:38:51
Gemini clearly solves that for 99% of knowledge retrieval queries.
1:38:58
Um I I would be I I think I'm going to be using that a lot unless they really botch it and I don't know how they're going to botch it but yeah. >> Yeah.
1:39:07
I mean I I think the >> because anything's possible, you know, >> Apple's like hold my beer.
1:39:11
>> They're going to be like every for privacy reasons every time you press the button you have to e sign.
1:39:15
[laughter] And it's like why are we doing that? Yeah.
1:39:19
Like uh the original like Siri kind of vision was like this very conversational AI, right? Yeah.
1:39:24
>> Um but I I don't think Gemini has a um real-time voice model yet.
1:39:27
Like I'm pretty sure OpenAI is the only one that has that >> really. Huh.
1:39:31
>> Um >> I don't think it matters at all. >> Really? >> Yeah.
1:39:34
I I really think >> that that um form factor feels like that would be the best thing to be on.
1:39:37
You press the button on your iPhone talking >> and then it's just on. Yeah. >> Yeah.
1:39:43
I mean, I wouldn't be surprised if they can if they can like get that model that version of the model out because it's really just like distilled a little bit faster.
1:39:52
It's not some like uncanny breakthrough that Gemini that the Gemini team will not never be able to crack, right?
1:39:59
So, they just have to build that.
1:40:00
But honestly, like I don't know that that's certainly not how I how how I would use it for most things.
1:40:05
Most things I would say, okay, like like I want I have I have one question.
1:40:09
get me an answer within a reasonable amount of time and maybe read it off to me or produce like an article that's, you know, a pretty readable article summarizing the answer to my question.
1:40:22
Um, and then yeah, may maybe there is like a back and forth, but I don't know. We'll see.
1:40:26
Oh, you're you're getting you're getting truth zoned in the chat.
1:40:28
Gemini does have a real-time voice feature. Gemini live.
1:40:32
>> Yeah, I think it's on the app. I've used >> Try that. Yeah, I don't have app.
1:40:34
So, >> large journalistic force headed towards you. Stand by.
1:40:40
Uh Tyler, Anthropic is acquiring Bun. What do you have to say?
1:40:46
>> Um yeah, I mean this is definitely in line with their like, you know, focus on dev stuff. >> Okay.
1:40:52
>> Uh >> what what what is bund?
1:40:54
>> Bun is a it's like a I don't know.
1:40:59
[snorts] It's like a bundler for JavaScript. It's like a very dev.
1:41:02
>> It has dramatically improved the JavaScript and Typescript developer experience.
1:41:06
They're going to make Claude code even better.
1:41:08
I like that Gabriel from OpenAI here is OMG in the chat, which is like uh a pretty crazy thing to [laughter] say.
1:41:18
Uh but I I I appreciate it.
1:41:18
So, uh Claude is one of the world's smartest, most capable AI models for developers, startups, and enterprises.
1:41:25
Cloud Code represents a new era of agent coding, fundamentally changing how teams build software.
1:41:30
In November, Cloud Code achieved a significant milestone just six months after becoming available to the public. That's crazy.
1:41:35
It's only been six months.
1:41:38
>> [laughter] >> uh it reached a $1 billion revenue run rate.
1:41:41
We were always struggling to understand what that meant, right? >> Yeah.
1:41:45
Well, so there there are two ways to pay for cloud code.
1:41:47
There's either with your cloud subscription where you get like cloud pro or cloud max and there's a certain amount of tokens you can use and then there's also you can just directly wire up APIs uh calls essentially to cloud code and then you're being charged like directly based on usage.
1:42:01
So that's probably what that revenue is from. >> Yeah.
1:42:05
And then yeah, I also have thought about this thing where like oh you can break down the number of tokens uh from the subscription.
1:42:11
So it's like your $20 subscription, three4s of your tokens are on cloud code.
1:42:15
So that means three4s of your $20 is counts as cloud code revenue. >> Okay. Yeah. Yeah. I'm not sure exactly. >> Yeah.
1:42:21
I mean I'm sure they can they can uh uh account for it.
1:42:23
Uh so was founded by Jared Summer Sumner in 2021.
1:42:26
Bun is dramatically faster than leading competition.
1:42:30
They say it's a breakthrough JavaScript runtime. Does it compete with V8?
1:42:32
I I'm I'm very interested in like uh it's Node. js. It competes with Node.
1:42:38
The thing that Tyler needs [laughter] to learn. >> What was that?
1:42:41
What was that company that OpenAI acquired earlier this year for like a billion?
1:42:46
>> I know the one you're talking about analytics or something. >> Yeah.
1:42:49
But again, I I remember at the time people were like, oh, like OpenAI's competitors are not going to be happy about this acquisition.
1:42:56
So >> that comment from Gabriel.
1:42:58
Um >> congratulations to the BUN team.
1:43:02
Congratulations to Anthropic and everyone on the Claude code team.
1:43:05
Uh very excited that you're getting to work together uh for your massive deal.
1:43:08
Uh speaking of other ma massive deals, >> speaking of size, >> Alfred Lynn >> hit the get that gong ready, John. What did he do?
1:43:20
>> Alfred Lynn comes in uh on the board of Door Dash, buys $100 million of Door Dash. >> Calling it Lynsanity.
1:43:32
He's not done stewarding Door Dash.
1:43:32
He's he's continuing to steward the company with a hund00 million buy >> and of course sends the stock up uh almost 6% on that.
1:43:44
>> Pretty pretty excited about it. >> He ripped. He ripped. Um what is this? Uh another 4.
1:43:50
6 million to donated to shrimp welfare. >> Okay.
1:43:55
So So they [laughter] Yeah.
1:43:58
Uh basically the story is um Enthropic they were doing some like research and uh about um uh smart contracts and so they had Claude code try to figure out like um you know issues in smart contracts and then I I'm not sure exactly where the like money came from.
1:44:16
Maybe it was for like bounties.
1:44:16
Um but there was some way in which Claude Code basically generated like $4.
1:44:20
6 million in like >> cash from uh finding these exploits.
1:44:24
So then they just >> did it actually generate real money or is this like the the hypothetical?
1:44:32
>> This is simulated testing. >> Okay. Huh. Simul. >> So So yeah.
1:44:37
So it probably means that they could have like basically stolen $4 million from people, but they don't want to do that.
1:44:43
>> Maybe they should have if they really want to get up the uh >> in other anthropic news, uh David Sax says he's still waiting on Daario's support uh after the uh New York Times piece was published.
1:44:55
Sam Alman of course came in and said, "David Saxs really understands AI and cares about the US leading in innovation.
1:45:00
I'm grateful we have him."
1:45:02
Of course, Daario and Sachs not the biggest fans of each other, so I don't expect that one uh coming through uh anytime soon.
1:45:12
Um while we wait for our first uh guest, Matt Garmin, uh let's pull up this clip from uh Huberman Lab. Uh if we can play this.
1:45:22
Uh Rob Moore is highlighting Dr.
1:45:26
Jeffrey tells Huberman that LED lighting in buildings is a public health crisis that could be on par with the use of asbestos.
1:45:33
Many building contractors/designers are coming to him worried they're going to be sued and asking how to start fixing the issue.
1:45:41
So uh let's pull this up when we have a second lighting because I am very concerned about the amount of short wavelength light that people are exposed to nowadays especially kids.
1:45:50
The group of us that are shuffling around, some of them are saying this is an issue on the same level as asbestos.
1:45:57
This is a public health issue and it's big.
1:46:00
LEDs came in and people won the Nobel Prize for this very rightly at the time because they save a lot of energy.
1:46:10
The LED has got a big blue spike in it, although we tend not to see that.
1:46:16
And that is even true of warm LEDs. And there is no red.
1:46:18
The light found in LEDs when we use them certainly when we use them on the retiny looking at mice.
1:46:26
We can watch the mitochondria gently go downhill.
1:46:28
They're far less responsive.
1:46:32
They their membrane potentials are coming down.
1:46:35
The mitochondria are not breathing very well.
1:46:38
Can watch that in real time >> under LED lighting.
1:46:42
>> Under LED lighting at the same energy levels that we would find in a domestic or or a commercial environment.
1:46:50
This is why I want to rig the studio with incandescent light. >> Incandescent.
1:46:54
We're going back to candles. >> Candle max. Let's do candle lights. How about a >> hearth? This is the way.
1:46:57
Uh >> if we put a hearth, so we have lights above our heads that I'm sure are LEDs killing us slowly and softly.
1:47:03
Uh if we put like somehat a bonfire right above us, >> that' be the way.
1:47:11
>> And then uh we just when when the wood kind of burns out, the show's over.
1:47:13
We just go until the >> That would be good. I like that.
1:47:16
Uh let me tell you about turbo puffer serverless vector and full text search built from first principles and object storage.
1:47:22
Fast 10x cheaper and extremely scalable.
1:47:25
Uh our turbo is at reinvent. >> Amazing. So >> fantastic.
1:47:30
Well, we are joined by the CEO of Amazon Web Services, Matt Garmin.
1:47:35
Thank you so much for taking the time to come and chat with us. How are we doing? >> Hi.
1:47:39
Thanks guys for having me.
1:47:41
>> Uh please take us through uh some of the highlevel announcements.
1:47:43
Obviously, it's uh it's reinvent. Very exciting.
1:47:48
Congratulations on all the progress.
1:47:48
Uh would love to know uh what's at the top of your mind, what's on the top of your uh presentations over the over the over the course of the event.
1:47:58
>> Yeah, we had a couple of really exciting announcements today.
1:48:00
Uh a couple I'd highlight.
1:48:02
First, we uh introduced these idea of frontier agents. Yeah.
1:48:05
>> Uh these are agents both uh in Kirao uh for software development as well as uh in operations and security.
1:48:10
And these frontier agents are meant to accomplish much much more than customers were a ever able to do uh in the past where we have these autonomous agents that can help customers really turbocharge their software environment.
1:48:22
So super excited about that.
1:48:24
Um we had some announcements around Nova which is our frontier um uh AI models that we announced.
1:48:29
We announced Nova 2 uh and our new sets of models.
1:48:33
Um, and one of the things I'm in particular really excited about, um, is Nova Forge, which allows customers to actually bring their own data to pre-training checkpoints, mix in their data with Amazon data, finish training the model, and at the end of it have a custom model that deeply understands their own enterprise data um, and is uh, and is just for them.
1:48:51
Um, so that that's another thing that I'm excited about.
1:48:53
Um and then the third thing is uh we announced a new chip around tranium 3 um to really turbocharge uh the next generation of training and inference uh for our customers and so quite excited to to get that and that went G today as well. >> That's very exciting.
1:49:08
I let's go back and start with the first one.
1:49:10
Let's talk about uh coding agents and uh the your own proprietary models.
1:49:15
How are you thinking about positioning those to potential buyers?
1:49:20
uh are you do you like the benchmarks these days?
1:49:23
Do you think that uh we're sort of like post all the benchmarks or do you think those are still useful tools uh for a buyer who's making a decision? Is it about integration? Is it about cost?
1:49:34
How are you positioning them?
1:49:37
>> Yeah, when you think about software development, it's it's not about pure benchmarks.
1:49:40
It's really about what is going to allow you to get the most amount of work done.
1:49:43
And when you think about our offering which is called Curo, yeah, >> um it's really focused on in a enterprise uh or environment where somebody's doing high velocity um development, they actually need more structure.
1:49:56
People love vibe coding and it's exciting, but you can actually get down a path where you get stuck. Yeah.
1:50:01
>> And you'll often find actually that you spend just as much time trying to get back to where you were before as if you had just coded it from the beginning.
1:50:07
>> We have this idea of specs that gives you structure to what you're trying to build.
1:50:11
And so you can have agents go and start to build around those specs together with you and your team.
1:50:15
And it gives you the structure that allows you to go really fast, can undo if you need to, um can make sure that you're hitting your design requirements.
1:50:23
And it and it really allows you and the agents to operate um in conjunction with each other and move really really fast.
1:50:28
Um and we're starting to build these much more capable agents that can go and actually do longunning tasks for you on your behalf.
1:50:36
But all of it is kind of ties into this structure.
1:50:37
And we view that as a way to deliver kind of real development that's going to be meaningful on a large codebase with large teams in enterprises um where they have existing things not just um kind of single individual people sitting there kind of doing vibe coding which you know you can do vibe coding on on Kira as well by the way we think that that's just not sufficient for what makes development's going to need. >> Yeah.
1:50:58
and uh talk to me about what it actually looks like to set an agent off and say, "Hey, I got a task for you.
1:51:05
Come back to me uh in a few days," which it sounds like that's where we're going.
1:51:10
Uh we've been tracking the meter benchmark and it seems like we've been seeing doublings there, but again, a lot of those have been the benchmark has been h how long would it take a human to do this task?
1:51:21
The actual agent might have done it faster.
1:51:23
Um, and so you it's not necessarily that you're actually letting something cook over the weekend.
1:51:30
Uh, what's the experience been like and and what have people been reporting about uh these longunning agents?
1:51:33
about uh these longunning agents? Yeah, I think the first and actually most important thing is thinking about how you actually kind of have a mind change on how you think about software development where you think about not about do this task, get it back, look at it, do this task, but how are you thinking about directing a lot of agents to go out there and do lots of different things and and let those run for long
1:51:52
periods of time where they can kind of have amorphous tasks like instead of go write me this function like try to go solve this problem for me and then it'll come back and uh and and then but you can but if you send out two or three or
1:52:04
10 or 20 or 50 of those things then your job as a software developer and as a product leader is actually much more around coordinating those when they come back troubleshooting make sure that you know directing them course correcting etc. Um and so I'm excited about that.
1:52:16
Um and so I'm excited about that.
1:52:18
We've already seen these um these processes go off and work for multiple hours at a time um on on particularly like really hard tricky amorphous tasks and um and we think those things are are going to continue and be more the norm of how software developer teams change what they accomplish.
1:52:32
Yeah, >> we think hero is going to be the engine that's going to drive a lot of that. >> Yeah.
1:52:35
Yesterday we were talking to uh Vincent from Prime Intellect and they do some of this like fine-tuning on smaller models and he has this thesis I think that you share that uh a lot of businesses will need to take a a pre-train and then and then bring their own data fine-tune it not just because it's important from performance and output but also from cost.
1:52:58
But I'm interested in understanding um how you think the market will shape out.
1:53:03
Do you see implementation partners and like consulting firms coming in and doing that?
1:53:10
I was asking him like uh >> yeah, >> you know, there's a lot of tech startups that are going to be able to do that.
1:53:15
They're going to understand I need to build an RL environment around my app.
1:53:18
Uh but for larger legacy companies, they might not understand.
1:53:21
So how are they going to wind up uh using that tool in particular?
1:53:26
I I think they will and and actually just want to highlight one piece there where some of what we announced today. Yeah.
1:53:31
>> Um is a little bit different. We announced this idea.
1:53:32
It's it's an open training model with Nova.
1:53:34
Um and so the difference and what you just said is people take a pre-trained model and they'll do RL after the fact and they'll try to do some some fine-tuning.
1:53:41
Um which is great, but there is actually limits to where that does.
1:53:44
In fact, if you do too much post- training, often times those those models will forget what they've done at the beginning.
1:53:50
They'll start to lose some of their reasoning and their core intelligence. Yeah.
1:53:53
Yeah. I mean this is an unsolved problem um except when you go and insert your data in the pre-training phase and so what we do with Nova is we expose checkpoints you can take a 60% trained or an 80% trained um model pre-trained model insert your data into that pre-training phase mix it in we then
1:54:11
expose actually Amazon training data to you via an API that you can then mix it together and so it's like you said here's my all my corpus of corporate data here's everything that I need to know about my industry We then mix that in and then [clears throat] finish pre-training the model. So you get a
1:54:25
So you get a pre-trained model that totally understands your company and your data.
1:54:30
And then you can go do fine-tuning.
1:54:30
You can go do reinforcement learning gyms.
1:54:34
After that you can shrink them down and distill them.
1:54:35
You can do all those things but on a pre-trained model that deeply understands what your company does.
1:54:40
>> And is that called mid training now?
1:54:40
Is that the right buzz word for that?
1:54:44
>> It's not like we're and mid training is a different thing.
1:54:45
the first time that anyone's ever exposed this idea to to to deliver pre-training checkpoints where we can mix in your data.
1:54:52
No one's ever done this before. It's first time. >> Uh, great. Yeah. Well, then yeah.
1:54:55
On on market structure uh do you think it's self-s served enough that you know large corporations will do it or do you need like do you need an AI lab?
1:55:06
Do you need an AI scientist?
1:55:07
Do you need someone to who can you know write TensorFlow or PyTorch or something to implement this or is it something where you know just a normal software engineer at a large company could go and pull this off the shelf and implement it? >> Yeah, we we'll see.
1:55:20
I think we're going to keep working on the tools today.
1:55:21
Um I do think that for some enterprises they'll want to have some consulting um folks that help them with this.
1:55:27
I think we'll have some some people where you have some experts that can come and and teach how to do this.
1:55:31
And I think we'll quickly get the tools to a point where you know it's not somewhere where uh you know a non-technical person is going to go do this for sure.
1:55:38
But um but it may be a software developer that that tends to be a little bit more on the the AI or ML side um that we hope is going to be able to go do this without having to have a whole bunch of expertise about how to go pre-train a frontier model. >> Yeah.
1:55:52
Uh on the cost side obviously you're working you announced a new chip.
1:55:57
Uh I imagine that there's you know the emergence of some synergies across the models that you're developing the software you're deploying the cloud and then also the chips.
1:56:06
Uh how are you positioning the like the tranium ecosystem?
1:56:11
Is this something that you're you're planning on really doubling down on across the entire stack?
1:56:16
Uh or do you want to be more chip agnostic?
1:56:19
Are we going to see you buying TPUs in the future?
1:56:23
Uh, no, we we definitely um well, a couple of things that that there to unpack.
1:56:28
The first is we're very excited about Tranium and think it has enormous potential and we absolutely think there's a benefit to optimizing every single layer of that stack where we have um the best cost performance um that we can deliver at at Trrenium.
1:56:40
We have optimized models for you to use and applications and agents at the top of that that we talked about.
1:56:47
>> So, we think that whole um optimization of that stack is going to be critically important.
1:56:51
And of course, we're gonna support choice for our customers as well.
1:56:54
And so we'll continue to offer um GPUs from Nvidia as an example and um and we have a very tight partnership there.
1:57:01
But but we do think and we're quite excited about what Trrenium 3 is going to offer for customers.
1:57:06
And I do think that we're going to see an explosion of that ecosystem as more and more people um get access to those chips and are able to take advantage um of the pretty significant cost performance benefits that you can get from running on training.
1:57:17
How are you thinking about uh opensource the open source ecosystem that you need to build around tranium?
1:57:25
That's the big discussion with the TPU right now.
1:57:27
The question of you know Google has some amazing folks.
1:57:32
They have some amazing uh software folks.
1:57:34
It seems like uh they don't necessarily need to open source everything.
1:57:39
Uh and so a lot of people are waiting to see how much the the industry, you know, builds open source alternatives independently uh versus how much does Google just give away.
1:57:47
What's your thought process on building a an open- source ecosystem or even just giving developers access to closed source software to run efficiently on Tranium? >> Yeah.
1:57:59
No, we're we're all in favor of having a an open set of software to run on Tranium.
1:58:03
In fact, we've we have our uh Neuron uh SDK um which is open source today and allows everyone to to contribute to that.
1:58:11
We we think that the the more that we can collaborate on that software ecosystem to make it easier for people to to use chips and we of course support um the broad set of whether it's PyTorch or or other kind of open frameworks as well.
1:58:23
Um so we collaborate across the industry on that and and are big advocates of um contributing to and uh and supporting that um open ecosystem.
1:58:33
Jordy >> uh love to get uh your insight on just like general constraints for for AWS as a business.
1:58:40
What you guys are doing on the power side is that is that a real constraint?
1:58:44
Uh anything that you can share there? >> Yeah.
1:58:48
Uh you know, look, it's um as we're scaling incredibly rapidly, we've um you know, we recently announced that we've added um 3.
1:58:53
8 gawatts of data center capacity in the last year alone, which is just an insane amount of data center capacity. >> Thank you.
1:59:01
Um, [laughter] oh, you're welcome. I don't know.
1:59:03
And uh and and so it's it's ramping incredibly fast and it's it is a constraint.
1:59:08
You know, we have more demand than we have supply today for uh for AI. Sure.
1:59:13
>> Um and as we ramp up the supply chain, we think about all of the constraints.
1:59:18
We think about chip constraints, we think about networking constraints, we think about power constraints, we think about networking constraints, um data centers, etc.
1:59:24
And so we're we're working really hard to to try to remove every single one of those.
1:59:28
And uh when with an industry that's growing as rapidly as the AI one is, there's always going to be some constraint and uh and we work really hard to keep removing blockers every every time so we can keep growing fast. >> Makes sense.
1:59:41
Um well, we have a hard stop.
1:59:43
So, thank you so much for taking the time on such a busy day to come chat with us.
1:59:47
Uh would love to have you back on the show and go way deeper, but thanks so much and uh congratulations on all the massive releases.
1:59:52
We're excited to dig in deeper and keep chatting about them.
1:59:56
Uh but have a great rest of your day.
1:59:58
We'll talk to you soon on coming on. Cheers.
2:00:01
>> Um, let me tell you about public. com.
2:00:03
[music] Investing for those that take it seriously.
2:00:04
They got multiass investing and they're trusted by millions.
2:00:06
Um, we have Take Kim, author of the Nvidia way and a Baron's senior writer joining the show.
2:00:16
Take him the author of the NVIDIA way >> for joining the show.
2:00:20
And I'm sorry it [music] took us so long.
2:00:22
Uh we've exchanged uh posts on on X many times and uh we wanted to have you on the show earlier, but I'm so glad you got to ask right away.
2:00:31
>> We did because it's the perfect time to talk to you.
2:00:33
>> Do you have roommates? [laughter] >> Not. >> No roommates.
2:00:39
>> Guys, longtime listener, first time caller.
2:00:42
I'm so excited to be on this.
2:00:44
>> I'm so excited to have you here.
2:00:45
Reminder, everyone, go buy the book.
2:00:48
Seriously, uh Christmas is coming.
2:00:48
I can't imagine a better gift for everyone.
2:00:53
>> For a four-year-old for even >> I know what my son's getting.
2:00:54
He's getting the Nvidia A Jensen one teachable copies. >> Multiple copies.
2:01:00
No, seriously, get get 10 copies.
2:01:02
Give them that teenagers.
2:01:02
A lot of teenagers have have read it.
2:01:04
Like it's amazing and they reach out and uh it's an inspirational entrepreneurial book that a lot of parents are giving to their kids.
2:01:13
So definitely >> I love your uh your headset by the way because we're actually developing our own TVPN over here headset because this is just like this is the this is the ideal ideal setup.
2:01:26
>> I was telling my friend I don't care if I look like a dork.
2:01:27
My hearing is going >> so like I get locked in when I have the headset I can hear everything. >> No. And it's wired.
2:01:34
The worst the worst is AirPods. AirPods have a tiny lag.
2:01:38
Like you're doing Zoom calls. It just it it ruins it.
2:01:40
You feel like you feel like you're right here at the table with us.
2:01:44
>> It's not code red over there. He's Baja blasting.
2:01:45
You know, he's Baja blasting.
2:01:48
Um anyway, uh let's let's uh we we we I mean we have some time.
2:01:52
Let let's run through uh I want [snorts] to know a little bit more about your perception of Jensen, your perception of Nvidia, and just set the table for us.
2:02:01
We know how what is his management style?
2:02:04
How does he he has all the direct reports?
2:02:06
He reads everyone's like to-do lists every day.
2:02:10
They have tons of employees. They never fire people.
2:02:12
Like what makes Nvidia's culture unique? Set the table for us.
2:02:17
So then we can go into the opportunities and challenges with that framework in mind.
2:02:22
>> So So the first thing I found out about their culture is it's very blunt.
2:02:26
>> Like I think in most companies and you guys have done startups, but I don't know if you work for large corporations.
2:02:32
>> Um bureaucracy builds up process it gets oified. Yeah.
2:02:37
>> Uh, Nvidia is the complete opposite.
2:02:39
Like things are not going well, he'll chew you out in front of the whole company.
2:02:43
>> And that kind of blunt mentality, I think, you know, sparks better performance because you don't want to be embarrassed in front of Johnson in front of the whole company. Yeah.
2:02:51
But [snorts] also it it just sparks an agility like uh when I talk to people at Intel or Google like the biggest problem they have is meeting paralysis and you need to get signoffs from like five different executives at Nvidia like you have a meeting Jensen makes a call he seeks out the right information and you move.
2:03:11
So there's this agility at Nvidia.
2:03:13
Uh the the other thing is just a meritocracy.
2:03:17
Even from the beginning, like 30 years ago, Jensen's always asking who's the smartest person that you work with?
2:03:21
Who should I try to recruit?
2:03:25
>> And uh from the beginning, like Dwight Durks, uh who's one of the top people right now, he he recruited him because he talked to this other guy and he said, "Oh, Dwight's really smart.
2:03:35
Like I I really enjoyed working with him."
2:03:37
And he just almost ruthless in a sense.
2:03:39
He just goes after them and and recruits them, brings people aboard.
2:03:43
So this meritocracy, agility, speed and just getting rid of the internal politics I think really separates Nvidia.
2:03:53
>> Uh how has most of the team uh becoming millionaires affected the culture >> if at all?
2:04:00
>> Um I think a they had that thing in in the first 1015 years like people started getting sports cars and putting them in the parking lot but a lot of people >> cars let's go that's the best news I've ever heard. That's fantastic.
2:04:11
Sounds like it's had an incredible impact on the culture. >> Job's finished. Okay.
2:04:15
Yeah, you can end you can.
2:04:17
You don't need to respond.
2:04:17
You don't need to say anything.
2:04:19
>> I think [laughter] winning culture breeds winning and people want to stay with winners. Yeah. Right.
2:04:25
>> You want to win on the track then.
2:04:27
>> A lot of people I meet in terms of colleagues, they work for a company for five years and the chip doesn't work out. Yeah.
2:04:33
>> And you just wasted five, seven years of your life.
2:04:34
So, you want to stay with a winning company and Nvidia has been winning for 30 years.
2:04:38
So, it it it's kind of like winning begets winning.
2:04:40
kind of like winning begets winning. you get the talent and then then the talent stays like there's so many top executives at Nvidia that have stayed there for 25 30 years and >> yeah there's also there's there's there's some benefit of people when if if if an somebody doesn't have a
2:04:54
scarcity mindset right and they're just playing to win like they're just like they're they're they're they're no longer thinking like oh if we can just get to that next milestone and get this secondary sale and if I if I can participate and if I can vest my two years and sell into the next uh tender offer then I'll you know they're just like we're good. All that matters is
2:05:11
All that matters is just being as elite as we possibly can be and just uh and doing it uh for the love of the game basically.
2:05:21
>> And a lot of these people that like they made it they could retire they could have retired 10 20 years ago but they stay and work still work 80 hours 100 hours a week because that's what's expected.
2:05:31
Like even the marketing people inv 85 hours a week and that that kind of mentality I think is different at the companies. Mhm.
2:05:39
Okay, let's uh let's shift into the competitive dynamic.
2:05:42
I mean, Nvidia's been uh we I was revisiting the performance of the Mag 7 since the dawn of ChateBT.
2:05:48
It's been three years exactly.
2:05:50
Nvidia by far the winner, up 10x uh on market cap.
2:05:54
Uh the next closest company, I think maybe 4xed uh by comparison.
2:06:01
And so, uh, the clear AI winner in the public markets, the most obvious AI trade that just completely ripped.
2:06:08
Now, you know, there's this whole narrative of like, uh, how strong is their moat?
2:06:12
Uh, what is what is the TPU mean?
2:06:15
Is the TPU going to be significantly competitive?
2:06:19
Is there going to be margin compression?
2:06:21
How have you been processing this new narrative that uh Nvidia might face serious competitive threats because they're so on top of the world that everyone owes them so much money [laughter] that people are saying I got to get a discount from somewhere and I'll go to Google maybe
2:06:38
>> I want to talk about this 10x move like it hasn't been like straight up to the right there's always been >> every three six months there's always a reason to sell Nvidia like >> the H100 problem the transition to Black Whale, uh, China, ASIC, Broadcom competition. So, this this stuff has
2:06:54
So, this this stuff has been happening this entire 10x move up and the media loves to latch on to the latest thing to worry about, right?
2:07:02
We had deepse earlier this year that the entire media establishment was it's over for the AI tree.
2:07:08
A AI models have become so efficient when it was actually the opposite because the reasoning models and there was an exponential demand for computer.
2:07:17
So I I find it amusing like the whole world kind of discovered that Google had a really good chip in the TPU >> which they've been working on for a decade too.
2:07:27
>> Yeah, [laughter] they've had it for 10 years, right?
2:07:29
They've offered it to uh clients 2018. This is nothing new.
2:07:32
The Ironwood specs, you know, which I always take with a grain of salt with specs.
2:07:39
Even even Nvidia, they talk about 25x improvement with Blackwell when it's more like >> I'm really just focused on the name Ironwood is goes pretty hard. It's pretty good.
2:07:47
They they got some good names over there in Google.
2:07:51
>> Ironwood came out all the specs came out in April.
2:07:54
Like this is not all this stuff is isn't new.
2:07:57
And uh another thing I want to say is TPUs their chips.
2:08:02
Morgan Stanley estimates there was a huge decline in TPU shipments in 2025.
2:08:07
And Google at Google Cloud Nvidia GPUs took more share than TPUs this year.
2:08:14
It's like no one talks about this, right?
2:08:15
And now everyone's going Ironwood is going to take over the world and Nvidia's in trouble.
2:08:20
>> Is that just because we're >> I mean I think it was a Gemini 3.
2:08:21
It was a No, it was a Gemini 3 thing.
2:08:23
People are like Gemini 3 is the best model in the world. >> Yeah.
2:08:28
>> On by the world for a whopping six days and chatbt is still number one on the app store.
2:08:33
Let's >> But I mean six days it was unseated by Claude which was also anthropic which is also TP potentially in the future. >> Yes.
2:08:41
>> Yes. and chat GPT um Microsoft the head of Microsoft AI cloud code refer that open eye is training um their next models on the GB300 and the L72 uh that what just went live in October actually earlier today the Nvidia CFO said it's going to take six months so I was a
2:08:59
little disappointed in that >> six months for the training run >> yeah well she said the first models on Blackwell like on the superclusters are going to take six months so >> it's a singularity >> [Β __Β ] It's going to be another another >> hop AI is going to get there. The the
2:09:13
The the Claude and Gemini uh benchmark gains with uh pre-training. Yep.
2:09:19
>> That's the most bullish thing for the whole AI industry, right?
2:09:21
AI AI adoption is the scaling laws are intact.
2:09:27
Everything's going to work out and open AI is going to get there when they uh build their next training uh on the next model.
2:09:33
So, so going back to TPUs, uh, thank goodness, a shout out to semi analysis.
2:09:40
They do the best channel work in the industry.
2:09:43
Everyone freaked out on Friday, right?
2:09:45
They read that semi analysis. No.
2:09:47
Oh no, total cost of ownership.
2:09:50
>> They're going to destroy everything.
2:09:50
But like people that actually know the industry, >> it was flaming bullish for Nvidia.
2:09:54
Like it it just it was like so obvious in my face because uh Dylan and and the Sammy analysis they said, "Wait a minute.
2:10:03
The next TPU V8 is not going to be that great.
2:10:07
They lost a ton of people and the the set function up in performance is not going to be that great."
2:10:12
So you know what's going to be great?
2:10:14
Nvidia's Vera Rubin, which comes out at the end of next year.
2:10:18
>> So no company is >> People are saying it's the Rick Rubin of chips.
2:10:21
[laughter] >> Are they related? >> Yes.
2:10:25
it and so anyway, so Bar Rubin is going to be dramatically better at the end of next year and even >> you know the the Ironwood which just became generally available and they're ramping right now.
2:10:34
Um it's it's it's no one's going to switch over for one it's a huge uh endeavor to put workloads uh from CUDA Nvidia GPUs and put them on there's always problems when you put put them on on a new chip. >> Yes.
2:10:50
>> U and let's talk about TPU customers, right?
2:10:53
Everyone freaked out that Meta might spend a few billion dollars in 2027.
2:10:58
>> That sounds like a lot, right?
2:11:00
>> That's less than 1% of Nvidia's uh expected revenue. >> Sure, it nothing.
2:11:05
>> And Ben Thompson was very smart and astute.
2:11:07
He's like, who's going to buy the TPU?
2:11:08
Who are the biggest buyers of AI chips? >> Yep.
2:11:12
>> They're the hyperscalers, right?
2:11:12
So, so Met maybe Meta will put a portion of their workloads 1% Nvidia's revenue.
2:11:19
>> Is Amazon going to buy TPUs?
2:11:19
Well, John just asked the CEO of AWS if he was going to buy TPUs. He dodged that question. >> He didn't say yes.
2:11:27
>> I mean, there's no way in hell they're going to buy TPUs.
2:11:29
They have their own tranium.
2:11:31
They're not going to support their number one like one of the number one >> arch rival.
2:11:37
>> Yeah, they're not doing that.
2:11:37
So, is Microsoft going to buy?
2:11:40
>> You should have been like, "Yes, I'm going to buy one so I can like study it."
2:11:42
[laughter] >> Microsoft's not going to buy TPUs.
2:11:46
They're the number two player in cloud computing and they're not.
2:11:48
Are the neoclouds gonna buy TPUs?
2:11:49
Now you're gonna say yes, Google got has some neoclouds.
2:11:54
You know what happened with those Neoclouds?
2:11:56
They're financially backstopping those Neoclouds.
2:11:57
So Google is financially giving money and and backs stopping the debt for those Neocloud.
2:12:03
So So there's a handful of small NeoClouds, but is Core Weave going to buy TPUs? Probably not, right?
2:12:10
>> Who are the other customers of AI chips?
2:12:13
>> Enterprises, companies, sovereign AI. Yeah.
2:12:15
Anybody that wants to run like a fine tuned model, some small model, something like that. >> Yeah. >> 90.
2:12:22
[laughter] Um, >> so like if you just go down on a first principles basis and look at the customers of AI chips, like they're going to stick with Nvidia.
2:12:32
The millions of developers know CUDA.
2:12:34
So you don't have you really need like Dylan talked about this is you really need like top-notch like software sophisticated engineers that can like work with TPUs and learn learn jacks and all that stuff.
2:12:47
So most people aren't don't have those cracker jackack engineers, right?
2:12:52
So they're going to stick with Nvidia because everyone's used to Nvidia.
2:12:55
Nvidia is backwards compatible and forwards compatible.
2:12:58
So like 20 30 years of this stuff. >> Mhm.
2:13:01
And if you buy it, Nvidia Collat app the CFO talked about this morning.
2:13:08
It's um you can use it for training and you can use it for inference.
2:13:10
It's all on the same architecture and it's going to work.
2:13:13
Like I I I talked to an AI startup CEO a few months ago. He tried AWS training.
2:13:18
Oh, it looks a lot cheaper.
2:13:22
>> Total cost of ownership, but then it crashed.
2:13:24
There were bugs, the reliability, they they couldn't figure out what happened.
2:13:27
And there's like they just threw up their hands. I give up.
2:13:30
Like no one is going to like if you have reliability problems, bugs, crashes, the best thing about Nvidia is all that stuff has been ironed out over the last 1015 years.
2:13:39
If you have a problem, you can figure it out because >> yeah, it's like giving giving an F1 giving an F1 driver like a car that that is unreliable and saying like, "Hey, go race, go race, have have fun out there."
2:13:51
And then it's like, you know, >> and the specs DNF immediately, right?
2:13:56
>> Specs specs look awesome.
2:13:56
It look seems great, but then when actually build your business on it, you you put the future of your business onto something.
2:14:02
You the number one thing, it's not price.
2:14:04
It's like it better [Β __Β ] better work.
2:14:08
[laughter] It better work. >> And it works. >> Yeah.
2:14:12
But what what uh react to the this idea that Dylan Patel was uh joking about as uh TPU is a stocking horse.
2:14:17
So this idea that Sam Alman is already saving 30% on his Nvidia purchases effectively because just the threat of going to TPU is enough to get Nvidia to make an investment or slightly discount in one way or another.
2:14:35
I >> I don't think that's reality and I don't think that math actually works because I think he's confusing the AMD deal where AMD gave, you know, free warrants to Open AI, >> right?
2:14:46
First of all, the deal is not done. >> Sure.
2:14:48
It's a letter of intent has >> none of these deals are done.
2:14:50
They're they're not done.
2:14:52
>> AMD AMD's AMD is done.
2:14:52
They they they signed an agreement where they're giving away a percentage of their company through these free warrants.
2:14:58
The Nvidia deal hasn't been signed yet.
2:15:00
And they actually language in the Tank Q that it might not happen.
2:15:03
Doesn't OpenAI have to buy AMD chips in order to get the warrants? >> Yes. Yes. Yes.
2:15:11
>> And so it's still it still could be that they don't actually end up going through with the purchase and then they wouldn't get they would point. >> Yeah. Yeah.
2:15:16
So, so the whole like let's do a side step here with the circularity and all that stuff.
2:15:22
>> All this stuff it's like one gigat at a time.
2:15:25
There's a milestones on open AI there's milestones on AMD technical milestones they have to achieve certain targets.
2:15:31
So all this talk about, you know, everyone loves the the big number that adds up five years of capex a lot like that it it's it could get the leverage could be up or down depending on how things happen every each year of the way.
2:15:48
So you know it might not be that big number if open AAI doesn't come out with an amazing model or AMD isn't able to hit the milestones they said for their next 450MI chip, right?
2:15:57
So like it, you know, don't worry about five years.
2:16:03
Like take it one year at a time.
2:16:05
Right now demand is off the charts.
2:16:05
Now going back to the Nvidia 30% discount, like that's not how equity investments work.
2:16:12
If Nvidia does invest 10 billion, 10 billion up to 100 billion.
2:16:15
Say that say that >> Nvidia gets ownership of the company.
2:16:20
It's not like a freebie, right?
2:16:20
You're giving away uh ownership of your company.
2:16:24
So it's not really a discount.
2:16:26
you're getting you're getting uh uh ownership of the company.
2:16:28
So I I don't really believe in this 30% discount thing because um Nvidia Jensen will say they're they're investing to accelerate open AI and uh they would they they they're looking forward to you know open AI going >> I mean it's definitely creative.
2:16:41
It's definitely a new structure.
2:16:43
I'm just trying to I would I would steal man in that like if I'm an entrepreneur and somebody comes to me and they're like I'm going to invest $und00 million in your company over a series of milestones and you're also going to buy something from me.
2:16:56
I'm like yeah I'm taking some dilution but realistically like this is a way less of a headache.
2:17:01
Like where else was I going to get hundred billion dollars from if in in O OpenAI's case like it's a great it's a great source of funding that yes it will be diluted but the whole structure is all diluted all the time because of all the different ownerships >> except for this kind of sentiment thing that we had the last few weeks.
2:17:19
So open AI hasn't had any problem in raising money for venture capital. >> Yeah. Yeah. It's true.
2:17:24
It's not just it's not like Nvidia is the only source of funding for open AI like everyone wants in the revenue run rate it's like 5 billion to 20 billion at the end of this year.
2:17:36
>> Um >> so what was your take on the code red what do you think about the code red?
2:17:41
>> So uh I saw that you you you showed that Ashley Vance uh >> interview really interesting Chen was talking about how they kind of focused a little too much on reasoning and their pre-training muscle wasn't there.
2:17:53
Um I reasoning we could talk about this later.
2:17:58
Reasoning is like the biggest kind of accelerant of AI demand in the past year.
2:18:02
So I I think it's actually really good and supposedly Ilia was you know doing the research for reasoning.
2:18:08
Reasoning is awesome, right?
2:18:08
But they they kind of focus on reasoning the the past year with 01 and 03.
2:18:14
>> Um and now they're like okay we have to go back to pre-training.
2:18:16
So, o OpenAI knows that pre-training still works because Gemini 3 had great pre-training results and and Cloud Opus uh 4. 5 did.
2:18:27
So, now they're going to do the pre-training.
2:18:30
>> So, they had their focus on one thing and now they're going to do the other thing and make their their model much better.
2:18:35
I I do agree that Open AI has been a little too maybe >> diluted like they're doing apps. >> Sure.
2:18:43
>> They're doing hardware.
2:18:43
They want to compete in AI infrastructure against Microsoft and Oracle.
2:18:48
They want to compete in AI chips against Nvidia.
2:18:50
Like I I I thought it was really interesting like Satia repeatedly said um he wouldn't name who he's talking to but it's like >> I think it's important that we realize this is not a zero sum game and this could be a win partnership.
2:19:06
That was during the anthropic Nvidia uh deal with Microsoft. Right. Sure.
2:19:10
He said that a couple times and I think the the person that he's talking to is Sam >> Alman, >> right?
2:19:18
>> Let's let's Nvidia, Microsoft um made Open AI as successful.
2:19:22
They were they were the partners like why are you competing with your the partners that brought you to the dance, right?
2:19:28
Let's let's go back >> focus on making the best AI model in the world and uh don't compete with Nvidia and Microsoft.
2:19:37
maybe maybe 5 years from now, but like it seems a little aggressive to compete with them uh right now.
2:19:43
>> Let's talk about China.
2:19:43
Uh there's been a ton of debate over Nvidia selling chips to China, legacy chips, older chips.
2:19:50
Um we've gone back and forth on it so many times.
2:19:54
Uh what's your current thinking about the the the the best policy for Nvidia exporting chips to China?
2:20:04
Generally, I >> I think the best thing is to keep it one or two generations behind >> uh the current state of the art. >> Yeah.
2:20:12
>> Like this is a really nuanced policy that people, you know, everyone's either hawkish or dubbish. >> Totally. >> Whatever.
2:20:17
The best policy is to keep I don't want to use the word that Howard Lutnik used that got China very upset and forced [laughter] >> forced China to like tell his companies not to buy H2.
2:20:29
>> Um but the best policy is to get China still on the Nvidia stack.
2:20:34
So Nvidia gets $50 billion of revenue per year that can help R&D and fund R&D and make the chips even better.
2:20:44
Like >> Nvidia and the US already won, right?
2:20:49
>> They have 95% market share.
2:20:49
>> They have 95% market share. like why are we going to give $50 billion of oxygen to Huawei and all these other Chinese AI chips uh companies that now Chinese uh companies that need to buy AI chips are going to buy Chinese AI chips like why not keep China on the Nvidia tech stack
2:21:10
one or two generations behind don't give them the best stuff but maybe one generation behind I think that'll be the best compromise um >> for for for both sides but I don't know >> what do you think what do you I think what do you where do you place a likelihood that uh that the Chinese market has opened up again at at some point in the next 12 months? >> Maybe 50 5050. I mean that sounds like a >> Maybe 50 5050.
2:21:29
I mean that sounds like a copout.
2:21:32
Like I was much more positive 6 months ago, but >> um you know this has been just so crazy.
2:21:38
First the Trump administration banned the H20. >> Yeah.
2:21:41
>> Then they didn't ban it. They said it was fine.
2:21:43
But then China was like, "No, you hurt our feelings.
2:21:45
Um we're not going to let companies buy the H20." >> Yep. And then they ban it. >> And then >> Yeah.
2:21:50
And then maybe Trump is going to let Nvidia sell the H200 >> or a Chinese specific version of Blackwell that's kind of like hobbled a little bit. >> Who knows?
2:22:00
Like Nvidia needs to convince the Trump administration and then China to to buy the chips.
2:22:06
The the the worst part of it is China was willing to buy the H20 and it's just all the kind of geopolitics and hurt feelings.
2:22:14
Um you know that that ship has sailed.
2:22:17
So, I don't know what's going to happen.
2:22:19
Um, but I do think it the ideal situation is Nvidia could uh sell one generation behind, make $50 billion a year and and keep u the competition from ch Chinese AI startups uh out of the way. >> Yeah.
2:22:34
>> And even understanding that at some point in the future, China's buying effectively zero chips from Nvidia, but it would be 5 10 years in the future.
2:22:45
Um >> it's like you have to you have to assume you have to assume that like >> right now >> go for it.
2:22:54
>> Uh the the amazing thing is [laughter] sorry uh it's 0% right and Nvidia's revenue accelerated for the first time in two years.
2:23:03
Nvidia's revenue accelerated in this latest quarter and this is like not talked about enough.
2:23:10
This is the first quarter that the NVL72 AI server has been available in volume and then revenue just skyrocketed without China which is incredible right and that's why I'm so bullish over the next few years because next few quarters let's say that >> because this product cycle is going to
2:23:27
last at least three four quarters um the the key tell is the revenue acceleration first time in two years and and the other key tell is that the networking segment for Nvidia was up 162% year-over-year a year and typically a lot of these data centers um and these neoclouds buy the networking stuff 6 months ahead of time. So the next six
2:23:45
So the next six months from now like the the the GPU numbers for for Blackwell and the NVL72 server is going to be it's just going to be bonkers.
2:23:53
It's going to be off the charts.
2:23:55
And people don't talk about these NVL72 AI servers.
2:23:57
They're 3 to4 million, right?
2:24:00
There's 72 GPUs, 144 dies, um, uh, one and a half tons, 5,000 cables, and the, and the prior version was 8 GPUs.
2:24:09
So, so these these AI servers are I I call it the iPhone 3G moment.
2:24:15
Do you guys remember the iPhone 3G?
2:24:16
Like, >> this is big for the Christmas shoppers out there.
2:24:19
Uh, if you if you want a gift, it's a step up from the Nvidia way, I recommend.
2:24:24
Or you want to bundle something, >> picking up an NVL72, >> it's only $4 million. >> Yeah.
2:24:30
But for the right 12-year-old in your life or or the potentially the intern, I think Tyler, >> it won't fit under the tree, but it could be fun. Keep it in the garage.
2:24:39
>> It's like it's like a Lexus.
2:24:39
You put it in the in the in the driveway with the bow on top. That's the way it does.
2:24:45
>> With the fork, you drive the forklift. Come drop off. >> Exactly. I got you an NVL72. Enjoy.
2:24:50
[laughter] Um and and and then we have the reasoning model thing where exponential compute and like companies are actually seeing like huge cursor 40% productivity gains. >> Yeah.
2:25:02
>> Uh CH Robinson 40% shipments.
2:25:05
>> Um >> rocket mortgage 80% reduction in paperwork cost processing.
2:25:08
Um this is like the next year because of AI reasoning because of the MDLMD2.
2:25:17
That's why uh Amazon and Microsoft said every quarter this year they raised their capex and everyone's like they're going to cut their capex.
2:25:24
They're no every single quarter they raise their capex.
2:25:27
That's because they're seeing the demand and that's why Amazon and Microsoft are going to double their data center capacity over the next two years.
2:25:34
I mean that that that's crazy, right? >> Yeah.
2:25:37
>> In September quarter more leasing there are more data centers leased than entirety of 2024.
2:25:41
This is like exponential step function up and and people aren't talking about it.
2:25:47
They they want to talk about TPUs like destroying Nvidia.
2:25:51
>> What about talking about >> what what about some of the kind of demand guarantees that have been happening?
2:25:57
Is that a is that a concern at all?
2:26:00
Do you think about it much?
2:26:00
at all? Do you think about it much? Is it >> I not really I mean demand guarantees like you're talking about Nvidia and coreweed it it's like when that happens analysts every every quarter or on the conference call like did you use that
2:26:14
you know demand guarantee like it's not happening yet like coreweave's 5-year-old GPUs that everyone says are useless are 100% us utilized right H100 a massive cluster before uh it expired probably a three-year contract they got like 95% % of the pricing. This is like This is like unheard of.
2:26:33
And and the reason is there's overwhelming AI demand and there's not enough capacity.
2:26:41
Overwhelming AI demand, not enough capacity.
2:26:43
And and and people just are are just they don't care about what's happening in the real market.
2:26:48
This is real life facts, evidence, numbers.
2:26:53
Nvidia going from 56% revenue growth to 62% revenue growth on 57 billion with zero China re I mean these are bonkers numbers [laughter] we we talk about the stock price being up 10x their revenue is up 10x in like two years this is like
2:27:11
beyond history the last 30 years of following technology >> I love how you're the you're the only person without Nvidia fatigue you're just like you're not bullish >> David Gogggins of the >> [laughter] >> I I mean they can't keep running. They They can't keep running.
2:27:26
This is not me just like making stuff up.
2:27:28
This is like the numbers are there right in front of me.
2:27:31
>> You make you make good points. You make good points. I I like it a lot.
2:27:33
Uh we'll have to have you back on the show soon. This is a lot of fun. >> Yeah, let's do it.
2:27:37
Let's let's make this a regular thing.
2:27:39
Super fun to have you on finally.
2:27:42
>> Thanks so much for all your time.
2:27:42
Uh the book is The Nvidia Way.
2:27:46
Get it at wherever books are sold. Get 10 copies.
2:27:51
Give it to everyone in your life.
2:27:51
Also, give the gift of Finn.
2:27:55
AI, the number one AI agent for customer service.
2:27:58
Automate your most complex customer service queries on every channel.
2:28:03
Um, >> up next, >> oh yeah, we we can just go straight into our next guest.
2:28:10
Uh, >> let's bring in >> we have Tark from Kelshi with some massive news. Tark, great to see you.
2:28:16
How are you doing stream? Good to see you.
2:28:21
>> Hey guys, thanks for having me. Very excited to be here. >> You are locked in. Look at that backdrop. Fantastic.
2:28:24
Uh please uh introduce yourself. You've been introduced. Give give us the update. What's the news? Let's ring the gong.
2:28:33
>> Uh well, we just raised our series E.
2:28:33
Um we just raised a billion dollar billion [laughter] valuation. >> Great, great wind up. Great wind up.
2:28:41
>> Honestly, I was waiting for the gong. >> Congratulations. >> [Β __Β ] sick moment. >> How's it going, guys?
2:28:46
>> Uh yeah, great to have you on.
2:28:46
Uh I don't I was thinking over the I don't know if anybody had a crazier uh Thanksgiving holiday than you.
2:28:54
It was there was a lot there's a lot going on last week.
2:28:58
So nice nice to come out of that with a with a with a big announcement.
2:29:01
But um but yeah, maybe maybe kind of just update us on um uh I think everybody's has been following the prediction market wars.
2:29:09
The the more important story I think is just how >> some people are calling blood bath actually. >> Yeah.
2:29:15
[laughter] I mean just like it's been a battlefield on the timeline.
2:29:17
But um but yeah, I think like the what's happening in the background is like this explosion of this you know new asset class that um you know again I think uh in your announcement earlier you were saying >> few years ago there nobody really cared at all and now it they you know you and and the the industry broadly have million millions of users.
2:29:36
So um it's pretty unprecedented.
2:29:38
Um but yeah, what what's what's uh what's been the latest on on your mind?
2:29:45
I >> mean, I think the the thing that's happening right now is prediction markets, I think, have gone mainstream.
2:29:50
Um I think every inch of evidence is pointing towards that.
2:29:52
And I think that the >> the thing that we're seeing is there's sort of one of these rare shifts in consumer behavior that you you don't see often like they they don't happen like changing the behavior of a customer, the habits of a customer is is a rare thing and it's unique.
2:30:07
And when you you see it, you have to really go after it with all your might.
2:30:10
Um, and it's, you know, there's like a number of things that have to align for that to happen.
2:30:15
And I think they're aligning for prediction markets.
2:30:17
I think it's it's happening.
2:30:19
And I think there's, you know, one factor is the fact that people are not really trusting um the sort of legacy media and legacy sources of information and they go to prediction markets to get smarter.
2:30:29
The other one is that they're legal now.
2:30:31
you know, Kashi has took on this sort of battle over years to legalize this entire market and and you know, set it up as a legitimate financial asset so that anyone can participate.
2:30:40
Um, and three, I mean, I think we're all kind sort of we sort of caught wildfire um this this this year.
2:30:47
I mean, I think the um we're seeing people there's a little bit of this phenomena where you cannot watch a sports game without looking at the KI odds live uh and the KI charts.
2:30:55
you cannot talk debate about a topic about the future without um you know uh uh talenting somebody to put a position on Kali on the app.
2:31:04
So um it's it's a huge announcement.
2:31:07
We're very excited about it.
2:31:08
Um and it honestly really feels like we're just we're just scratching the surface of what prediction markets can be.
2:31:14
>> One thing I've noticed when uh I'm watching a sports game is there's sometimes an integration with KHI, sometimes with a competitor.
2:31:21
Uh what's actually going on?
2:31:24
>> You know, legacy sports book. >> Yeah.
2:31:26
What what is actually going on?
2:31:26
I feel like a lot of people who are just passively observing the timeline are seeing a lot of like announcements and partnerships with >> the partnership economy >> the part and people are joking about it like what's actually going on?
2:31:38
What's at stake with some of these partnerships?
2:31:41
What have you done and what does it actually mean?
2:31:42
Because it feels like if you do a partnership with a specific league that doesn't necessarily mean that I can't get odds on that event somewhere else.
2:31:51
So what what is actually going on with the partnership economy?
2:31:56
Um I mean I'll tell you kind of our approach to this.
2:31:58
So so we are building you know our focus is building on a business.
2:32:02
It's very metrics driven you know and sort of for context.
2:32:04
So we're doing a billion and a half of volume a week now. >> Wow.
2:32:08
>> Um and you know we're market leader by meaningful margin.
2:32:11
I think depending on sort of how you measure it.
2:32:12
So we're something around 80 to 90% market share now.
2:32:16
And I think any partnership we do, we bucket them in a bunch of categories, but they're all focused on like actually driving legitimate volume and legitimate use case into the product.
2:32:25
So our partnership with, you know, platforms like Robin Hood, I mean, Coinbase leaked, it's coming in December.
2:32:29
Um, uh, and Price Fix, Weeble are kind of in that bucket.
2:32:34
Then we have partnerships with >> um, a series of partnerships coming around news.
2:32:38
Um, one of them leaked this morning in the New York Times article.
2:32:42
Uh but they're also very [laughter] >> one sentence 10 leaks.
2:32:48
>> Everything leaks these days.
2:32:48
I you know we just like nothing is news anymore.
2:32:51
It's like sort of you know it's it's all leaks.
2:32:53
But but the point is >> we're focused on things that drive legitimate use to the products.
2:32:56
Um and and and then drive legitimate utility to >> uh the partner.
2:33:04
And so you know whether it's a broker obviously you know this could be a big revenue line for them.
2:33:09
And if it's a news network, it's a complement to the reporting that actually makes the reporting more accurate.
2:33:14
And you know, um, reporters love truth and prediction markets bring truth.
2:33:17
So you could see the synergies and how they fit. >> Okay. Yeah. Yeah. Yeah.
2:33:19
That makes a lot of sense. >> Uh, yeah.
2:33:22
What uh yeah, I think the uh some of the some of the big news out of last week is that Robin Hood is is entering uh and kind of potentially trying to verticalize the product experience on their side.
2:33:36
What can you say about the I guess like how you see the structure of the market evolving?
2:33:42
You guys are an exchange.
2:33:44
Robin Hood is a brokerage.
2:33:47
Sounds like they're trying to actually build uh an underlying exchange themselves.
2:33:50
Uh how much should how much should uh sort of observers of the industry look to how the how stock trading and stock markets, stock exchanges have evolved versus prediction markets?
2:34:04
like what does this market kind of look like in 5 years, 10 years as much as you can uh kind of pull out a crystal ball for us?
2:34:13
I mean maybe the basics is like and you've seen this a little bit in AI right after you see the success we've had um it's it's basically indicative of like okay there's a massive market opportunity ahead of us um and when that happens I think you're going to inevitably see a ton of competition um
2:34:32
and generally in those markets like the the the sort of massive surge of competition whether it's brokers there are some of the sports books like Draxkins and FanDuel coming in Um it's it's just usually a sign that there's a lot of good things to come for that market, right? It's it's a sign that
2:34:45
It's it's a sign that like you have big companies rep prioritizing their entire road maps to go all in after this.
2:34:49
Um and that's a positive for us like we're market leader in a market that you know everybody is starting to believe is going to be ginormous.
2:34:57
ginormous. Um in terms of the specific question of market structure I mean like you know we we have obviously the exchange we also have our direct product um in some ways are is competitive with some of our partners and I think you know the same way that we're working with a lot of different brokers over time some brokers are going to sort of
2:35:13
diversify and work with number of different exchanges um and that's how these sort of market structure evolve over time um and the only thing that matters that kind of the thing that stands out is similar to any other market is product and product velocity is are you putting out products faster than anybody else and are you putting putting out products better than everybody else. And I think Kashi has
2:35:29
And I think Kashi has had a pretty incredible track record of setting the pace in the industry.
2:35:33
At least if you look at the last year, we've set the pace in the industry and everyone's following and I feel pretty good about us continuing to do so in the next 24 months.
2:35:40
>> How do you think about the market structure?
2:35:41
I think everyone's wondering like obviously this is a new market.
2:35:45
It's unlocking entirely new sort of asset classes.
2:35:48
Uh and and it's it's obviously big.
2:35:52
Everyone's excited about the numbers, but uh is this a natural monopoly? Is this duopoly?
2:35:56
Like how many winners will there be?
2:35:58
How do you even think about the market structure?
2:36:00
Is there some return to scale?
2:36:07
>> It's it's interesting like I I kind of like don't think much about that.
2:36:08
Like I think investors love to sort of investors do this thing where they're sort of going to rationalize all of it in five years.
2:36:14
You know, everybody's going to be super smart about how they like all figured out.
2:36:18
>> But like look, I I think that like it's a very nent thing, right?
2:36:19
It's it's a it's you know like it's it has some similarities to ride share.
2:36:24
It has some similarities to the drafting fanduel era when that happened.
2:36:29
It has some similarities to the online brokerage industry >> and it has some similarities to financial exchanges like CME.
2:36:36
>> So where does it fall?
2:36:36
It probably somewhere in between all of these sort of buckets.
2:36:39
>> Uh and probably not exactly the same as any of the any of these other buckets.
2:36:43
Um >> um and I think that you'll see more of um >> I think with enough scale in financial services, but also true for any industry, everyone gets into everyone's territory.
2:36:53
And so well the only thing that matters again is sort of what companies are going to rise above the others in terms of product velocity and product quality. >> Yeah.
2:37:00
>> Um and I that's just what we're narrowly focused on.
2:37:03
>> There's a question from the chat.
2:37:03
Uh can you explain how external market making works on Kelshi >> that's been a for some reason a hot topic recently but [clears throat] you know market makers are part of any financial market.
2:37:14
you you kind of need them um to basically have liquidity in in markets and um actually Koshi and prediction markets have less customer to market maker flow than traditional markets.
2:37:26
If you look at options for example, it's like the vast majority is >> you know Jordy to a market maker like Citadel whereas on cash actually the vast majority is you know Jordi versus John and then some of it [clears throat] goes to market makers and it's an open transparent order book where everyone's competing on price.
2:37:41
>> Um and we have actually a separate company called cash trading that trades on the exchange but they're very small percentage of any liquid markets.
2:37:47
Really their function has been for new markets are a little bit weird. >> Liquid markets.
2:37:52
Yeah, that makes sense cuz it's if it's some really really niche thing, who's going to put in the first 500 bucks?
2:37:56
Like you take the risk >> and they're not very profitable.
2:37:58
It's actually we really like they're really focused on providing a good customer experience so that we bootstrap markets. >> Yeah.
2:38:05
>> Um rather than like any meaningful part of the business model today.
2:38:08
Um and if we took it out, it'd be actually a worse experience.
2:38:11
So I I think it's definitely not positive for the ecosystem.
2:38:13
But it's a bit like Uber, you know, when the adverse interests got impacted. >> Sure.
2:38:19
>> The taxis, they were coming up with all these reasons, right?
2:38:20
like you know about all these kind of random reasons but I I don't think there's much truth to it. >> Yeah.
2:38:25
So so uh I'm sure you can't comment on any specific lawsuit.
2:38:27
Uh there's there's a number of them.
2:38:30
Uh but uh what what what has been the uh I think there was quite a lot of prediction markets experts that have looked at some recent lawsuits again against prediction markets and said uh they they clearly don't understand how this works.
2:38:46
like can you comment at all on some kind of like misunder misunderstandings broadly? >> Yeah.
2:38:51
So we what what KCI has done is first regulate prediction markets as a financial instrument uh under this agency called the CFTC.
2:38:59
People have been hearing more about the CFC recently because it also regulates crypto. >> Mh.
2:39:04
>> Um and that's one of the main financial agencies.
2:39:06
There's the SEC that does stocks and CF that does commodities.
2:39:10
>> Um and then we did the same thing with elections and now we did the same thing with sports.
2:39:13
Um, and the way that it works is like financial markets, those are regulated at the federal level.
2:39:18
And so the law around these markets is just federal.
2:39:23
They they kind of report to a federal government and federal regulator, not a state government and state regulator.
2:39:27
And there's a bunch of reasons for that, but you know, it's kind of how the constitution was formed, which is some stuff makes sense at the federal level, and some stuff is more local and makes sense at the state level.
2:39:36
And we are one of these things that fall under the federal level and federal law preempts state law.
2:39:41
So if you are okay on the federal side, state law doesn't really kind of apply to that exchange.
2:39:47
Um and that's why we have one regulator which is the CFTC, our federal regulator.
2:39:51
And again, like I think it's normal with like when something so disruptive happens to an industry, the people that are adversely impacted are going to come after it and come up with all sorts of arguments for why it shouldn't exist or why, you know, Airbnb was terrible and all these different things.
2:40:05
But at the end of the day, the thing that drives it longterm is is this a great product and are consumers loving it and using it and the answer is yes in those ca in this case.
2:40:14
>> Um off of the success of Khi and Poly Market, there's been a bunch of net new prediction market startups that are created.
2:40:24
Is there a possibility that that uh this market like ends up having these sort of like niche uh maybe more like vertical marketplaces or do you think that the platforms with the greatest liquidity and and the deepest liquidity will will uh ultimately just absorb those submarkets.
2:40:45
Um it depends on how narrowly we define prediction markets versus broadly like I I really think of prediction markets as kind of just like a next gener like like expansion of financial markets to touch anything.
2:40:56
Calcium means everything in Arabic but really if like if markets kind of progressively grew over time what we did is just like kind of widened that set dramatically over what it could touch.
2:41:08
So I could see some you know startups innovating on like specific verticals over time and doing reasonably well but there is real concentration of liquidity and concentration of volume that happens in those type in in these types of markets.
2:41:20
Uh that is hard I would say to battle with um and so I think at least from from that aspect like I think that the cards are probably mostly shuffled uh already. >> That makes sense. >> Last question.
2:41:33
Um, there was a viral clip of you talking about uh Donald Trump Jr.
2:41:37
is do you have anything more to share on his involvement?
2:41:39
Uh, because I was watching that and I was like, "Yeah, there it's kind of like uh, hey, where are we going with this thing?"
2:41:45
Uh, it seems like politicians have a deep insight on how campaigns use these uh, prediction markets, but uh, can you share anything more about his involvement in the company?
2:41:57
>> Yeah, I mean, look, f first of all, that clip is a clip and you know how these clips are taken.
2:42:01
Uh but you know >> who needs context?
2:42:05
>> Yeah, it's [laughter] like we don't need context.
2:42:06
context. It's a completely you know but anyways look I think that um >> uh you know uh we have done like one of the main products that took us mainstream was an election market and that brings a lot of attention from politicians on both sides of the aisle and you see it you know Trump at the time was using his prediction markets all during the election >> and actually Mamani more recently was using his calcia odds pretty
2:42:28
consistently during his election >> and and so in some ways like you're going to see a lot more like prediction markets are going to touch financial markets going to touch the news and going to touch the political process because they bring more truth to all of the above all of these categories and in some ways it's good that like we get more and more I would say like um politicians involved and like engage with these markets. Um the one thing
2:42:47
Um the one thing I'll say about this and like >> again it's it's very it's in the same bucket as as the you know as as the other things that we discussed where like there's industry dissident that are against prediction markets that find all these different reasons for why prediction markets might be bad.
2:43:01
prediction markets might be bad. But the thing that happened is not this administration necessarily even though this administration is pro innovation is we won that lawsuit on the election market which has really redefined what the landscape what the boundaries of what the financial market is and that lawsuit was one you know was is a is in
2:43:17
the court of appeals in DC with relative it's a very progressive panel it was a a panel of democratic judges where we won three zero so people want to make it out to be a partisan issue even though I I don't think truth needs to be a partisan issue it's just you know uh these markets people love them and they generate a lot of insight out of them. Um, and I think that will win the win
2:43:35
Um, and I think that will win the win the day at the end of the day.
2:43:38
>> Last question from my side.
2:43:38
How does the CFTC view when a market participant uh has some type of alpha or or non-public information and they're uh they're they're betting they're betting on a market, you know, based on that information.
2:43:55
From my view as somebody who like gets data from, you know, we work with Poly Market, we we look at we use Poly Market data on the show.
2:44:02
If somebody has sight information and they're they're they're trading on that information, it actually makes the markets more accurate.
2:44:10
So, in some ways, as a user who's just like viewing markets, it's I want people that have inside information on global events to be trading so that the markets actually better reflect reality.
2:44:23
Uh, but what is like the CFTC's view on that type of activity because like things get thrown around all the time, insider trading this or that, but I don't actually know like what the actual law says. >> Yeah.
2:44:38
>> Yeah, that's a great like that's actually a great question.
2:44:39
It's a it's a point of debate um in in this land, but I think there's some distinction.
2:44:44
So, so cash is a regulated exchange.
2:44:45
So everything we do in some ways a lot of the laws and the rules are very similar to what you would expect in a New York stock exchange in some of the traditional financial markets.
2:44:54
Um the question of insider trading is interesting because what you just said could also apply to the stock market, right?
2:45:00
Like if you want to accurately price a stock, maybe we should let insider trading happen. Sure.
2:45:05
>> And the reason why it's actually not allowed is because it makes the game unfair.
2:45:10
It makes the market unfair.
2:45:10
And if the market is unfair, liquidity dries up.
2:45:15
people just stop participating. >> Yeah. >> Right.
2:45:17
And and that's why you have to have reasonable rules of the road where people can reasonably expect to be treated fairly in this marketplace where there's no kind of asymmetric uh or structural advantage for for one participant versus the other.
2:45:29
And we take the similar approach here.
2:45:31
So if you actually have insider information, which is information that like you're not supposed to reveal to the public, you're not supposed to trade on it because trading on it is a way to reveal it to the public.
2:45:42
Um and and so and so that makes kind of the more balanced, more fair marketplace and I think we're very focused on that.
2:45:48
Um but it's a very interesting question.
2:45:51
It's it's one the industry is battling with.
2:45:52
But we we take a hard stance on insider trading. >> Yeah.
2:45:55
Because if somebody goes and and uh they go and they vote in a local election and they see like, okay, I talked to I talked to somebody there and they said they were voting this way and I talked to another person, they all says that they were voting this way.
2:46:07
And then somebody trades on that information.
2:46:10
like is it actually like is is that you know how how how do you define that type of activity, right?
2:46:18
It's like anybody could go down >> to the polling, you know, any anyone could go down to the polls and and kind of like uh uh uh or or voting center and just see like ask the same question, right?
2:46:30
So anyways, uh >> well, I was going to say is it's the same as the stock market, right?
2:46:34
If you go and sit in front of Walmart and count everybody that's going in and out and then you know during the day and forecast their sales from that, that's actually fair game.
2:46:41
Now, if you call your cousin at Walmart and ask them for information they have internally that they're not supposed to reveal to the public, that's inside of trading.
2:46:47
And I think we have a very similar line here. >> Yeah. Yeah. Yeah.
2:46:50
That makes that makes sense. >> Um uh very cool. Well, super helpful.
2:46:52
Um and yeah, congrats to the whole team.
2:46:57
It's pretty massive milestone. >> Huge.
2:47:01
>> And uh yeah, great great getting the update.
2:47:03
Thanks so much for taking the time to help us. Thanks for having me.
2:47:05
We will talk to you soon. >> Talk soon. >> Have a good one. >> adquick.
2:47:08
com out of home advertising made easy and measurable.
2:47:11
Plan, buy and measure out of home with precision.
2:47:14
Our next guest is Matt Mulleng from automatic.
2:47:16
He is in the reream waiting room.
2:47:19
Let's bring him in >> parent company of WordPress. com.
2:47:26
Tumblr [music] >> I think are going to remain.
2:47:28
>> Welcome to the show [snorts] Matt. >> How are you doing?
2:47:31
I think we have you on a hot mic.
2:47:35
>> We might have you on a hot mic. Hopefully not. Welcome to the show.
2:47:37
>> They're about to come on the screen. >> Yes. >> All right.
2:47:39
Well, a little little drum roll experience TVPN.
2:47:44
>> So, Matt, our audience.
2:47:46
>> We're streaming on them. They're streaming on us. How are you doing? >> We'll clap, too. Fantastic. How are you doing? Good to meet you. >> Howdy. Howdy.
2:47:54
[laughter] Uh, thank you so much for I know this is a little non-traditional, so we're we're kind of like two hours into like our big annual address, the state of the word.
2:48:03
It's kind of like our state of the union speech and um but thank you so much for allowing us to connect them.
2:48:07
I'm kind of imagine a lot of folks in the room have never heard or seen TVBN before.
2:48:11
So, this will bring a lot of new folks into your world and I'm excited for some of your world to learn about WordPress. >> Yeah.
2:48:18
Give me the state of the word.
2:48:18
Uh and then also I want your your personal word of the year.
2:48:23
We've been debating what the word of the year should be over here. >> Oh.
2:48:28
>> Uh, so state of the word.
2:48:28
And I'll say the state of the word is strong. >> Okay, >> there we go. >> That's good. Let's hit the gong.
2:48:34
>> We're hitting the gong for that one.
2:48:39
>> The strong state of the word.
2:48:42
[applause] >> Congratulations.
2:48:44
>> We actually just did a live release of WordPress 6. 9.
2:48:46
So WordPress does major releases three times per year.
2:48:49
We were able to do it right here on stage.
2:48:51
We had a little button that we pushed.
2:48:52
We got to get it going next time. >> I love it.
2:48:56
>> That was uh It was pretty fun.
2:48:56
Don't worry, I didn't just ship it again.
2:48:58
It's [laughter] >> But um you know uh one of the things about WordPress is is it's a it's not just built by one company, but it's a community of in WordPress 6.
2:49:07
9 over 900 contributors from all over the world, different countries, different languages, different companies, all coming together.
2:49:13
And so that was pretty exciting.
2:49:15
Uh my word of the year and actually a theme we were just talking about is I'm going to choose freedom.
2:49:19
about is I'm going to choose freedom. Uh so powerful >> as technology like starts to influence more and more of our lives you know how we travel who we date the things we learn the news we're exposed to um you know the sort of freedoms that are embedded in an open source license I like to refer to open source licenses
2:49:37
sort of like a bill of rights for software um gives you inalable rights that no company or person can take away from you and that freedom and agency I think is really really important and something that um I think you know as technologists or builders that we should try to embed into everything that we give us an update on Beeper. I was super I was super fascinating.
2:49:54
I was super fascinated by that product.
2:49:56
Uh I love I love walled gardens.
2:50:00
I also love tearing down the walls of gardens.
2:50:01
Uh it seems like a a good shot across the bow of the uh the iMessage uh walled garden.
2:50:08
Uh how's the progress going there?
2:50:11
Are you using the service personally daily?
2:50:13
Is are we going to see a lot of growth there?
2:50:18
>> Um well, obviously I'm using it daily.
2:50:20
Um, so I would think of it not as like a replacing a wall garden, but more like allowing your gardens to come together.
2:50:27
>> Um, so I'm sure you like me, I have friends on lots of different networks, and some of them always love to use WhatsApp, and some of them always love to use, you know, Instagram or LinkedIn DM.
2:50:37
Sometimes I even get some interesting stuff there.
2:50:40
>> Um, and I hate it when I miss these messages, you know, because, you know, checking all the different apps sometimes or in the notifications I might miss something.
2:50:47
So, think of it not unlike how email clients, you know, can bring in lots of different email accounts.
2:50:52
Beeper takes all the different networks where your friends already are and maps them together.
2:50:55
Um, now the plus and minus is that you're it's not going to replace the networks.
2:51:00
Like I still keep all the different sort of specialized messaging apps because like for example, if someone sends you an Instagram story, when you click on that, you're going to want to load Instagram for example.
2:51:10
So, I think of it as complimentary and hopefully even increasing the usage in a very small way right now.
2:51:14
It's pretty nent, but in the future, think of it as like sort of a different interface.
2:51:18
So, you might still have like the dedicated apps, but then having this all in one inbox that you can sort of manage everything, uh, tag people, have folders, and does cool features like scheduled messaging across all platforms, or even just like weird huristics that are pretty simple to do, but like show me all the not don't just show me unread, but show me all the people I've messaged that haven't messaged me back yet. >> Oh, yeah. Sure. Sure.
2:51:41
uh we we've talked to some young hackers, some startups uh who are building, you know, sort of Bieber competitors and their whole value prop is like we've figured out a way to get it into the iMessage ecosystem.
2:51:52
Uh do you think that uh we need a new regulation there or some sort of law change or some result to actually open up iMessage or do you think that uh with enough tricky hacking it can be done?
2:52:09
Um well technically it's it's not hard.
2:52:13
Well it is hard but it's very possible to reverse engineer these networks.
2:52:17
>> Um however as we saw with sort of a previous iteration of Beeper >> if uh the network really really doesn't want you to do that >> um it's probably not good to pick a fight with a trillion dollar company. >> Yeah.
2:52:30
So um perhaps these things might happen through open source or something but as a commercial company I think ultimately you have to be somewhat respectful and try to complement these networks.
2:52:40
>> Um so how beeper works today is we don't support iMessage on the mobile or Android.
2:52:46
>> Um >> in theory we could but Apple has indicated that's something they don't want.
2:52:49
We do support on the Mac OS client.
2:52:52
We have a way to integrate with sort of iMessage using some APIs that are available in Mac OS.
2:52:55
And so on Mac OS we can bring in your iMessage.
2:52:57
Got Um but again I'm building this for the long term and we are a commercial company as well.
2:53:05
>> So um you know we we want to work with the networks and um you know perhaps there can be uh regulations like the European DMA or things that can encourage interoperability.
2:53:14
Um but ultimately I think that the the sort of people who run these networks have to see a longerterm benefit for them and for things like um you know some of the other networks I mentioned that Bieber works with I think um their business model and everything the increased usage is really useful for them.
2:53:34
I think for today Apple's business model of particularly in the US kind of the lock in effect to the device business which is of course where they make uh a lot of money from iMessage probably indicates that unless forced to I I doubt uh they will adopt u sort of iMessage interoperability uh but who knows sort of like they used lightning for a while and eventually got USBC and all of our lives got better um who knows what'll happen in the future.
2:54:02
>> Talk about links on the internet.
2:54:02
I feel like we're at a point in time where social media platforms are trying to keep users in in their own applications so that they can monetize them to the fullest extent.
2:54:14
Meanwhile, you have LLMs which are ultimately doing a lot of the same thing.
2:54:19
They're taking content from all over the internet trying to keep users in the individual applications.
2:54:22
uh feels like WordPress in in many ways is uh uh making moves to kind of like almost fight back against that.
2:54:32
I might have that incorrect, but I feel like it's important if you're running a business independently online, it's great to have people like on your own website so you can develop a a deep uh relationship with them.
2:54:45
Uh but what what is your view on that?
2:54:47
We're very much anchored around X as a as a business.
2:54:52
Obviously X has had uh issues with links or you know chosen to um demote them in the algorithm over the last couple years but uh give us kind of the state of the union on on links. >> That's a broad one.
2:55:03
Well, I will say X is actually a great example and I've I've talked to Nikita about this.
2:55:07
So they now um they've shifted some of the balancing of links and they now have this really nice in sort of app browser.
2:55:16
So, you've probably noticed that now that when you load a link, you actually still have the ability to like like and reblog and everything.
2:55:20
And I think that's kind of the future.
2:55:22
Um, so I I do think that there you can have things that are complimentary because so much of the like great content and everything is more on this open web.
2:55:31
It doesn't have to be like fully embedded in an app.
2:55:34
Um, but that is sort of a technological change.
2:55:36
So, I would say actually point to X as some place where I think things are going in the right direction.
2:55:40
Although I do agree that sort of time when links got really deboosted and everyone had to do it as like a reply was kind of weird and sucked.
2:55:48
>> Um so for for WordPress publishers, you know, we support so many different types of websites and different types of websites I think might have different motivations.
2:55:55
So for example, um a popular plugin for WordPress is called Woo Commerce.
2:56:00
It's an e-commerce plugin.
2:56:02
It actually runs on about 8.
2:56:02
9% of all websites in the world are now running this e-commerce plugin.
2:56:06
You can think of it like an open source Shopify.
2:56:07
And when you if you're selling something a merchant >> um you don't you just want to sell the product you don't might not necessarily care that someone comes to your website to buy it.
2:56:18
>> So some of the new things that are happening with in partnership with OpenAI and others where we're allowing products to actually be like browsed and bought inside of the LLM are pretty exciting.
2:56:28
>> Um I also think that the incentives of these uh open source chat bots in particular um are very complimentary to the open web.
2:56:35
So, for example, like if you're on Amazon, Amazon really wants you to say or eBay or Etsy or something like that, they want you to stay in their marketplace on their system.
2:56:43
But when you think of how Google works and sort of the growth of Google in the open web, you know, they they have their search pages, but they also would link out and that was whole part of their business model and how they grew.
2:56:54
Um, we're seeing that with the chat bots as well.
2:56:57
And in fact, something I talked about a little bit earlier is that the traffic from bots, both from them crawling, but also user initiated actions is exploding and has already surpassed sort of human traffic and it'll be interesting to see where that goes in the future.
2:57:11
So, you know, there's never a better time, I think, to invest in having a domain, but also invest in publishing.
2:57:16
And, you know, just like you might have a direct relationship, like for example, I suppose I could get like a, you know, chat GBT to summarize today's TVPN episode, but it's more exciting to watch it.
2:57:26
I think that creators developing a direct relationship and brand is going to be um part of the future as well. >> Very very cool.
2:57:32
Well, there's so many more things that I want to ask, but uh I know you're in the midst of of your own presentation.
2:57:38
So, uh thank you for for tuning in.
2:57:41
Come back on uh soon and uh thank you for having us.
2:57:44
The the view is spectacular as well.
2:57:46
So, [laughter] >> it's a pleasure for me.
2:57:48
I love to come down and hang out when we're I'm in LA next. So, >> fantastic. Thanks so much. We'll talk to you soon. >> Great chatting.
2:57:54
>> Have a good rest of your day. >> See you. Hi. >> Uh, first. com.
2:58:00
>> Exceptional sleep without exception.
2:58:02
Fall asleep faster, sleep deeper, wake up energized. That's our sponsor.
2:58:03
Uh, we have >> 94 last night.
2:58:07
John, I think I smoked you again. Lost my phone. Lost your phone.
2:58:11
Well, we have Jason Freed in the ream waiting for bring to the TV. There he is.
2:58:17
Jason, how are you doing? Good to see you. >> Good. And you? >> Congratulations. Massive news today. Break it down for us. What's up?
2:58:24
>> Was there big news today? I I missed the news. What was the news? >> Oh, your news.
2:58:26
>> You're you're just calling out every news. >> Trello naming names.
2:58:28
He's [clears throat] naming names.
2:58:30
A lot of a lot of people don't do that.
2:58:31
A lot of people say, "Oh, the competitors, the best-in-class solutions, the Gartner hype cycle." No, you called them out.
2:58:38
You you put them on the map. >> We had some fun. Yes.
2:58:40
So, we launched a new product today called Fizzy, which is kind of a a fresh take on Conbon. An old idea.
2:58:46
obviously been around for a long time, >> but we, you know, we have a different spin on things, different take on things and felt like it was time to do something new and kind of bring it back to the basics and also add some fun and color and vibrancy, which is missing in the software industry.
2:58:58
I feel like um the people might be colorful in a sense, but the products are very much the same and so we wanted to do something different and um that was uh that was what we did today. >> Why new name?
2:59:08
Why not uh you know a new new tab in an existing product, >> right?
2:59:14
Well, Base Camp, which is our biggest product, has uh conbon in it.
2:59:16
We call it card table there.
2:59:19
But, you know, the thing is is that um Base Camp is very popular, but it's, you know, let's say there's 100,000 accounts, right?
2:59:27
100,000 companies use it.
2:59:27
It's a small number in the end.
2:59:31
>> And um there's a lot of people who can use something like Fizzy that are not going to use Base Camp.
2:59:34
Base Camp is a much bigger system.
2:59:36
It's for bigger projects and there's a lot of small things that people need to do and organize and track.
2:59:41
And so building a small standalone thing just feels like it makes more sense frankly for for this kind of thing.
2:59:48
>> So do I mean do you have an idea of like who is the target market startups individuals like like you use this to plan your Thanksgiving dinner?
2:59:56
>> Yeah I mean the target market is me and us basically we build things for ourselves.
3:00:01
I don't think about who we're making things for because we're making things for us always.
3:00:04
And um the idea is that you know >> well actually let me just say this.
3:00:09
I I find the best products in the world are made by the person who's making them for themselves.
3:00:13
That's been my experience, like enthusiast products.
3:00:17
>> And then other people find them and other people discover them and you find out that you're like other people and other people are like you and they kind of >> dig it, you know?
3:00:24
And so >> I I always think about I said to someone this morning that that I feel like TVPN is that way where sometimes when I'm driving home, I want to watch I want to I want to watch TVPN, but I'm like, we just we just made it. I just lived it.
3:00:37
I should probably, you know, watch something else.
3:00:40
So, I I never I never watch watch the show myself, but I >> just call me and say, "Hey, we >> just make we do a podcast on the fly.
3:00:48
>> We just talk about tech news more."
3:00:48
Uh I'd love to I'd love to know about uh the actual process for building the product.
3:00:55
Uh who who was staffed on the team? How many people? What time period? When did you start?
3:01:00
Do you have a designer, developer?
3:01:03
Is it all just what's the prompt?
3:01:04
I imagine you just use one prompt for this.
3:01:06
What prompt was all all was >> all you needed. [laughter] >> Yeah.
3:01:11
Um so you know it's what's interesting is we actually also open sourced this.
3:01:14
So this is fully open source.
3:01:16
It's a SAS product and fully open source.
3:01:18
So you can run it yourself for free which means you can go into GitHub actually and look back at the very first commit about 18 months ago >> and see everything we did along the way.
3:01:30
All the changes we made, all the dead ends, all the starts and stops, exactly who was involved on our team over time. And it's changed.
3:01:36
So we had typically we have two designers, one or two designers on something.
3:01:40
Then there's other people who chime in here and there who jump in here and there.
3:01:43
Different programmers jump in at different times.
3:01:44
But it's fully documented, which is very rare.
3:01:48
You'll almost never ever see this in commercial soft. Basically almost never.
3:01:51
Sometimes, but almost never, especially going back to day one.
3:01:54
What ends up happening is you can do this thing where you can basically on launch day you can clear the log basically and then from that point on people can see what you're doing.
3:02:02
But we opened it up from day one about 18 months ago.
3:02:04
So, it's actually all in there.
3:02:06
Um, the team sized in total probably about six people worked on it here and there over 18 months, but for the most part, it's usually two or three people working on something at a time.
3:02:19
>> How do you think about pricing these?
3:02:21
>> Yeah, I feel like as as in in uh 37 signals fashion, pricing will be opinionated.
3:02:27
So, I'm excited to hear uh how you how you guys approach this one.
3:02:32
I you know we don't really well we have a price but I don't know if it's the right price never do um it's 20 bucks a month unlimited users unlimited usage one price no chart no table no contact us just a price tag like if you went and bought a pair of jeans or peanut butter
3:02:49
it'd be like talk to the sales rep they're going to look you up and down they're going to say well how much how much should this person pay >> right what watch are you wearing all the things right so it's 20 bucks but we give you a th000 cards for free so there's There's no time limit on the trial. You get a,000 cards for free. And
3:03:02
You get a,000 cards for free.
3:03:02
And if you never use a card is like a, you know, like a to-do item or something. >> Sure.
3:03:07
>> If you never use them up, it's free forever. >> Okay.
3:03:10
>> And you can also run it for free if you want to run it yourself. >> Open source. Yeah. >> Yeah.
3:03:14
So, we're basically just serving as a host.
3:03:16
If you want >> to just turn it on, sign up, and be going.
3:03:19
We'll host it for 20 bucks currently.
3:03:21
Look, this is an introductory price.
3:03:22
We could change the price 6 months from now.
3:03:24
If we do, we'll let people lock in where they were.
3:03:26
We're not going to change prices on them, but we might raise it.
3:03:30
We I don't even know what we'll do, but we wanted to pick a number that was fair.
3:03:32
The other thing I want to do is I want to price this more like an accessory.
3:03:37
>> This is not the only tool.
3:03:37
Like, you know, the software industry is interesting because it thinks that whatever it makes, it's the only thing anyone ever needs, right?
3:03:42
The thing is is people need a lot of different things.
3:03:46
And so, >> Fizzy is not going to be the only thing you have.
3:03:49
It might be one of the many things you might use.
3:03:51
And so, we kind of price it that way. It's like an accessory.
3:03:54
20 bucks a month, kind of a no-brainer.
3:03:56
unlimited users, >> uh, cancel any time, no upfront anything, and it just feels like that's the right place to start.
3:04:01
We'll see where we end up, but that feels good for now.
3:04:05
>> If you if I pay you to host it, where is it hosted? >> Um, China.
3:04:10
[laughter] >> No, so it's hosted.
3:04:13
We have we have we have a few different data centers, so it's not in the Well, it's in it's in our >> Yeah.
3:04:19
What I'm getting at is like is like it would be easy to just throw this on AWS, but like you're the one company that doesn't just do that, right? That's right.
3:04:26
So, we have a data center in [laughter] in Chicago.
3:04:28
We have one in in Amsterdam.
3:04:30
We have one in uh North Carolina.
3:04:33
>> Uh so, we have in a few different spots.
3:04:35
Um and uh it's all on our hardware and other people's data centers where we rent space and data centers. >> Yeah.
3:04:40
>> That said, again, um you can also if you just don't trust us, don't want us to do it, you can put on your own stuff, including like a simple droplet like a digital ocean something, whatever you can find that that will host something basic will work for for this as well.
3:04:52
I mean, you actually can host it in Alibaba cloud if you want. It's open source.
3:04:57
That's the whole point of open source.
3:04:59
I could put it on >> I hope someone does.
3:05:01
>> There's a AI company that uh recently had a code red.
3:05:04
Uh have you ever had a code red ever once? >> Not like that.
3:05:10
Uh not like a competitive pressure code red.
3:05:15
Let's make sure we kind of focus on this competitor, but we've like screwed up >> and had all hands on deck to fix something.
3:05:22
something. I mean there was a moment I think probably >> did you learn did you ever learn the the like did you ever get overly fixated on a competitor and sort of like learn that because because there's there's like that's like YC like uh uh just law right like don't overly focus on competitors
3:05:40
like you're probably not going to die as a company because of your competitor you die because of I think they say like indigestion or something like that >> right most wounds are self-inflicted I mean but but sometimes you have to to actually have a a lesson be fully ingrained, you you have to learn it the hard way. I'm curious if if that was
3:05:58
I'm curious if if that was ever the case.
3:06:02
>> Um I I think there was one time when way back when um we used to have a product, we still have a new product now called Campfire, but way back in 2006, we launched Campfire, which is a real-time chat, group chat. >> Yeah.
3:06:13
>> And back then, we could not shove this down people's throats.
3:06:15
Like nobody understood group chat for a business.
3:06:16
It just was very very hard to sell and to move and was a very small product for us.
3:06:23
>> And then Slack came out >> and I saw it.
3:06:24
I remember oh [Β __Β ] like yeah >> they nailed it. Like we we just >> Yeah.
3:06:30
>> It was crazy because them nailing it was it was IRC.
3:06:32
Like I used IRC back in the day and the hashtag channels like everything like there were all the primitives had been like battle tested in IRC.
3:06:40
The other thing is Slack doesn't feel like that outside of the world in ter even from a I'm sure you have opinions on Slack's like design, but it it it doesn't even feel that like you guys probably could see that and be like, "Oh, that that's like >> like the the design was opinionated and you know, fun. >> It felt fun. Slack felt fun."
3:06:59
Um I mean IRC of course was is very geeky and whatever, but yeah, the fundamentals were there, but Slack had a wonderful onboarding experience. It felt fun.
3:07:07
They had great integrations.
3:07:09
they just kind of like totally leaprogged us in that world and and that was like fine, but it it did it was the first time I felt like I felt that sort of nervousness in my stomach.
3:07:19
Um, now I didn't feel it against our business because Base Camp is very different kind of product and it was fine, but it was Campfire specifically cuz I was frustrated.
3:07:28
I was trying to figure out how to make it better and then I saw them come out like oh [Β __Β ] like yeah that that that's how you do it.
3:07:36
So, that was one time, but but I I just don't think there's any reason to focus on competitors.
3:07:40
I I just don't You can't control them.
3:07:42
You don't know what they're going to do.
3:07:44
You don't know if they're going to be around in 3 months or 3 years.
3:07:46
You don't have the same economics as they do. >> Yeah.
3:07:49
>> Um so, it doesn't really make sense.
3:07:51
Like, for example, I'll take Hey, our email service, hey. com.
3:07:55
>> Um we have 40 some odd thousand paying customers for Hey, right.
3:07:58
Which is >> if if we if you were Gmail, it'd be an absolute abject failure to only have 40,000 paying customers if you're the Google shut down years ago >> in seconds, right?
3:08:08
But for us, it's a multi-million dollar business cuz we have 60 people here.
3:08:11
So for us, it's a great business.
3:08:13
So like I can't go, well, Gmail is killing us. They're not killing us.
3:08:17
They're doing their thing. We're doing our thing.
3:08:18
So I think >> you've got to, in my opinion, the only person you actually compete with are your own economics.
3:08:23
Like that's not a person, but the only thing you compete with are your own economics.
3:08:27
If you can make it work, you can make it viable, you're fine. You can't >> your cost.
3:08:33
You compete with your cost.
3:08:34
>> Competing with your costs. Yeah.
3:08:34
Every business needs an AI noteaker.
3:08:35
What are your opinions on AI notetakers?
3:08:38
If they join the call, are you admitting them or are you letting them sit? >> I'm pretty harsh.
3:08:43
I always let them sit out in the cold. I never let them in.
3:08:48
[laughter] >> We don't we don't have meetings.
3:08:49
We don't So, I don't I don't even I couldn't even invite one in if I wanted to.
3:08:53
We just we don't we don't do that.
3:08:55
But I have I will say I have been in a few calls recently that other people have set up and there's been like an AI transcript and it has been quite handy.
3:09:02
It's really pretty impressive when it works really well.
3:09:03
Strangely, Apple can't seem to get voicemail transcriptions to work at all.
3:09:07
Have you >> I mean, Apple is is just struggling with all the all the basics on transcription.
3:09:14
Even just talking to your phone and like whisper works.
3:09:16
It works in the Chat GBD app.
3:09:19
It works everywhere else.
3:09:19
Apple just has not implemented it properly.
3:09:23
And it's and it's not it's not crazy AI god.
3:09:25
Like it's literally just take the words that I'm saying and write them down verbatim.
3:09:28
And that is a huge and that's a huge benefit because if you're in a business call sometimes I just want to search the actual transcript.
3:09:35
I don't even need you to summarize it or put action items or go do things for me.
3:09:38
Not agentic none of that.
3:09:41
Just actually write down exactly what I said so that when I say you know uh you know we had you know when I say AWS or whatever I can go search for when that happened in the transcript.
3:09:53
And a lot of a lot of companies just haven't even been able to implement that. It's been weird. >> I agree.
3:09:56
>> I agree. I think I think frankly that is one of the best use cases of I don't even know it's not even AI though it's just it's trans great transcription software is is very very handy and I think like this is the thing like it's it's transcription software has been around for a long time it's gotten
3:10:09
better and better and better but it's not like AI really you know in other ways it has been AI for 20 years it's been the original AI in many ways you know throw a bunch of data at it and and try and estimate what things are even like OCR these similar things they're just per they're not AGI, they're AI in the sense they're narrow. It's it's the
3:10:28
It's it's the recommendation algorithm on YouTube or Tik Tok or in Netflix or you know this specific you upload a you take a picture of a receipt.
3:10:38
Does it understand the text in there?
3:10:40
Even if it's kind of a dark photo, yes, that's specific narrow AI and that's great, but we need to actually get those things working on our phones.
3:10:47
[laughter] >> We left out busy by the way.
3:10:50
We made a conscious effort.
3:10:52
We actually had some for a while and pulled it out and had it back in and pulled it out.
3:10:57
>> I'm just like I want to remove from this.
3:10:58
I don't want to add intelligence.
3:11:00
I want to remove from the software.
3:11:00
So it's just so straightforward that it just works and you don't even feel like god I wish I had AI for this or for that. So V1, no AI.
3:11:08
We'll see what happens down the road. Again, it's open source.
3:11:13
>> So the really interesting thing with with Fizzy is that there is a world where you can just actually sit back and do nothing on AI.
3:11:18
And if AI is real and valuable to your users, they will get it stuff down their throats via their OS, via their browser, because Atlas is going to be trying to jump up.
3:11:29
Perplexity Comet is going to be trying puppeteering their their fizzy uh and and the rest of the the rest of the system that they're using, whether it's their phone or their laptop or their desktop, like it's going to bring the AI to bear with computer use.
3:11:42
And so you might never have to build it. >> Yeah, >> I think so.
3:11:45
In fact, this is actually interesting really quick.
3:11:47
Um, recently OpenAI added a basecam connector to chat GBT and we didn't we didn't even do anything.
3:11:54
So, they did all the work >> and they just sent us an email saying, "Hey, we're launching this base cam connector like in a few weeks." Like, great.
3:12:01
I'm like, "This is fantastic.
3:12:01
We don't have an MCP server. They just did it."
3:12:05
And so, I just think more to your point, I think more and more of that's going to happen, which is it's going to be available in the OS or someone else is going to do it or whatever.
3:12:11
and to to spend all this time to build it into the product specifically.
3:12:15
I just don't feel like it's the right the best use of initial an initial V1 should be focused on the product itself and not the other things that it could possibly do.
3:12:23
Again, later on maybe there's stuff that comes in.
3:12:26
Maybe people via the open source version submit some PRs that have some AI stuff.
3:12:31
I we'll see where it goes, but we didn't need it for V1.
3:12:34
>> Yeah, there's just always a question of where the AI lives.
3:12:35
where the AI lives. like do you need to go and pre-train your own model to answer questions or if you set up a good knowledge base will you just get sucked into the next pre-train automatically and you can just go to chatbt and ask about you and you'll be there anyway >> I want to I want to keep hanging out for
3:12:51
an hour but we we do we do have to wrap the show because we're going to look at a a studio >> we have one last question >> yeah one last question from our mutual friend David >> Cra we get a could we get a wrist check what are you rocking on launch day >> I might be the only person that coordinates their watch with their software. It's It's possible. It's It's possible.
3:13:10
>> So, I'm wearing today I'm just wearing a um I'll take it off cuz I I don't know how to quite hold it up otherwise.
3:13:13
Um >> this is a vint just a vintage Hoyer from 1974, which is birth a birth year watch. Let's see. Hang on. >> Whoa. >> Oh, birthier watch. Hang on. Hang on.
3:13:22
I love the Yeah, hang on. Let me see.
3:13:24
The focus is hard, but there you go. There you go. There you go. >> There we go.
3:13:31
Love the I love the orange. It's colorful. Fizzy is colorful. Fizzy is full of color.
3:13:37
It's the most colorful watch I own for the most colorful product we've ever made.
3:13:42
>> I knew I knew you literally said I knew you were going to match.
3:13:44
I knew I knew it was going to be intentional. >> This is so good. >> It's a little sad. A little bit. It's a little bit sad. Sad. It's fun. >> It's joy. This is joy. >> It's joy. >> This is amazing.
3:13:55
>> I didn't I didn't have like a what? Like a green What?
3:13:56
Why were you guys wearing a green jacket a few days ago? What was that about?
3:13:59
>> It was Shopify Black Friday. We did.
3:13:59
We were celebrating commerce online and so we wore just green solid green suit. No, it's Shopify.
3:14:06
Signature color is is green. >> Oh, shop. I didn't even know that. Yeah, green. Yeah. Yeah. Okay.
3:14:13
>> It is a little confusing because we use a dark green in our brand theme and so it actually paired up pretty nicely.
3:14:17
We also have yellow suits for when we uh for when there's big ramp news.
3:14:21
We will wear solid yellow.
3:14:23
Yeah, you probably seen those. Those are fun. >> I've seen that. >> Hard to hard to miss.
3:14:27
Uh, well, Jason, uh, open invite to the studio.
3:14:29
We'd love to hang out for like a full hour. Everybody >> loving it.
3:14:35
Everyone's having >> everybody.
3:14:36
Uh, >> let's do it sometime. I'd love to.
3:14:37
I think I got an email about about that.
3:14:39
So, we'll figure that out. >> Amazing. Awesome. >> Appreciate it. >> All right.
3:14:41
Thanks for having me on to the whole team on the line. Talk to you soon. Very exciting. >> Thank you. See you. >> Bye. >> Getbzzle. com.
3:14:46
Shop over 26,500 luxury watches that you're not going to you're not going to believe it, but this was actually the next ad read up.
3:14:51
Uh, fully authenticated inhouse by Bezel's team of experts.
3:14:55
And uh we got to close out the show.
3:14:59
So I'm going to tell you about wanderer. com.
3:15:00
Book a wander with inspiring views, hotel grade amenities, dreaming beds, top tier cleaning, 24/7 concier service.
3:15:04
There are so many more posts that I want to get to.
3:15:06
Uh there's a lot to them.
3:15:09
>> There's a new Arena Mag out.
3:15:09
You got to go to Arena Mag. Check it out.
3:15:11
We are featured in this Arena Mag issue 006, the three martini launch.
3:15:16
We had Julia on the show, of course, to talk about it, but now it's it's in print.
3:15:19
Uh there's a lot else going on >> and we will be back tomorrow. Yes. Sorry to cut it off. A lot of fun.
3:15:28
>> I would be in a very bad place if we weren't podcasting tomorrow, but fortunately we are.
3:15:31
So, we'll see you tomorrow.
3:15:34
>> Leave us five stars on Apple Podcast. Goodbye. Have a